Add DG2 OA test
Subsystems (2)
v6 cover Add DG2 OA test
Add tests for changes in perf OA for DG2.
v2:
- Rebase
- Reposting since some patches did no land in the ML
v3:
- Fix author/sign-off name mismatch in one patch
v4:
- Add review comments
- Drop below patches
"i915/perf: Bump timestamp tolerance for DG1" (Not able to reproduce it with latest drm-tip)
"i915/perf: Add support for 64-bit counters" "i915/perf: Add a test for non-power-of-2 oa reports" "i915/perf: Treat ticks as 64 bit" "i915/perf: Treat timestamp as 64 bit value" "i915/perf: Fix DG2 A0 report header" (Separate out 64-bit OA formats from this series)
v5:
- Add patches from lionel for updated metric sets
v6:
- ACM metrics patch message size is large and gets dropped by the email server. Break the patch into smaller chunks.
Lionel Landwerlin (21):
lib/i915/perf: implement report accumulation for new format lib/i915/perf: fixup conversion script for XEHPSDV lib/i915/perf: make warning message more helpful lib/i915/perf: expose new operators for codegen tools/i915-perf-recorder: add ability to select device lib/i915/perf: fixup report validity
lib/i915/perf: add a helper to read timestamps lib/i915/perf: store bit shifting required for OA timestamps lib/i915/perf: indentation fix tools/i915-perf-recorder: capture OA & CS frequencies tools/i915-perf: make timestamp range easier to compare tools/i915-perf: printout CPU clock used tools/i915-perf: record remaining perf data on exit lib/i915/perf: add support for new EuDualSubslicesTotalCount var lib/i915/perf-config: extend the device info i915/perf: update import script
lib/i915/perf: add a raw timestamp utility lib/i915/perf: add helper function to get report reason tools/i915-perf: add option to printout reports data lib/i915: prepare codegen for new ACM/DG2 variables lib/i915/perf: Add ACM GT3 metrics
Umesh Nerlige Ramappa (15):
i915/perf: Check regularly if we are done reading reports i915/perf: Fix OA short_reads test
i915/perf: Check return value from getparam i915/perf: Limit sseu-config tests for gen11 i915/perf: Account for OA sampling interval in polling test i915/perf: Define OA report types and fix oa-formats test i915/perf: Use ARRAY_SIZE consistently for num_properties i915/perf: Use gt in perf tests and lib i915/perf: Explicitly state rendercopy needs for a test i915/perf: Skip tests that use rendercopy i915/perf: Add OA formats for DG2
i915/perf: Fix CS timestamp vs OA timstamp mismatch i915/perf: Wait longer for rc6 residency in DG2 lib/i915/perf: Add ACM GT1 metrics
lib/i915/perf: Add ACM GT2 metrics
include/drm-uapi/i915_drm.h | 10 + lib/i915/perf-configs/codegen.py | 110 +- lib/i915/perf-configs/guids.xml | 1909 + lib/i915/perf-configs/mdapi-xml-convert.py | 244 +- lib/i915/perf-configs/oa-acmgt1.xml | 76252 +++++ lib/i915/perf-configs/oa-acmgt2.xml | 118222 +++++++ lib/i915/perf-configs/oa-acmgt3.xml | 250108 +++++++++++++++ lib/i915/perf-configs/oa-bdw.xml | 2 +- lib/i915/perf-configs/oa_guid_registry.py | 15 +- .../perf-configs/perf-metricset-codegen.py | 13 + lib/i915/perf-configs/update-guids.py | 23 +- lib/i915/perf.c | 332 +- lib/i915/perf.h | 115 +- lib/i915/perf_data_reader.c | 29 +- lib/meson.build | 5 +- tests/i915/perf.c | 260 +- tools/i915-perf/i915_perf_reader.c | 117 +- tools/i915-perf/i915_perf_recorder.c | 112 +- 18 files changed, 447626 insertions(+), 252 deletions(-) create mode 100644 lib/i915/perf-configs/oa-acmgt1.xml create mode 100644 lib/i915/perf-configs/oa-acmgt2.xml create mode 100644 lib/i915/perf-configs/oa-acmgt3.xml
Signature · 1 line
2.25.1
PATCH v6 1/36 i915/perf: Check regularly if we are done reading reports
If we pass a user buffer as large as the OA buffer, the OA unit will return all available reports. In the second part of the buffer-fill/enable-disable tests, we are collecting all reports within a time window, but we only check the time window in the outer while loop. We may have read well past that window if OA returns more reports. Fix this rare failure by checking for the window within the inner for loop.
--- tests/i915/perf.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
tests/i915/perf.c +10 -0
@@ -2522,6 +2522,9 @@ test_buffer_fill(void) first_timestamp = report[1 ]; last_timestamp = report[1 ]; +if (((last_timestamp - first_timestamp) * oa_period) < (fill_duration /2 )) +break ; +if (oa_report_is_periodic (oa_exponent, report)) {memcpy (last_periodic_report, report,sizeof (last_periodic_report));
@@ -2537,6 +2540,8 @@ test_buffer_fill(void)do_ioctl (stream_fd,I915_PERF_IOCTL_DISABLE ,0 ); +igt_debug ("first ts = %u, last ts = %u\n" , first_timestamp, last_timestamp); + igt_debug("%f < %zu < %f\n", report_size * n_full_oa_reports * 0.45, n_periodic_reports * report_size,
@@ -2732,6 +2737,9 @@ test_enable_disable(void)oa_report_is_periodic (oa_exponent, report), oa_report_get_ctx_id(report)); +if (((last_timestamp - first_timestamp) * oa_period) < (fill_duration /2 )) +break ; +if (oa_report_is_periodic (oa_exponent, report)) {memcpy (last_periodic_report, report,sizeof (last_periodic_report));
@@ -2755,6 +2763,8 @@ test_enable_disable(void)do_ioctl (stream_fd,I915_PERF_IOCTL_DISABLE ,0 ); +igt_debug ("first ts = %u, last ts = %u\n" , first_timestamp, last_timestamp); + igt_debug("%f < %zu < %f\n", report_size * n_full_oa_reports * 0.45, n_periodic_reports * report_size,
Signature · 1 line
2.25.1
PATCH v6 2/36 i915/perf: Fix OA short_reads test
In reply to Umesh Nerlige RamappaWith 5000 ns oa exponent, report lost errors can occur. A report lost header can be read by the user even when there are no valid OA report samples, so we get an 8 byte read. This defeats the EFAULT check.
When checking for EFAULT on a sampled report, discard any report lost errors.
--- tests/i915/perf.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
tests/i915/perf.c +6 -1
@@ -2853,8 +2853,13 @@ test_short_reads(void)/* A read that can't return a single record because it would result * in a fault on buffer overrun should result in an EFAULT error... +* +* Make sure to weed out all report lost errors before verifying EFAULT. */ - ret =read (stream_fd, pages + page_size -16 , page_size); + header = (void *)(pages + page_size -16 ); +do { + ret =read (stream_fd, header, page_size); + }while (ret >0 && header->type ==DRM_I915_PERF_RECORD_OA_REPORT_LOST );igt_assert_eq (ret,-1 );igt_assert_eq (errno,EFAULT );
Signature · 1 line
2.25.1
PATCH v6 3/36 i915/perf: Check return value from getparam
In reply to Umesh Nerlige RamappaIn some cases getparam could fail. Check return from getparam and fail early on.
--- lib/i915/perf.c | 28 +++++++++++++++------------- 1 file changed, 15 insertions(+), 13 deletions(-)
lib/i915/perf.c +15 -13
@@ -287,19 +287,16 @@ intel_perf_for_devinfo(uint32_t device_id,return perf; } -static uint32_t -getparam (int drm_fd,uint32_t param) +static int +getparam (int drm_fd,uint32_t param,uint32_t *val) { -struct drm_i915_getparam gp; -int val =-1 ; - -memset (&gp,0 ,sizeof (gp)); - gp.param = param; - gp.value = &val; +struct drm_i915_getparam gp; -perf_ioctl (drm_fd,DRM_IOCTL_I915_GETPARAM , &gp); +memset (&gp,0 ,sizeof (gp)); + gp.param = param; + gp.value = (int *)val; -return val; +return perf_ioctl (drm_fd,DRM_IOCTL_I915_GETPARAM , &gp); } static bool
@@ -415,9 +412,9 @@ open_master_sysfs_dir(int drm_fd)struct intel_perf *intel_perf_for_fd (int drm_fd) { -uint32_t device_id =getparam (drm_fd,I915_PARAM_CHIPSET_ID ); -uint32_t device_revision =getparam (drm_fd,I915_PARAM_REVISION ); -uint32_t timestamp_frequency =getparam (drm_fd,I915_PARAM_CS_TIMESTAMP_FREQUENCY ); +uint32_t device_id; +uint32_t device_revision; +uint32_t timestamp_frequency;uint64_t gt_min_freq;uint64_t gt_max_freq;struct drm_i915_query_topology_info *topology;
@@ -434,6 +431,11 @@ intel_perf_for_fd(int drm_fd) }close (sysfs_dir_fd); +if (getparam (drm_fd,I915_PARAM_CHIPSET_ID , &device_id) || +getparam (drm_fd,I915_PARAM_REVISION , &device_revision) || +getparam (drm_fd,I915_PARAM_CS_TIMESTAMP_FREQUENCY , ×tamp_frequency)) +return NULL ; + topology =query_topology (drm_fd);if (!topology)return NULL ;
Signature · 1 line
2.25.1
PATCH v6 4/36 i915/perf: Limit sseu-config tests for gen11
In reply to Umesh Nerlige RamappaOn Gen11, the Media driver will need to limit dispatch to only big subslices (usually half the EUs depending on EU fusing). If media, render and OA are running in parallel, then OA will lose the NOA config because half of the EUs will be turned off by media. To avoid this with concurrent use cases, an sseu config to keep all EUs powered on is passed to OA in the i915_perf_open_ioctl params. This configuration is made part of the render contexts, so that it's applied when 3d contexts are running. Since this is an issue on gen11, limit sseu-config tests to gen11.
--- tests/i915/perf.c | 1 + 1 file changed, 1 insertion(+)
tests/i915/perf.c +1 -0
@@ -5112,6 +5112,7 @@ igt_main igt_subtest_group { igt_fixture {igt_require (i915_perf_revision (drm_fd) >=4 ); +igt_require (intel_gen (devid) ==11 ); }igt_describe ("Verify invalid SSEU opening parameters" );
Signature · 1 line
2.25.1
PATCH v6 5/36 i915/perf: Account for OA sampling interval in polling test
In reply to Umesh Nerlige RamappaThe polling test reads all the reports and then expects EAGAIN on a subsequent read since it has consumed all available data. This works well for OA sampling periods in the order of 10s of ms. For smaller sampling periods like 500 us, we end up reading valid data and the test fails. Fix the check to account for the OA sampling period.
--- tests/i915/perf.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)
tests/i915/perf.c +12 -4
@@ -2220,14 +2220,22 @@ test_polling(uint64_t requested_oa_period, bool set_kernel_hrtimer, uint64_t ker n_extra_iterations++;/* At this point, after consuming pending reports (and hoping -* the scheduler hasn't stopped us for too long we now -* expect EAGAIN on read. +* the scheduler hasn't stopped us for too long) we now expect +* EAGAIN on read. While this works most of the times, there are +* some rare failures when the OA period passed to this test is +* very small (say 500 us) and that results in some valid +* reports here. To weed out those rare occurences we assert +* only if the OA period is >= 40 ms because 40 ms has withstood +* the test of time on most platforms (ref: subtest: polling). */ while ((ret =read (stream_fd, buf,sizeof (buf))) <0 && errno ==EINTR ) ; -igt_assert_eq (ret,-1 ); -igt_assert_eq (errno,EAGAIN ); + +if (requested_oa_period >=40000000 ) { +igt_assert_eq (ret,-1 ); +igt_assert_eq (errno,EAGAIN ); + } n++; }
Signature · 1 line
2.25.1
PATCH v6 6/36 i915/perf: Define OA report types and fix oa-formats test
In reply to Umesh Nerlige RamappaWith newer gen12 platforms, formats used for OAG buffer and MI_REPORT_PERF_FORMAT differ slightly. This causes issues when the oa-formats test does a sanity check. Run oa-formats for OAG reports only.
OAR formats must be tested with the gen12-mi-rpc and gen12-unprivileged-single-ctx-counters.
v2: Clarify commit message
--- tests/i915/perf.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
tests/i915/perf.c +11 -0
@@ -95,6 +95,13 @@ struct accumulator {uint64_t deltas[MAX_RAW_OA_COUNTERS ]; }; +enum { +OAG , +OAR , + +MAX_OA_TYPE , +}; +struct oa_format {const char *name;size_t size;
@@ -108,6 +115,7 @@ struct oa_format {int n_b;int c_off;int n_c; +int oa_type; };static struct oa_format hsw_oa_formats[I915_OA_FORMAT_MAX ] = {
@@ -1483,6 +1491,9 @@ test_oa_formats(void)if (!format.name)/* sparse, indexed by ID */ continue ; +if (format.oa_type !=OAG )/* sparse, indexed by ID */ +continue ; +igt_debug ("Checking OA format %s\n" , format.name); open_and_read_2_oa_reports(i,
Signature · 1 line
2.25.1
PATCH v6 7/36 i915/perf: Use ARRAY_SIZE consistently for num_properties
In reply to Umesh Nerlige RamappaReplace all variations of calculating the num_properties with ARRAY_SIZE.
--- tests/i915/perf.c | 62 +++++++++++++++++++++++------------------------ 1 file changed, 30 insertions(+), 32 deletions(-)
tests/i915/perf.c +30 -32
@@ -86,8 +86,6 @@ IGT_TEST_DESCRIPTION("Test the i915 perf metrics streaming interface");#define MAX_OA_BUF_SIZE (16 * 1024 * 1024) -#define NUM_PROPERTIES (p) (sizeof(p) / (2 * sizeof(uint64_t))) -struct accumulator {#define MAX_RAW_OA_COUNTERS 62enum drm_i915_oa_format format;
@@ -1025,7 +1023,7 @@ test_system_wide_paranoid(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_FD_NONBLOCK , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1051,7 +1049,7 @@ test_system_wide_paranoid(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_FD_NONBLOCK , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };write_u64_file ("/proc/sys/dev/i915/perf_stream_paranoid" ,0 );
@@ -1082,7 +1080,7 @@ test_invalid_open_flags(void) };struct drm_i915_perf_open_param param = { .flags = ~0 ,/* Undefined flag bits set! */ - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1104,7 +1102,7 @@ test_invalid_oa_metric_set_id(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_FD_NONBLOCK , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1138,7 +1136,7 @@ test_invalid_oa_format_id(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_FD_NONBLOCK , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1170,7 +1168,7 @@ test_missing_sample_flags(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1310,7 +1308,7 @@ open_and_read_2_oa_reports(int format_id, };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1777,7 +1775,7 @@ test_invalid_oa_exponent(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1812,7 +1810,7 @@ test_low_oa_exponent_permissions(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };uint64_t oa_period, oa_freq;
@@ -1875,7 +1873,7 @@ test_per_context_mode_unprivileged(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -1951,8 +1949,8 @@ test_blocking(uint64_t requested_oa_period, bool set_kernel_hrtimer, uint64_t ke .flags = I915_PERF_FLAG_FD_CLOEXEC | I915_PERF_FLAG_DISABLED, .num_properties = set_kernel_hrtimer ? -NUM_PROPERTIES (properties) : - NUM_PROPERTIES(properties) - 1, +ARRAY_SIZE (properties) /2 : + (ARRAY_SIZE(properties) / 2) - 1, .properties_ptr = to_user_pointer(properties), };uint8_t buf[1024 *1024 ];
@@ -2111,8 +2109,8 @@ test_polling(uint64_t requested_oa_period, bool set_kernel_hrtimer, uint64_t kerI915_PERF_FLAG_DISABLED | I915_PERF_FLAG_FD_NONBLOCK, .num_properties = set_kernel_hrtimer ? -NUM_PROPERTIES (properties) : - NUM_PROPERTIES(properties) - 1, +ARRAY_SIZE (properties) /2 : + (ARRAY_SIZE(properties) / 2) - 1, .properties_ptr = to_user_pointer(properties), };uint8_t buf[1024 *1024 ];
@@ -2302,7 +2300,7 @@ static void test_polling_small_buf(void) .flags = I915_PERF_FLAG_FD_CLOEXEC | I915_PERF_FLAG_DISABLED | I915_PERF_FLAG_FD_NONBLOCK, - .num_properties = NUM_PROPERTIES(properties), + .num_properties = ARRAY_SIZE(properties) / 2, .properties_ptr = to_user_pointer(properties), };uint32_t test_duration =80 *1000 *1000 ;
@@ -2402,7 +2400,7 @@ gen12_test_oa_tlb_invalidate(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_DISABLED , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };int num_reports1, num_reports2, num_expected_reports;
@@ -2593,7 +2591,7 @@ test_non_zero_reason(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };struct drm_i915_perf_record_header *header;
@@ -2678,7 +2676,7 @@ test_enable_disable(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_DISABLED ,/* Verify we start disabled */ - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };int buf_size =65536 * (256 +sizeof (struct drm_i915_perf_record_header ));
@@ -2827,7 +2825,7 @@ test_short_reads(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };size_t record_size =256 +sizeof (struct drm_i915_perf_record_header );
@@ -2919,7 +2917,7 @@ test_non_sampling_read_error(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };int ret;
@@ -2955,7 +2953,7 @@ test_disabled_read_error(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_DISABLED ,/* XXX: open disabled */ - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };uint32_t oa_report0[64 ];
@@ -3112,7 +3110,7 @@ test_mi_rpc(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };struct buf_ops *bops =buf_ops_create (drm_fd);
@@ -3215,7 +3213,7 @@ hsw_test_single_ctx_counters(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -4196,7 +4194,7 @@ test_rc6_disable(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };unsigned long rc6_start, rc6_end, rc6_enabled;
@@ -4252,7 +4250,7 @@ test_stress_open_close(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_DISABLED ,/* XXX: open disabled */ - .num_properties =NUM_PROPERTIES (properties), + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };
@@ -4348,8 +4346,8 @@ test_global_sseu_config_invalid(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_DISABLED ,/* XXX: open disabled */ - .num_properties =NUM_PROPERTIES (properties), - .properties_ptr =to_user_pointer (properties), + .num_properties =ARRAY_SIZE (properties) /2 , + .properties_ptr =to_user_pointer (properties), };memset (&default_sseu,0 ,sizeof (default_sseu));
@@ -4423,8 +4421,8 @@ test_global_sseu_config(void)struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC |I915_PERF_FLAG_DISABLED ,/* XXX: open disabled */ - .num_properties =NUM_PROPERTIES (properties), - .properties_ptr =to_user_pointer (properties), + .num_properties =ARRAY_SIZE (properties) /2 , + .properties_ptr =to_user_pointer (properties), };memset (&default_sseu,0 ,sizeof (default_sseu));
@@ -4856,7 +4854,7 @@ test_i915_ref_count(void) };struct drm_i915_perf_open_param param = { .flags =I915_PERF_FLAG_FD_CLOEXEC , - .num_properties =sizeof (properties) /16 , + .num_properties =ARRAY_SIZE (properties) /2 , .properties_ptr =to_user_pointer (properties), };unsigned baseline, ref_count0, ref_count1;
Signature · 1 line
2.25.1
PATCH v6 8/36 i915/perf: Use gt in perf tests and lib
In reply to Umesh Nerlige RamappaEnable perf tests to use gt information for sysfs attributes.
--- lib/i915/perf.c | 36 ++++++++++++++++++++++++++++++++++-- tests/i915/perf.c | 27 ++++++++++++++++++++------- 2 files changed, 54 insertions(+), 9 deletions(-)
lib/i915/perf.c +34 -2
@@ -409,6 +409,35 @@ open_master_sysfs_dir(int drm_fd)return sysfs; } +typedef enum { +RPS_MIN_FREQ_MHZ , +RPS_MAX_FREQ_MHZ , + +RPS_MAX_ATTR , +}intel_sysfs_attr_id ; + +static const char *intel_sysfs_attr_name[2 ][RPS_MAX_ATTR ] = +{ + { +"gt_min_freq_mhz" , +"gt_max_freq_mhz" , + }, + { +"gt/gt0/rps_min_freq_mhz" , +"gt/gt0/rps_max_freq_mhz" , + }, +}; + +static const char * +intel_sysfs_attr_id_to_name (int sysfs_dirfd,intel_sysfs_attr_id id) +{ +assert (id <RPS_MAX_ATTR ); + +return !faccessat (sysfs_dirfd,"gt" ,O_RDONLY ,0 ) ? + intel_sysfs_attr_name[1 ][id] : + intel_sysfs_attr_name[0 ][id]; +} +struct intel_perf *intel_perf_for_fd (int drm_fd) {
@@ -424,8 +453,11 @@ intel_perf_for_fd(int drm_fd)if (sysfs_dir_fd <0 )return NULL ; -if (!read_sysfs (sysfs_dir_fd,"gt_min_freq_mhz" , >_min_freq) || - !read_sysfs (sysfs_dir_fd,"gt_max_freq_mhz" , >_max_freq)) { +#define read_sysfs_rps (fd, id, value) \ + read_sysfs(fd, intel_sysfs_attr_id_to_name(fd, id), value) + +if (!read_sysfs_rps (sysfs_dir_fd,RPS_MIN_FREQ_MHZ , >_min_freq) || + !read_sysfs_rps (sysfs_dir_fd,RPS_MAX_FREQ_MHZ , >_max_freq)) {close (sysfs_dir_fd);return NULL ; }
tests/i915/perf.c +20 -7
@@ -345,11 +345,11 @@ try_sysfs_read_u64(const char *path, uint64_t *val) }static unsigned long -sysfs_read (const char *path) +sysfs_read (enum i915_attr_id id) {unsigned long value; -igt_assert (igt_sysfs_scanf (sysfs, path,"%lu" , &value) ==1 ); +igt_assert (igt_sysfs_rps_scanf (sysfs, id,"%lu" , &value) ==1 );return value; }
@@ -4177,7 +4177,7 @@ gen12_test_single_ctx_render_target_writes_a_counter(void)static unsigned long rc6_residency_ms (void ) { -return sysfs_read ("power/rc6_residency_ms" ); +return sysfs_read (RC6_RESIDENCY_MS ); } static void
@@ -4200,7 +4200,7 @@ test_rc6_disable(void)unsigned long rc6_start, rc6_end, rc6_enabled; rc6_enabled =0 ; -igt_sysfs_scanf (sysfs,"power/rc6_enable" ,"%lu" , &rc6_enabled); +igt_sysfs_rps_scanf (sysfs,RC6_ENABLE ,"%lu" , &rc6_enabled);igt_require (rc6_enabled);/* Verify rc6 is functional by measuring residency while idle */
@@ -4837,6 +4837,19 @@ done:return ref_count; } +static int perf_sysfs_open (int i915) +{ +int dirfd, gt; + +/* use the first available sysfs interface */ +for_each_sysfs_gt_dirfd (i915, dirfd, gt) +break ; + +igt_assert (dirfd !=-1 ); + +return dirfd; +} +/* check that an open i915 perf stream holds a reference on the drm i915 module * including in the corner case where the original drm fd has been closed. */
@@ -4872,7 +4885,7 @@ test_i915_ref_count(void) drm_fd =__drm_open_driver (DRIVER_INTEL );igt_require_intel (drm_fd); devid =intel_get_drm_devid (drm_fd); - sysfs =igt_sysfs_open (drm_fd); + sysfs =perf_sysfs_open (drm_fd);/* Note: these global variables are only initialized after calling * init_sys_info()...
@@ -4974,14 +4987,14 @@ igt_mainigt_require_gem (drm_fd); devid =intel_get_drm_devid (drm_fd); - sysfs =igt_sysfs_open (drm_fd); + sysfs =perf_sysfs_open (drm_fd);igt_require (init_sys_info ());write_u64_file ("/proc/sys/dev/i915/perf_stream_paranoid" ,1 );write_u64_file ("/proc/sys/dev/i915/oa_max_sample_rate" ,100000 ); - gt_max_freq_mhz =sysfs_read ("gt_boost_freq_mhz" ); + gt_max_freq_mhz =sysfs_read (RPS_RP0_FREQ_MHZ ); render_copy =igt_get_render_copyfunc (devid);igt_require_f (render_copy,"no render-copy function\n" );
Signature · 1 line
2.25.1
PATCH v6 9/36 i915/perf: Explicitly state rendercopy needs for a test
In reply to Umesh Nerlige RamappaLet tests explicitly require rendercopy before running. This helps in temporarily blocking these tests until rendercopy is available on new platforms.
--- tests/i915/perf.c | 25 ++++++++++++++++++++++--- 1 file changed, 22 insertions(+), 3 deletions(-)
tests/i915/perf.c +22 -3
@@ -1545,6 +1545,9 @@ static void load_helper_set_load(enum load load)static void load_helper_run (enum load load) { +if (!render_copy) +return ; +/* * FIXME fork helpers won't get cleaned up when started from within a * subtest, so handle the case where it sticks around a bit too long.
@@ -1577,12 +1580,20 @@ static void load_helper_run(enum load load)static void load_helper_stop (void ) { +if (!render_copy) +return ; +kill (lh.igt_proc.pid,SIGUSR1 );igt_assert (igt_wait_helper (&lh.igt_proc) ==0 ); }static void load_helper_init (void ) { +if (!render_copy) { +igt_info ("Running test without render_copy\n" ); +return ; + } + lh.devid =intel_get_drm_devid (drm_fd);/* MI_STORE_DATA can only use GTT address on gen4+/g33 and needs
@@ -1603,7 +1614,12 @@ static void load_helper_init(void)static void load_helper_fini (void ) { -int i915 =buf_ops_get_fd (lh.bops); +int i915; + +if (!render_copy) +return ; + + i915 =buf_ops_get_fd (lh.bops);if (lh.igt_proc.running)load_helper_stop ();
@@ -4997,7 +5013,6 @@ igt_main gt_max_freq_mhz =sysfs_read (RPS_RP0_FREQ_MHZ ); render_copy =igt_get_render_copyfunc (devid); -igt_require_f (render_copy,"no render-copy function\n" ); }igt_subtest ("non-system-wide-paranoid" )
@@ -5100,6 +5115,7 @@ igt_main igt_subtest("unprivileged-single-ctx-counters") {igt_require (IS_HASWELL (devid)); +igt_require_f (render_copy,"no render-copy function\n" );hsw_test_single_ctx_counters (); }
@@ -5113,6 +5129,7 @@ igt_main* For gen12 implement a separate test that uses only OAR */ igt_require (intel_gen (devid) >=8 &&intel_gen (devid) <12 ); +igt_require_f (render_copy,"no render-copy function\n" );gen8_test_single_ctx_render_target_writes_a_counter (); }
@@ -5128,8 +5145,10 @@ igt_maingen12_test_oa_tlb_invalidate ();igt_describe ("Measure performance for a specific context using OAR in Gen 12" ); - igt_subtest("gen12-unprivileged-single-ctx-counters") + igt_subtest("gen12-unprivileged-single-ctx-counters") { +igt_require_f (render_copy,"no render-copy function\n" );gen12_test_single_ctx_render_target_writes_a_counter (); + } }igt_subtest ("rc6-disable" )
Signature · 1 line
2.25.1
PATCH v6 10/36 i915/perf: Skip tests that use rendercopy
In reply to Umesh Nerlige RamappaRunning tests that require rendercopy sometimes leaves the DUT in an unsusable state. Further tests fail to run and DUT requires reboot. Since it affects CI efficiency, skip tests is rendercopy is missing.
--- tests/i915/perf.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-)
tests/i915/perf.c +16 -1
@@ -41,6 +41,7 @@#include "i915/gem.h" #include "i915/perf.h" #include "igt.h" +#include "igt_perf.h" #include "igt_sysfs.h" #include "drm.h"
@@ -4968,6 +4969,19 @@ static int i915_perf_revision(int fd)return value; } +static bool has_class_instance (int i915,uint16_t class,uint16_t instance) +{ +int fd; + + fd =perf_i915_open (i915,I915_PMU_ENGINE_BUSY (class, instance)); +if (fd >=0 ) { +close (fd); +return true; + } + +return false; +} +igt_main { igt_fixture {
@@ -5012,7 +5026,8 @@ igt_main gt_max_freq_mhz =sysfs_read (RPS_RP0_FREQ_MHZ ); - render_copy =igt_get_render_copyfunc (devid); +if (has_class_instance (drm_fd,I915_ENGINE_CLASS_RENDER ,0 )) + render_copy =igt_get_render_copyfunc (devid); }igt_subtest ("non-system-wide-paranoid" )
Signature · 1 line
2.25.1
PATCH v6 11/36 i915/perf: Add OA formats for DG2
In reply to Umesh Nerlige RamappaAdd OA formats for DG2.
v2:
- 40-bit A counter accumulation was broken due to incorrect DG2 format definition. Introduce first_a40 to fix the sanity check.
- Drop 64-bit OA formats
- Set the correct OAR format for mi-rpc
--- include/drm-uapi/i915_drm.h | 4 +++ tests/i915/perf.c | 50 +++++++++++++++++++++++++++++++------ 2 files changed, 46 insertions(+), 8 deletions(-)
include/drm-uapi/i915_drm.h +4 -0
@@ -2535,6 +2535,10 @@ enum drm_i915_oa_format {I915_OA_FORMAT_A12_B8_C8 , I915_OA_FORMAT_A32u40_A4u32_B8_C8, +/* DG2 */ + I915_OAR_FORMAT_A32u40_A4u32_B8_C8, + I915_OA_FORMAT_A24u40_A14u32_B8_C8, + I915_OA_FORMAT_MAX /* non-ABI */ };
tests/i915/perf.c +42 -8
128 lines not on this page · raw message
Signature · 1 line
2.25.1
PATCH v6 12/36 i915/perf: Fix CS timestamp vs OA timstamp mismatch
In reply to Umesh Nerlige RamappaCS timestamp can have an additional configuration that shifts the timestamp value by a specified number of bits. On DG2, this additional configuration is ignored by OA, so OA timestamps are missing the shift.
Add newly defined interface to get the exact OA timestamp frequency so that the deltas can be accurately calculated.
--- include/drm-uapi/i915_drm.h | 6 ++++++ lib/i915/perf.c | 6 +++++- tests/i915/perf.c | 25 ++++++++++++++++++++++++- 3 files changed, 35 insertions(+), 2 deletions(-)
include/drm-uapi/i915_drm.h +6 -0
@@ -749,6 +749,12 @@ typedef struct drm_i915_irq_wait {/* Query if the kernel supports the I915_USERPTR_PROBE flag. */ #define I915_PARAM_HAS_USERPTR_PROBE 56 +/* +* Frequency of the timestamps in OA reports. This used to be the same as the CS +* timestamp frequency, but differs on some platforms. +*/ +#define I915_PARAM_OA_TIMESTAMP_FREQUENCY 57 +/* Must be kept compact -- no holes and well documented */ typedef struct drm_i915_getparam {
lib/i915/perf.c +5 -1
@@ -464,7 +464,11 @@ intel_perf_for_fd(int drm_fd)close (sysfs_dir_fd);if (getparam (drm_fd,I915_PARAM_CHIPSET_ID , &device_id) || -getparam (drm_fd,I915_PARAM_REVISION , &device_revision) || +getparam (drm_fd,I915_PARAM_REVISION , &device_revision)) +return NULL ; + +/* if OA_TIMESTAMP_FREQUENCY is not supported, fall back to CS_TIMESTAMP_FREQUENCY */ +if (getparam (drm_fd,I915_PARAM_OA_TIMESTAMP_FREQUENCY , ×tamp_frequency) &&getparam (drm_fd,I915_PARAM_CS_TIMESTAMP_FREQUENCY , ×tamp_frequency))return NULL ;
tests/i915/perf.c +24 -1
@@ -442,6 +442,29 @@ gen8_read_report_reason(const uint32_t *report)return "unknown" ; } +static uint32_t +cs_timestamp_frequency (int fd) +{ +struct drm_i915_getparam gp = {}; +static uint32_t value =0 ; + +if (value) +return value; + + gp.param =I915_PARAM_CS_TIMESTAMP_FREQUENCY ; + gp.value = (int *)(&value); + +igt_assert_eq (igt_ioctl (fd,DRM_IOCTL_I915_GETPARAM , &gp),0 ); + +return value; +} + +static uint64_t +cs_timebase_scale (uint32_t u32_delta) +{ +return ((uint64_t )u32_delta *NSEC_PER_SEC ) /cs_timestamp_frequency (drm_fd); +} +static uint64_t timebase_scale (uint32_t u32_delta) {
@@ -4142,7 +4165,7 @@ static void gen12_single_ctx_helper(void)/* Sanity check that we can pass the delta to timebase_scale */ igt_assert (delta_ts64 <UINT32_MAX ); delta_oa32_ns =timebase_scale (delta_oa32); - delta_ts64_ns =timebase_scale (delta_ts64); + delta_ts64_ns =cs_timebase_scale (delta_ts64);igt_debug ("oa32 delta = %u, = %uns\n" , delta_oa32, (unsigned )delta_oa32_ns);
Signature · 1 line
2.25.1
PATCH v6 13/36 i915/perf: Wait longer for rc6 residency in DG2
In reply to Umesh Nerlige Ramappaigt@perf@rc6-disable waits for 50ms to check if we entered rc6. One some failures this seems to take longer to enter rc6, about 100ms. Tweak the test for DG2 to use a longer wait time.
--- tests/i915/perf.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
tests/i915/perf.c +6 -3
@@ -4272,7 +4272,10 @@ test_rc6_disable(void) .properties_ptr = to_user_pointer(properties), };unsigned long rc6_start, rc6_end, rc6_enabled; +uint32_t rc6_duration; +/* On DG2, we wait longer to enter rc6 */ + rc6_duration =IS_DG2 (devid) ?100000 :50000 ; rc6_enabled =0 ;igt_sysfs_rps_scanf (sysfs,RC6_ENABLE ,"%lu" , &rc6_enabled);igt_require (rc6_enabled);
@@ -4280,7 +4283,7 @@ test_rc6_disable(void)/* Verify rc6 is functional by measuring residency while idle */ gem_quiescent_gpu (drm_fd); rc6_start =rc6_residency_ms (); -usleep (50000 ); +usleep (rc6_duration); rc6_end =rc6_residency_ms ();igt_require (rc6_end != rc6_start);
@@ -4288,7 +4291,7 @@ test_rc6_disable(void) stream_fd =__perf_open (drm_fd, ¶m, false); rc6_start =rc6_residency_ms (); -usleep (50000 ); +usleep (rc6_duration); rc6_end =rc6_residency_ms ();igt_assert_eq (rc6_end - rc6_start,0 );
@@ -4297,7 +4300,7 @@ test_rc6_disable(void)/* But once OA is closed, we expect the device to sleep again */ rc6_start =rc6_residency_ms (); -usleep (50000 ); +usleep (rc6_duration); rc6_end =rc6_residency_ms ();igt_assert_neq (rc6_end - rc6_start,0 ); }
Signature · 1 line
2.25.1
PATCH v6 14/36 lib/i915/perf: implement report accumulation for new format
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
Available on XEHPSDV+.
--- lib/i915/perf.c | 33 +++++++++++++++++++++++++++++++++ 1 file changed, 33 insertions(+)
lib/i915/perf.c +33 -0
@@ -676,6 +676,39 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,memset (acc,0 ,sizeof (*acc));switch (oa_format) { +case I915_OA_FORMAT_A24u40_A14u32_B8_C8: +accumulate_uint32 (start +1 , end +1 , deltas + idx++);/* timestamp */ +accumulate_uint32 (start +3 , end +3 , deltas + idx++);/* clock */ + +/* 4x 32bit A0-3 counters... */ +for (i =0 ; i <4 ; i++) +accumulate_uint32 (start +4 + i, end +4 + i, deltas + idx++); + +/* 20x 40bit A4-23 counters... */ +for (i =0 ; i <20 ; i++) +accumulate_uint40 (i +4 , start, end, deltas + idx++); + +/* 4x 32bit A24-27 counters... */ +for (i =0 ; i <4 ; i++) +accumulate_uint32 (start +28 + i, end +28 + i, deltas + idx++); + +/* 4x 40bit A28-31 counters... */ +for (i =0 ; i <4 ; i++) +accumulate_uint40 (i +28 , start, end, deltas + idx++); + +/* 5x 32bit A32-36 counters... */ +for (i =0 ; i <5 ; i++) +accumulate_uint32 (start +36 + i, end +36 + i, deltas + idx++); + +/* 1x 32bit A37 counter... */ +accumulate_uint32 (start +46 , end +46 , deltas + idx++); + +/* 8x 32bit B counters + 8x 32bit C counters... */ +for (i =0 ; i <16 ; i++) +accumulate_uint32 (start +48 + i, end +48 + i, deltas + idx++); +break ; + +case I915_OAR_FORMAT_A32u40_A4u32_B8_C8:case I915_OA_FORMAT_A32u40_A4u32_B8_C8:accumulate_uint32 (start +1 , end +1 , deltas + idx++);/* timestamp */ accumulate_uint32 (start +3 , end +3 , deltas + idx++);/* clock */
Signature · 1 line
2.25.1
PATCH v6 15/36 lib/i915/perf: fixup conversion script for XEHPSDV
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
--- lib/i915/perf-configs/mdapi-xml-convert.py | 56 ++++++++++++++++++---- 1 file changed, 48 insertions(+), 8 deletions(-)
lib/i915/perf-configs/mdapi-xml-convert.py +48 -8
@@ -62,6 +62,24 @@ gen8_11_chipset_params = { 'config_reg_blacklist': { 0x2364, # OACTXID }, + 'register_offsets': { + 0x1f0: 'PERFCNT 0', + 0x1f8: 'PERFCNT 1', + } +} + +xehpsdv_chipset_params = { + 'a_offset': 16, + 'b_offset': 192, + 'c_offset': 224, + 'oa_report_size': 256, + 'config_reg_blacklist': { + 0x2364, # OACTXID + }, + 'register_offsets': { + 0x1b0: 'PERFCNT 0', + 0x1b8: 'PERFCNT 1', + } } chipsets = {
@@ -161,14 +179,9 @@ def underscore(name): def print_err(*args): sys.stderr.write(' '.join(map(str,args)) + '\n') -read_register_offsets = { - 0x1f0: 'PERFCNT 0', - 0x1f8: 'PERFCNT 1', -} - def read_value(chipset, offset): - if offset in read_register_offsets: - return read_register_offsets[offset] + if offset in chipsets[chipset]['register_offsets']: + return chipsets[chipset]['register_offsets'][offset] print_err("Unknown offset register at offset {0}".format(offset)) assert 0
@@ -185,6 +198,7 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): offset = int(offset_str, 16) if raw_offsets: + # Location in the HW reports a_offset = chipsets[chipset]['a_offset'] b_offset = chipsets[chipset]['b_offset'] c_offset = chipsets[chipset]['c_offset']
@@ -199,7 +213,15 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): else: assert 0 elif offset < b_offset: - return "A " + str(int((offset - a_offset) / 4)) + " READ" + a_cnt_offset = int((offset - a_offset) / 4) + if chipset == "XEHPSDV": + # Most A counters are in a contiguous array, except + # this A37. + if a_cnt_offset == 42: + return "A 37 READ" + return "A " + str(a_cnt_offset) + " READ" + else: + return "A " + str(a_cnt_offset) + " READ" elif offset < c_offset: return "B " + str(int((offset - b_offset) / 4)) + " READ" elif offset < report_size:
@@ -207,6 +229,7 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): else: return "{0} READ".format(read_value(chipset, offset)) else: + # Location in the accumulated deltas idx = int(offset / 8) if chipset == "HSW": # On Haswell accumulated counters are assumed to start
@@ -222,6 +245,23 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): return "C " + str(idx - 54) + " READ" else: return "{0} READ".format(read_value(chipset, offset)) + elif chipset == "XEHPSDV": + # For XEHPSDV the array of accumulated counters is + # assumed to start with a GPU_TIME then GPU_CLOCK, + # then 38 A counters, then 8 B counters and finally + # 8 C counters. + if idx == 0: + return "GPU_TIME 0 READ" + elif idx == 1: + return "GPU_CLOCK 0 READ" + elif idx < 40: + return "A " + str(idx - 2) + " READ" + elif idx < 48: + return "B " + str(idx - 40) + " READ" + elif idx < 56: + return "C " + str(idx - 48) + " READ" + else: + return "{0} READ".format(read_value(chipset, offset)) else: # For Gen8+ the array of accumulated counters is # assumed to start with a GPU_TIME then GPU_CLOCK,
Signature · 1 line
2.25.1
PATCH v6 16/36 lib/i915/perf: make warning message more helpful
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
This helped debugging some XEHPSDV stuff but is also more readable.
--- lib/i915/perf-configs/mdapi-xml-convert.py | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-)
lib/i915/perf-configs/mdapi-xml-convert.py +9 -4
@@ -884,10 +884,15 @@ for arg in args.xml: # delta equations are expected to be identical so warn if that's # not true... if bool(raw_read_eq) ^ bool(delta_read_eq) or raw_read_eq != delta_read_eq: - print_err("WARNING: Inconsistent raw and delta report equations for " + \ - mdapi_set.get('ShortName') + " :: " + mdapi_counter.get('SymbolName') + \ - "(" + mdapi_counter.get('ShortName') + ")" + ": raw=\"" + str(raw_read_eq) + \ - "\" delta=\"" + str(delta_read_eq) + "\" (SKIPPING)") + print_err(("WARNING: Inconsistent raw and delta report equations for {0} :: {1} ({2}): " + + "raw=\"{3}\" / \"{4}\" delta=\"{5}\" / \"{6}\" (SKIPPING)") + .format(mdapi_set.get('ShortName'), + mdapi_counter.get('SymbolName'), + mdapi_counter.get('ShortName'), + str(raw_read_eq), + mdapi_counter.get('SnapshotReportReadEquation'), + str(delta_read_eq), + mdapi_counter.get('DeltaReportReadEquation'))) set.remove(counter) continue
Signature · 1 line
2.25.1
PATCH v6 17/36 lib/i915/perf: expose new operators for codegen
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
Newer platform started to use operators we didn't implement.
--- lib/i915/perf-configs/codegen.py | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+)
lib/i915/perf-configs/codegen.py +28 -0
@@ -150,11 +150,17 @@ class Gen: self.ops["<<"] = (2, self.emit_lshft) self.ops[">>"] = (2, self.emit_rshft) self.ops["AND"] = (2, self.emit_and) + self.ops["UGTE"] = (2, self.emit_ugte) + self.ops["UGT"] = (2, self.emit_ugt) + self.ops["ULTE"] = (2, self.emit_ulte) + self.ops["ULT"] = (2, self.emit_ult) self.exp_ops = {} # (n operands, splicer) self.exp_ops["AND"] = (2, self.splice_bitwise_and) self.exp_ops["UGTE"] = (2, self.splice_ugte) + self.exp_ops["UGT"] = (2, self.splice_ugt) + self.exp_ops["ULTE"] = (2, self.splice_ulte) self.exp_ops["ULT"] = (2, self.splice_ult) self.exp_ops["&&"] = (2, self.splice_logical_and)
@@ -238,6 +244,22 @@ class Gen: self.c("uint64_t tmp{0} = {1} & {2};".format(tmp_id, args[1], args[0])) return tmp_id + 1 + def emit_ulte(self, tmp_id, args): + self.c("uint64_t tmp{0} = {1} <= {2};".format(tmp_id, args[1], args[0])) + return tmp_id + 1 + + def emit_ult(self, tmp_id, args): + self.c("uint64_t tmp{0} = {1} < {2};".format(tmp_id, args[1], args[0])) + return tmp_id + 1 + + def emit_ugte(self, tmp_id, args): + self.c("uint64_t tmp{0} = {1} >= {2};".format(tmp_id, args[1], args[0])) + return tmp_id + 1 + + def emit_ugt(self, tmp_id, args): + self.c("uint64_t tmp{0} = {1} > {2};".format(tmp_id, args[1], args[0])) + return tmp_id + 1 + def brkt(self, subexp): if " " in subexp: return "(" + subexp + ")"
@@ -250,12 +272,18 @@ class Gen: def splice_logical_and(self, args): return self.brkt(args[1]) + " && " + self.brkt(args[0]) + def splice_ulte(self, args): + return self.brkt(args[1]) + " <= " + self.brkt(args[0]) + def splice_ult(self, args): return self.brkt(args[1]) + " < " + self.brkt(args[0]) def splice_ugte(self, args): return self.brkt(args[1]) + " >= " + self.brkt(args[0]) + def splice_ugt(self, args): + return self.brkt(args[1]) + " > " + self.brkt(args[0]) + def resolve_variable(self, name, set): if name in self.hw_vars: return self.hw_vars[name]['c']
Signature · 1 line
2.25.1
PATCH v6 18/36 tools/i915-perf-recorder: add ability to select device
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
We're starting to have system with multiple Intel GPUs.
--- tools/i915-perf/i915_perf_recorder.c | 50 ++++++++++++++++++++++++---- 1 file changed, 43 insertions(+), 7 deletions(-)
tools/i915-perf/i915_perf_recorder.c +43 -7
@@ -250,16 +250,37 @@ find_intel_render_node(void)return -1 ; } +static void +print_intel_devices (void ) +{ +fprintf (stdout,"Available devices:\n" ); +for (int i =0 ; i <128 ; i++) { +if (read_device_param ("card" , i,"vendor" ) ==0x8086 ) { +uint32_t devid =read_device_param ("card" , i,"device" ); +const struct intel_device_info *devinfo = +intel_get_device_info (devid); +fprintf (stdout," %i: %s (0x%04hx)\n" , i, + devinfo ? devinfo->codename :"unknwon" , + devid); + } + } +} +static int -open_render_node (uint32_t *devid) +open_render_node (uint32_t *devid,int card) {char *name;int ret;int fd; +int render; -int render =find_intel_render_node (); -if (render <0 ) -return -1 ; +if (card <0 ) { + render =find_intel_render_node (); +if (render <0 ) +return -1 ; + }else { + render =128 + card; + } ret =asprintf (&name,"/dev/dri/renderD%u" , render);assert (ret !=-1 );
@@ -722,6 +743,9 @@ usage(const char *name)"Recording tool for i915-perf.\n" "\n" " --help, -h Print this screen\n" +" --device, -d <value> Device to use\n" +" (value=list to list devices\n" +" value=1 to use /dev/dri/card1)\n" " --correlation-period, -c <value> Time period of timestamp correlation in seconds\n" " (default = 1.0)\n" " --perf-period, -p <value> Time period of i915-perf reports in seconds\n"
@@ -772,6 +796,7 @@ main(int argc, char *argv[]) {const struct option long_options[] = { {"help" , no_argument,0 ,'h' }, + {"device" , required_argument,0 ,'d' }, {"correlation-period" , required_argument,0 ,'c' }, {"perf-period" , required_argument,0 ,'p' }, {"metric" , required_argument,0 ,'m' },
@@ -798,7 +823,7 @@ main(int argc, char *argv[])struct timespec now;uint64_t corr_period_ns, poll_time_ns;uint32_t circular_size =0 ; -int opt; +int opt, dev_node_id =-1 ;bool list_counters = false;FILE *output =NULL ;struct recording_context ctx = {
@@ -812,7 +837,7 @@ main(int argc, char *argv[]) .poll_period = 5 * 1000 * 1000, }; -while ((opt =getopt_long (argc, argv,"hc:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) { +while ((opt =getopt_long (argc, argv,"hc:d:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) {switch (opt) {case 'h' :usage (argv[0 ]);
@@ -820,6 +845,12 @@ main(int argc, char *argv[])case 'c' : corr_period =atof (optarg);break ; +case 'd' : +if (!strcmp (optarg,"list" )) + dev_node_id =-2 ; +else + dev_node_id =atoi (optarg); +break ;case 'p' : perf_period =atof (optarg);break ;
@@ -865,7 +896,12 @@ main(int argc, char *argv[]) } } - ctx.drm_fd =open_render_node (&ctx.devid); +if (dev_node_id ==-2 ) { +print_intel_devices (); +return EXIT_SUCCESS ; + } + + ctx.drm_fd =open_render_node (&ctx.devid, dev_node_id);if (ctx.drm_fd <0 ) {fprintf (stderr,"Unable to open device.\n" );return EXIT_FAILURE ;
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 18/36] tools/i915-perf-recorder: add ability to select device
Quoted 5 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
We're starting to have system with multiple Intel GPUs.
Signed-off-by Lionel Landwerlin <[email protected]>
nit: I would use a --list option with arguments like 'devices' or 'metrics', but this works too, so
Umesh
Quoted 116 lines
---
tools/i915-perf/i915_perf_recorder.c| 50 ++++++++++++++++++++++++----1 file changed, 43 insertions(+), 7 deletions(-)tools/i915-perf/i915_perf_recorder.c +16
-0@@ -250,16 +250,37 @@ find_intel_render_node(void)return -1 ; } +static void +print_intel_devices (void ) +{ +fprintf (stdout,"Available devices:\n" ); +for (int i =0 ; i <128 ; i++) { +if (read_device_param ("card" , i,"vendor" ) ==0x8086 ) { +uint32_t devid =read_device_param ("card" , i,"device" ); +const struct intel_device_info *devinfo = +intel_get_device_info (devid); +fprintf (stdout," %i: %s (0x%04hx)\n" , i, + devinfo ? devinfo->codename :"unknwon" , + devid); + } + } +} +static int
tools/i915-perf/i915_perf_recorder.c +1
-1-open_render_node (uint32_t *devid) +open_render_node(uint32_t *devid, int card){ char *name; int ret; int fd;tools/i915-perf/i915_perf_recorder.c +8
-3+int render; -int render =find_intel_render_node (); -if (render <0 ) -return -1 ; +if (card <0 ) { + render =find_intel_render_node (); +if (render <0 ) +return -1 ; + }else { + render =128 + card; + }ret = asprintf(&name, "/dev/dri/renderD%u", render); assert(ret != -1);tools/i915-perf/i915_perf_recorder.c +18
-3@@ -722,6 +743,9 @@ usage(const char *name)"Recording tool for i915-perf.\n" "\n" " --help, -h Print this screen\n" +" --device, -d <value> Device to use\n" +" (value=list to list devices\n" +" value=1 to use /dev/dri/card1)\n" " --correlation-period, -c <value> Time period of timestamp correlation in seconds\n" " (default = 1.0)\n" " --perf-period, -p <value> Time period of i915-perf reports in seconds\n" @@ -772,6 +796,7 @@ main(int argc, char *argv[]) {const struct option long_options[] = { {"help" , no_argument,0 ,'h' }, + {"device" , required_argument,0 ,'d' }, {"correlation-period" , required_argument,0 ,'c' }, {"perf-period" , required_argument,0 ,'p' }, {"metric" , required_argument,0 ,'m' },@@ -798,7 +823,7 @@ main(int argc, char *argv[])struct timespec now;uint64_t corr_period_ns, poll_time_ns;uint32_t circular_size =0 ; -int opt; +int opt, dev_node_id =-1 ;bool list_counters = false;FILE *output =NULL ;struct recording_context ctx = {@@ -812,7 +837,7 @@ main(int argc, char *argv[]) .poll_period = 5 * 1000 * 1000, }; -while ((opt =getopt_long (argc, argv,"hc:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) { +while ((opt =getopt_long (argc, argv,"hc:d:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) {switch (opt) {case 'h' :usage (argv[0 ]);@@ -820,6 +845,12 @@ main(int argc, char *argv[])case 'c' : corr_period =atof (optarg);break ; +case 'd' : +if (!strcmp (optarg,"list" )) + dev_node_id =-2 ; +else + dev_node_id =atoi (optarg); +break ;case 'p' : perf_period =atof (optarg);break ;@@ -865,7 +896,12 @@ main(int argc, char *argv[]) } } - ctx.drm_fd =open_render_node (&ctx.devid); +if (dev_node_id ==-2 ) { +print_intel_devices (); +return EXIT_SUCCESS ; + } + + ctx.drm_fd =open_render_node (&ctx.devid, dev_node_id);if (ctx.drm_fd <0 ) {fprintf (stderr,"Unable to open device.\n" );return EXIT_FAILURE ;Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 18/36] tools/i915-perf-recorder: add ability to select device
In reply to Umesh Nerlige RamappaQuoted 18 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
We're starting to have system with multiple Intel GPUs.
Signed-off-by Lionel Landwerlin <[email protected]>--- tools/i915-perf/i915_perf_recorder.c | 50 ++++++++++++++++++++++++---- 1 file changed, 43 insertions(+), 7 deletions(-)tools/i915-perf/i915_perf_recorder.c +1
-0@@ -250,16 +250,37 @@ find_intel_render_node(void)return -1 ; } +static void
tools/i915-perf/i915_perf_recorder.c find_intel_render_node(void) · line 263
Quoted 10 lines
tools/i915-perf/i915_perf_recorder.c +10
-0+print_intel_devices (void) +{ +fprintf (stdout,"Available devices:\n" ); +for (int i =0 ; i <128 ; i++) { +if (read_device_param ("card" , i,"vendor" ) ==0x8086 ) { +uint32_t devid =read_device_param ("card" , i,"device" ); +const struct intel_device_info *devinfo = +intel_get_device_info (devid); + fprintf(stdout, " %i: %s (0x%04hx)\n", i, + devinfo ? devinfo->codename : "unknwon",
Typo, "unknown".
Is there a reason why this doesn't use the same device filtering/selection machinery that other tools and the tests themselves are using?
Signature · 1 line
Petri Latvala
Quoted 93 lines
tools/i915-perf/i915_perf_recorder.c +32
-7+ devid); + } + } +} +static int -open_render_node (uint32_t *devid) +open_render_node (uint32_t *devid,int card) {char *name;int ret;int fd; +int render; -int render =find_intel_render_node (); -if (render <0 ) -return -1 ; +if (card <0 ) { + render =find_intel_render_node (); +if (render <0 ) +return -1 ; + }else { + render =128 + card; + } ret =asprintf (&name,"/dev/dri/renderD%u" , render);assert (ret !=-1 );@@ -722,6 +743,9 @@ usage(const char *name)"Recording tool for i915-perf.\n" "\n" " --help, -h Print this screen\n" +" --device, -d <value> Device to use\n" +" (value=list to list devices\n" +" value=1 to use /dev/dri/card1)\n" " --correlation-period, -c <value> Time period of timestamp correlation in seconds\n" " (default = 1.0)\n" " --perf-period, -p <value> Time period of i915-perf reports in seconds\n" @@ -772,6 +796,7 @@ main(int argc, char *argv[]) {const struct option long_options[] = { {"help" , no_argument,0 ,'h' }, + {"device" , required_argument,0 ,'d' }, {"correlation-period" , required_argument,0 ,'c' }, {"perf-period" , required_argument,0 ,'p' }, {"metric" , required_argument,0 ,'m' },@@ -798,7 +823,7 @@ main(int argc, char *argv[])struct timespec now;uint64_t corr_period_ns, poll_time_ns;uint32_t circular_size =0 ; -int opt; +int opt, dev_node_id =-1 ;bool list_counters = false;FILE *output =NULL ;struct recording_context ctx = {@@ -812,7 +837,7 @@ main(int argc, char *argv[]) .poll_period = 5 * 1000 * 1000, }; -while ((opt =getopt_long (argc, argv,"hc:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) { +while ((opt =getopt_long (argc, argv,"hc:d:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) {switch (opt) {case 'h' :usage (argv[0 ]);@@ -820,6 +845,12 @@ main(int argc, char *argv[])case 'c' : corr_period =atof (optarg);break ; +case 'd' : +if (!strcmp (optarg,"list" )) + dev_node_id =-2 ; +else + dev_node_id =atoi (optarg); +break ;case 'p' : perf_period =atof (optarg);break ;@@ -865,7 +896,12 @@ main(int argc, char *argv[]) } } - ctx.drm_fd =open_render_node (&ctx.devid); +if (dev_node_id ==-2 ) { +print_intel_devices (); +return EXIT_SUCCESS ; + } + + ctx.drm_fd =open_render_node (&ctx.devid, dev_node_id);if (ctx.drm_fd <0 ) {fprintf (stderr,"Unable to open device.\n" );return EXIT_FAILURE ;Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 18/36] tools/i915-perf-recorder: add ability to select device
Quoted 32 lines · Petri Latvala
Quoted 18 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
We're starting to have system with multiple Intel GPUs.
Signed-off-by Lionel Landwerlin <[email protected]>--- tools/i915-perf/i915_perf_recorder.c | 50 ++++++++++++++++++++++++---- 1 file changed, 43 insertions(+), 7 deletions(-)tools/i915-perf/i915_perf_recorder.c +1
-0@@ -250,16 +250,37 @@ find_intel_render_node(void)return -1 ; } +static voidtools/i915-perf/i915_perf_recorder.c find_intel_render_node(void) · line 263
Quoted 10 lines
tools/i915-perf/i915_perf_recorder.c +10
-0+print_intel_devices (void) +{ +fprintf (stdout,"Available devices:\n" ); +for (int i =0 ; i <128 ; i++) { +if (read_device_param ("card" , i,"vendor" ) ==0x8086 ) { +uint32_t devid =read_device_param ("card" , i,"device" ); +const struct intel_device_info *devinfo = +intel_get_device_info (devid); + fprintf(stdout, " %i: %s (0x%04hx)\n", i, + devinfo ? devinfo->codename : "unknwon",Typo, "unknown".
Quoted 3 lines
Is there a reason why this doesn't use the same device filtering/selection machinery that other tools and the tests themselves are using?
afair, that was pulling in a lot more code/library than needed for the perf tool and/or GPUvis, but I might be wrong as I have not tried it recently.
Regards,
Umesh
Quoted 100 lines
Signature · 2 lines
-- Petri LatvalaQuoted 93 lines
tools/i915-perf/i915_perf_recorder.c +32
-8+ devid); + } + } +} +static int -open_render_node (uint32_t *devid) +open_render_node (uint32_t *devid,int card) {char *name;int ret;int fd; +int render; -int render =find_intel_render_node (); -if (render <0 ) -return -1 ; +if (card <0 ) { + render =find_intel_render_node (); +if (render <0 ) +return -1 ; + }else { + render =128 + card; + } ret =asprintf (&name,"/dev/dri/renderD%u" , render);assert (ret !=-1 );@@ -722,6 +743,9 @@ usage(const char *name)"Recording tool for i915-perf.\n" "\n" " --help, -h Print this screen\n" +" --device, -d <value> Device to use\n" +" (value=list to list devices\n" +" value=1 to use /dev/dri/card1)\n" " --correlation-period, -c <value> Time period of timestamp correlation in seconds\n" " (default = 1.0)\n" " --perf-period, -p <value> Time period of i915-perf reports in seconds\n" @@ -772,6 +796,7 @@ main(int argc, char *argv[]) {const struct option long_options[] = { {"help" , no_argument,0 ,'h' }, + {"device" , required_argument,0 ,'d' }, {"correlation-period" , required_argument,0 ,'c' }, {"perf-period" , required_argument,0 ,'p' }, {"metric" , required_argument,0 ,'m' },@@ -798,7 +823,7 @@ main(int argc, char *argv[])struct timespec now;uint64_t corr_period_ns, poll_time_ns;uint32_t circular_size =0 ; -int opt; +int opt, dev_node_id =-1 ;bool list_counters = false;FILE *output =NULL ;struct recording_context ctx = {@@ -812,7 +837,7 @@ main(int argc, char *argv[]) .poll_period = 5 * 1000 * 1000, }; -while ((opt =getopt_long (argc, argv,"hc:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) { +while ((opt =getopt_long (argc, argv,"hc:d:p:m:Co:s:f:k:P:" , long_options,NULL )) !=-1 ) {switch (opt) {case 'h' :usage (argv[0 ]);@@ -820,6 +845,12 @@ main(int argc, char *argv[])case 'c' : corr_period =atof (optarg);break ; +case 'd' : +if (!strcmp (optarg,"list" )) + dev_node_id =-2 ; +else + dev_node_id =atoi (optarg); +break ;case 'p' : perf_period =atof (optarg);break ;@@ -865,7 +896,12 @@ main(int argc, char *argv[]) } } - ctx.drm_fd =open_render_node (&ctx.devid); +if (dev_node_id ==-2 ) { +print_intel_devices (); +return EXIT_SUCCESS ; + } + + ctx.drm_fd =open_render_node (&ctx.devid, dev_node_id);if (ctx.drm_fd <0 ) {fprintf (stderr,"Unable to open device.\n" );return EXIT_FAILURE ; --2.25.1
PATCH v6 19/36 lib/i915/perf: fixup report validity
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
It appears that on Gfx12+ there is no more validity bit in the reports.
--- lib/i915/perf_data_reader.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-)
lib/i915/perf_data_reader.c +5 -4
@@ -45,13 +45,14 @@ oa_report_ctx_is_valid(const struct intel_perf_devinfo *devinfo, {const uint32_t *report = (const uint32_t *) _report; -if (devinfo->graphics_ver <8 ) { +if (devinfo->graphics_ver <8 )return false;/* TODO */ - }else if (devinfo->graphics_ver ==8 ) { +if (devinfo->graphics_ver >=12 ) +return true;/* Always valid */ +if (devinfo->graphics_ver ==8 )return report[0 ] & (1ul <<25 ); - }else if (devinfo->graphics_ver >8 ) { +if (devinfo->graphics_ver >8 )return report[0 ] & (1ul <<16 ); - }return false; }
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 19/36] lib/i915/perf: fixup report validity
Quoted 6 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
It appears that on Gfx12+ there is no more validity bit in the reports.
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 30 lines
---
lib/i915/perf_data_reader.c| 9 +++++----1 file changed, 5 insertions(+), 4 deletions(-)lib/i915/perf_data_reader.c +5
-4@@ -45,13 +45,14 @@ oa_report_ctx_is_valid(const struct intel_perf_devinfo *devinfo, {const uint32_t *report = (const uint32_t *) _report; -if (devinfo->graphics_ver <8 ) { +if (devinfo->graphics_ver <8 )return false;/* TODO */ - }else if (devinfo->graphics_ver ==8 ) { +if (devinfo->graphics_ver >=12 ) +return true;/* Always valid */ +if (devinfo->graphics_ver ==8 )return report[0 ] & (1ul <<25 ); - }else if (devinfo->graphics_ver >8 ) { +if (devinfo->graphics_ver >8 )return report[0 ] & (1ul <<16 ); - }return false; }Signature · 1 line
2.25.1
PATCH v6 20/36 lib/i915/perf: add a helper to read timestamps
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
We expect location/size of the timestamp might move around in the report.
--- lib/i915/perf.c | 21 +++++++++++++++++++++ lib/i915/perf.h | 4 ++++ lib/i915/perf_data_reader.c | 14 ++++++-------- lib/meson.build | 2 +- tools/i915-perf/i915_perf_reader.c | 20 ++++++++++++++++++++ 5 files changed, 52 insertions(+), 9 deletions(-)
lib/i915/perf.c +21 -0
@@ -737,3 +737,24 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc, } } + +uint64_t intel_perf_read_record_timestamp (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record) +{ +const uint32_t *report32 = (const uint32_t *)(record +1 ); +uint64_t ts; + +switch (metric_set->perf_oa_format) { +case I915_OA_FORMAT_A24u40_A14u32_B8_C8: +case I915_OA_FORMAT_A32u40_A4u32_B8_C8: +case I915_OA_FORMAT_A45_B8_C8 : + ts = report32[1 ]; +break ; + +default : +assert (0 ); + } + +return ts; +}
lib/i915/perf.h +4 -0
@@ -240,6 +240,10 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,const struct drm_i915_perf_record_header *record0, const struct drm_i915_perf_record_header *record1); +uint64_t intel_perf_read_record_timestamp (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record); + #ifdef __cplusplus };#endif
lib/i915/perf_data_reader.c +6 -8
@@ -65,12 +65,6 @@ oa_report_ctx_id(const struct intel_perf_devinfo *devinfo, const uint8_t *reportreturn ((const uint32_t *) report)[2 ]; } -static inline uint64_t -oa_report_timestamp (const uint8_t *report) -{ -return ((const uint32_t *)report)[1 ]; -} -static void append_record (struct intel_perf_data_reader *reader,const struct drm_i915_perf_record_header *header)
@@ -298,8 +292,12 @@ generate_cpu_events(struct intel_perf_data_reader *reader) last_ctx_id =oa_report_ctx_id (&reader->devinfo, start_report); current_ctx_id =oa_report_ctx_id (&reader->devinfo, end_report); - gpu_ts_start =oa_report_timestamp (start_report); - gpu_ts_end =oa_report_timestamp (end_report); + gpu_ts_start =intel_perf_read_record_timestamp (reader->perf, + reader->metric_set, + last_header); + gpu_ts_end =intel_perf_read_record_timestamp (reader->perf, + reader->metric_set, + current_header);if (last_ctx_id == current_ctx_id)continue ;
lib/meson.build +1 -1
@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.2.0') +pkgconf.set('i915_perf_version', '1.3.0') configure_file( input : 'i915-perf.pc.in',
tools/i915-perf/i915_perf_reader.c +20 -0
@@ -230,6 +230,26 @@ main(int argc, char *argv[])fprintf (stdout,"Context switches: %u\n" , reader.n_timelines);fprintf (stdout,"Timestamp correlation points: %u\n" , reader.n_correlations); +if (reader.n_correlations <2 ) { +fprintf (stderr,"Less than 2 CPU/GPU timestamp correlation points.\n" ); +return EXIT_FAILURE ; + } + +fprintf (stdout,"Timestamp correlation CPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , + reader.correlations[0 ]->cpu_timestamp, + reader.correlations[reader.n_correlations -1 ]->cpu_timestamp); +fprintf (stdout,"Timestamp correlation GPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , + reader.correlations[0 ]->gpu_timestamp, + reader.correlations[reader.n_correlations -1 ]->gpu_timestamp); + +fprintf (stdout,"OA data timestamp range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +intel_perf_read_record_timestamp (reader.perf, + reader.metric_set, + reader.records[0 ]), +intel_perf_read_record_timestamp (reader.perf, + reader.metric_set, + reader.records[reader.n_records -1 ])); +if (strcmp (reader.metric_set_uuid, reader.metric_set->hw_config_guid)) {fprintf (stdout, "WARNING: Recording used a different HW configuration.\n"
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 20/36] lib/i915/perf: add a helper to read timestamps
Quoted 6 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
We expect location/size of the timestamp might move around in the report.
Signed-off-by Lionel Landwerlin <[email protected]>
Eventually I would think the i915 query (not yet upstreamed) will be used for correlation, but this still works for now.
Umesh
Quoted 131 lines
---
lib/i915/perf.c | 21 +++++++++++++++++++++ lib/i915/perf.h | 4 ++++ lib/i915/perf_data_reader.c | 14 ++++++-------- lib/meson.build | 2 +- tools/i915-perf/i915_perf_reader.c | 20 ++++++++++++++++++++ 5 files changed, 52 insertions(+), 9 deletions(-)lib/i915/perf.c +21
-0@@ -737,3 +737,24 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc, } } + +uint64_t intel_perf_read_record_timestamp (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record) +{ +const uint32_t *report32 = (const uint32_t *)(record +1 ); +uint64_t ts; + +switch (metric_set->perf_oa_format) { +case I915_OA_FORMAT_A24u40_A14u32_B8_C8: +case I915_OA_FORMAT_A32u40_A4u32_B8_C8: +case I915_OA_FORMAT_A45_B8_C8 : + ts = report32[1 ]; +break ; + +default : +assert (0 ); + } + +return ts; +}lib/i915/perf.h +4
-0@@ -240,6 +240,10 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,const struct drm_i915_perf_record_header *record0, const struct drm_i915_perf_record_header *record1); +uint64_t intel_perf_read_record_timestamp (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record); + #ifdef __cplusplus };#endif lib/i915/perf_data_reader.c +0
-6@@ -65,12 +65,6 @@ oa_report_ctx_id(const struct intel_perf_devinfo *devinfo, const uint8_t *reportreturn ((const uint32_t *) report)[2 ]; } -static inline uint64_t -oa_report_timestamp (const uint8_t *report) -{ -return ((const uint32_t *)report)[1 ]; -} -static void
append_record(structintel_perf_data_reader*reader,
const structdrm_i915_perf_record_header*header)lib/i915/perf_data_reader.c +6
-2@@ -298,8 +292,12 @@ generate_cpu_events(struct intel_perf_data_reader *reader) last_ctx_id =oa_report_ctx_id (&reader->devinfo, start_report); current_ctx_id =oa_report_ctx_id (&reader->devinfo, end_report); - gpu_ts_start =oa_report_timestamp (start_report); - gpu_ts_end =oa_report_timestamp (end_report); + gpu_ts_start =intel_perf_read_record_timestamp (reader->perf, + reader->metric_set, + last_header); + gpu_ts_end =intel_perf_read_record_timestamp (reader->perf, + reader->metric_set, + current_header);if (last_ctx_id == current_ctx_id)continue ;lib/meson.build +1
-1@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.2.0') +pkgconf.set('i915_perf_version', '1.3.0')configure_file(
input : 'i915-perf.pc.in',tools/i915-perf/i915_perf_reader.c +20
-0@@ -230,6 +230,26 @@ main(int argc, char *argv[])fprintf (stdout,"Context switches: %u\n" , reader.n_timelines);fprintf (stdout,"Timestamp correlation points: %u\n" , reader.n_correlations); +if (reader.n_correlations <2 ) { +fprintf (stderr,"Less than 2 CPU/GPU timestamp correlation points.\n" ); +return EXIT_FAILURE ; + } + +fprintf (stdout,"Timestamp correlation CPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , + reader.correlations[0 ]->cpu_timestamp, + reader.correlations[reader.n_correlations -1 ]->cpu_timestamp); +fprintf (stdout,"Timestamp correlation GPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , + reader.correlations[0 ]->gpu_timestamp, + reader.correlations[reader.n_correlations -1 ]->gpu_timestamp); + +fprintf (stdout,"OA data timestamp range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +intel_perf_read_record_timestamp (reader.perf, + reader.metric_set, + reader.records[0 ]), +intel_perf_read_record_timestamp (reader.perf, + reader.metric_set, + reader.records[reader.n_records -1 ])); +if (strcmp (reader.metric_set_uuid, reader.metric_set->hw_config_guid)) {fprintf (stdout, "WARNING: Recording used a different HW configuration.\n"Signature · 1 line
2.25.1
PATCH v6 21/36 lib/i915/perf: store bit shifting required for OA timestamps
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
Some new platforms have the OA timestamps shifted to the right by a number of bits. This is configurable but since i915 appears to leave the default values, let's just hardcode this in the device info.
--- lib/i915/perf.c | 31 +++++++++++++++++++++++++----- lib/i915/perf.h | 18 ++++++++++++++++- lib/i915/perf_data_reader.c | 2 +- lib/meson.build | 4 ++-- tools/i915-perf/i915_perf_reader.c | 3 ++- 5 files changed, 48 insertions(+), 10 deletions(-)
lib/i915/perf.c +26 -5
@@ -208,6 +208,10 @@ intel_perf_for_devinfo(uint32_t device_id,/* Valid on most generations except Gen9LP. */ perf->devinfo.eu_threads_count =7 ; +/* Most platforms have full 32bit timestamps. */ + perf->devinfo.oa_timestamp_mask =0xffffffff ; + perf->devinfo.oa_timestamp_shift =0 ; +if (devinfo->is_haswell) {intel_perf_load_metrics_hsw (perf); }else if (devinfo->is_broadwell) {
@@ -663,7 +667,8 @@ accumulate_uint40(int a_index, }void intel_perf_accumulate_reports (struct intel_perf_accumulator *acc, -int oa_format, +const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set,const struct drm_i915_perf_record_header *record0,const struct drm_i915_perf_record_header *record1) {
@@ -675,9 +680,13 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,memset (acc,0 ,sizeof (*acc)); -switch (oa_format) { +switch (metric_set->perf_oa_format) {case I915_OA_FORMAT_A24u40_A14u32_B8_C8: -accumulate_uint32 (start +1 , end +1 , deltas + idx++);/* timestamp */ +/* timestamp */ +if (perf->devinfo.oa_timestamp_shift >=0 ) + deltas[idx++] += (end[1 ] - start[1 ]) << perf->devinfo.oa_timestamp_shift; +else + deltas[idx++] += (end[1 ] - start[1 ]) >> (-perf->devinfo.oa_timestamp_shift);accumulate_uint32 (start +3 , end +3 , deltas + idx++);/* clock */ /* 4x 32bit A0-3 counters... */
@@ -710,7 +719,10 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,case I915_OAR_FORMAT_A32u40_A4u32_B8_C8:case I915_OA_FORMAT_A32u40_A4u32_B8_C8: -accumulate_uint32 (start +1 , end +1 , deltas + idx++);/* timestamp */ +if (perf->devinfo.oa_timestamp_shift >=0 ) + deltas[idx++] += (end[1 ] - start[1 ]) << perf->devinfo.oa_timestamp_shift; +else + deltas[idx++] += (end[1 ] - start[1 ]) >> (-perf->devinfo.oa_timestamp_shift);accumulate_uint32 (start +3 , end +3 , deltas + idx++);/* clock */ /* 32x 40bit A counters... */
@@ -727,7 +739,11 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,break ;case I915_OA_FORMAT_A45_B8_C8 : -accumulate_uint32 (start +1 , end +1 , deltas);/* timestamp */ +/* timestamp */ +if (perf->devinfo.oa_timestamp_shift >=0 ) + deltas[0 ] += (end[1 ] - start[1 ]) << perf->devinfo.oa_timestamp_shift; +else + deltas[0 ] += (end[1 ] - start[1 ]) >> (-perf->devinfo.oa_timestamp_shift);for (i =0 ; i <61 ; i++)accumulate_uint32 (start +3 + i, end +3 + i, deltas +1 + i);
@@ -756,5 +772,10 @@ uint64_t intel_perf_read_record_timestamp(const struct intel_perf *perf,assert (0 ); } +if (perf->devinfo.oa_timestamp_shift >=0 ) + ts <<= perf->devinfo.oa_timestamp_shift; +else + ts >>= -perf->devinfo.oa_timestamp_shift; +return ts; }
lib/i915/perf.h +17 -1
@@ -52,6 +52,21 @@ struct intel_perf_devinfo {uint32_t devid;uint32_t graphics_ver;uint32_t revision; +/** +* Bit shifting required to put OA report timestamps into +* timestamp_frequency (some HW generations can shift +* timestamp values to the right by a number of bits). +*/ +int32_t oa_timestamp_shift; +/** +* On some platforms only part of the timestamp bits are valid +* (on previous platforms we would get full 32bits, newer +* platforms can have fewer). It's important to know when +* correlating the full 36bits timestamps to the OA report +* timestamps. +*/ +uint64_t oa_timestamp_mask; +/* Frequency of the timestamps in Hz */ uint64_t timestamp_frequency;uint64_t gt_min_freq;uint64_t gt_max_freq;
@@ -236,7 +251,8 @@ void intel_perf_add_metric_set(struct intel_perf *perf,void intel_perf_load_perf_configs (struct intel_perf *perf,int drm_fd);void intel_perf_accumulate_reports (struct intel_perf_accumulator *acc, -int oa_format, +const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set,const struct drm_i915_perf_record_header *record0,const struct drm_i915_perf_record_header *record1);
lib/i915/perf_data_reader.c +1 -1
@@ -209,7 +209,7 @@ correlate_gpu_timestamp(struct intel_perf_data_reader *reader,* Try to figure what portion of the correlation data the * 32bit timestamp belongs to. */ -uint64_t mask =0xffffffff ; +uint64_t mask = reader->perf->devinfo.oa_timestamp_mask;int corr_idx =-1 ;for (uint32_t i =0 ; i < reader->n_correlation_chunks; i++) {
lib/meson.build +2 -2
@@ -308,7 +308,7 @@ lib_igt_i915_perf_build = shared_library( dependencies: lib_igt_chipset, include_directories : inc, install: true, - soversion: '1') + soversion: '1.4') lib_igt_i915_perf = declare_dependency( link_with : lib_igt_i915_perf_build,
@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.3.0') +pkgconf.set('i915_perf_version', '1.4.0') configure_file( input : 'i915-perf.pc.in',
tools/i915-perf/i915_perf_reader.c +2 -1
@@ -271,7 +271,8 @@ main(int argc, char *argv[])fprintf (stdout,"hw_id=0x%x %s\n" , item->hw_id, item->hw_id ==0xffffffff ?"(idle)" :"" ); -intel_perf_accumulate_reports (&accu, reader.metric_set->perf_oa_format, +intel_perf_accumulate_reports (&accu, + reader.perf, reader.metric_set, i915_report0, i915_report1);for (uint32_t c =0 ; c < n_counters; c++) {
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 21/36] lib/i915/perf: store bit shifting required for OA timestamps
some nits, but lgtm,
Umesh
Quoted 43 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Some new platforms have the OA timestamps shifted to the right by a number of bits. This is configurable but since i915 appears to leave the default values, let's just hardcode this in the device info.
Signed-off-by Lionel Landwerlin <[email protected]>---
lib/i915/perf.c | 31 +++++++++++++++++++++++++----- lib/i915/perf.h | 18 ++++++++++++++++- lib/i915/perf_data_reader.c | 2 +- lib/meson.build | 4 ++-- tools/i915-perf/i915_perf_reader.c | 3 ++- 5 files changed, 48 insertions(+), 10 deletions(-)lib/i915/perf.c +6
-1@@ -208,6 +208,10 @@ intel_perf_for_devinfo(uint32_t device_id,/* Valid on most generations except Gen9LP. */ perf->devinfo.eu_threads_count =7 ; +/* Most platforms have full 32bit timestamps. */ + perf->devinfo.oa_timestamp_mask =0xffffffff ; + perf->devinfo.oa_timestamp_shift =0 ; +if (devinfo->is_haswell) {intel_perf_load_metrics_hsw (perf); }else if (devinfo->is_broadwell) {@@ -663,7 +667,8 @@ accumulate_uint40(int a_index, }void intel_perf_accumulate_reports (struct intel_perf_accumulator *acc, -int oa_format, +const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set,const struct drm_i915_perf_record_header *record0,const struct drm_i915_perf_record_header *record1) {@@ -675,9 +680,13 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,memset (acc,0 ,sizeof (*acc));
lib/i915/perf.c void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc, · line 689
Quoted 10 lines
lib/i915/perf.c +6
-2-switch (oa_format) { +switch (metric_set->perf_oa_format) {case I915_OA_FORMAT_A24u40_A14u32_B8_C8: -accumulate_uint32 (start +1 , end +1 , deltas + idx++);/* timestamp */ +/* timestamp */ +if (perf->devinfo.oa_timestamp_shift >=0 ) + deltas[idx++] += (end[1 ] - start[1 ]) << perf->devinfo.oa_timestamp_shift; +else + deltas[idx++] += (end[1 ] - start[1 ]) >> (-perf->devinfo.oa_timestamp_shift);
I am guessing the negative numbers are used to shift right and positive for shift left.
lib/i915/perf.c void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc, · line 725
Quoted 12 lines
lib/i915/perf.c +4
-1accumulate_uint32 (start +3 , end +3 , deltas + idx++);/* clock */ /* 4x 32bit A0-3 counters... */ @@ -710,7 +719,10 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,case I915_OAR_FORMAT_A32u40_A4u32_B8_C8:case I915_OA_FORMAT_A32u40_A4u32_B8_C8: -accumulate_uint32 (start +1 , end +1 , deltas + idx++);/* timestamp */ +if (perf->devinfo.oa_timestamp_shift >=0 ) + deltas[idx++] += (end[1 ] - start[1 ]) << perf->devinfo.oa_timestamp_shift; +else + deltas[idx++] += (end[1 ] - start[1 ]) >> (-perf->devinfo.oa_timestamp_shift);
These could use a helper ^.
Quoted 115 lines
lib/i915/perf.c +10
-1accumulate_uint32 (start +3 , end +3 , deltas + idx++);/* clock */ /* 32x 40bit A counters... */ @@ -727,7 +739,11 @@ void intel_perf_accumulate_reports(struct intel_perf_accumulator *acc,break ;case I915_OA_FORMAT_A45_B8_C8 : -accumulate_uint32 (start +1 , end +1 , deltas);/* timestamp */ +/* timestamp */ +if (perf->devinfo.oa_timestamp_shift >=0 ) + deltas[0 ] += (end[1 ] - start[1 ]) << perf->devinfo.oa_timestamp_shift; +else + deltas[0 ] += (end[1 ] - start[1 ]) >> (-perf->devinfo.oa_timestamp_shift);for (i =0 ; i <61 ; i++)accumulate_uint32 (start +3 + i, end +3 + i, deltas +1 + i);@@ -756,5 +772,10 @@ uint64_t intel_perf_read_record_timestamp(const struct intel_perf *perf,assert (0 ); } +if (perf->devinfo.oa_timestamp_shift >=0 ) + ts <<= perf->devinfo.oa_timestamp_shift; +else + ts >>= -perf->devinfo.oa_timestamp_shift; +return ts; }lib/i915/perf.h +17
-1@@ -52,6 +52,21 @@ struct intel_perf_devinfo {uint32_t devid;uint32_t graphics_ver;uint32_t revision; +/** +* Bit shifting required to put OA report timestamps into +* timestamp_frequency (some HW generations can shift +* timestamp values to the right by a number of bits). +*/ +int32_t oa_timestamp_shift; +/** +* On some platforms only part of the timestamp bits are valid +* (on previous platforms we would get full 32bits, newer +* platforms can have fewer). It's important to know when +* correlating the full 36bits timestamps to the OA report +* timestamps. +*/ +uint64_t oa_timestamp_mask; +/* Frequency of the timestamps in Hz */ uint64_t timestamp_frequency;uint64_t gt_min_freq;uint64_t gt_max_freq;@@ -236,7 +251,8 @@ void intel_perf_add_metric_set(struct intel_perf *perf,void intel_perf_load_perf_configs (struct intel_perf *perf,int drm_fd);void intel_perf_accumulate_reports (struct intel_perf_accumulator *acc, -int oa_format, +const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set,const struct drm_i915_perf_record_header *record0,const struct drm_i915_perf_record_header *record1);lib/i915/perf_data_reader.c +1
-1@@ -209,7 +209,7 @@ correlate_gpu_timestamp(struct intel_perf_data_reader *reader,* Try to figure what portion of the correlation data the * 32bit timestamp belongs to. */ -uint64_t mask =0xffffffff ; +uint64_t mask = reader->perf->devinfo.oa_timestamp_mask;int corr_idx =-1 ;for (uint32_t i =0 ; i < reader->n_correlation_chunks; i++) {lib/meson.build +1
-1@@ -308,7 +308,7 @@ lib_igt_i915_perf_build = shared_library( dependencies: lib_igt_chipset, include_directories : inc, install: true, - soversion: '1') + soversion: '1.4')
lib_igt_i915_perf= declare_dependency(
link_with:lib_igt_i915_perf_build,lib/meson.build +1
-1@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.3.0') +pkgconf.set('i915_perf_version', '1.4.0')configure_file(
input : 'i915-perf.pc.in',tools/i915-perf/i915_perf_reader.c +2
-1@@ -271,7 +271,8 @@ main(int argc, char *argv[])fprintf (stdout,"hw_id=0x%x %s\n" , item->hw_id, item->hw_id ==0xffffffff ?"(idle)" :"" ); -intel_perf_accumulate_reports (&accu, reader.metric_set->perf_oa_format, +intel_perf_accumulate_reports (&accu, + reader.perf, reader.metric_set, i915_report0, i915_report1);for (uint32_t c =0 ; c < n_counters; c++) {Signature · 1 line
2.25.1
PATCH v6 22/36 lib/i915/perf: indentation fix
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
--- lib/i915/perf_data_reader.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
lib/i915/perf_data_reader.c +2 -2
@@ -337,8 +337,8 @@ bool intel_perf_data_reader_init(struct intel_perf_data_reader *reader, int perf_file_fd) { -struct stat st; -if (fstat (perf_file_fd, &st) !=0 ) { +struct stat st; +if (fstat (perf_file_fd, &st) !=0 ) {snprintf (reader->error_msg,sizeof (reader->error_msg),"Unable to access file (%s)" ,strerror (errno));return false;
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 22/36] lib/i915/perf: indentation fix
Quoted 25 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Signed-off-by Lionel Landwerlin <[email protected]>---
lib/i915/perf_data_reader.c| 4 ++--1 file changed, 2 insertions(+), 2 deletions(-)lib/i915/perf_data_reader.c +2
-2@@ -337,8 +337,8 @@ bool intel_perf_data_reader_init(struct intel_perf_data_reader *reader, int perf_file_fd) { -struct stat st; -if (fstat (perf_file_fd, &st) !=0 ) { +struct stat st; +if (fstat (perf_file_fd, &st) !=0 ) {snprintf (reader->error_msg,sizeof (reader->error_msg),"Unable to access file (%s)" ,strerror (errno));return false;Signature · 1 line
2.25.1
PATCH v6 23/36 tools/i915-perf-recorder: capture OA & CS frequencies
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
CS & OA timestamp don't have the same accuracy, so we need to adjuste the correlated timestamps (capture from the CS unit) in the recording to fit the data in the OA reports.
--- tools/i915-perf/i915_perf_recorder.c | 43 ++++++++++++++++++++++++---- 1 file changed, 38 insertions(+), 5 deletions(-)
tools/i915-perf/i915_perf_recorder.c +38 -5
@@ -328,7 +328,7 @@ perf_ioctl(int fd, unsigned long request, void *arg) }static uint64_t -get_device_timestamp_frequency (const struct intel_device_info *devinfo,int drm_fd) +get_device_cs_timestamp_frequency (const struct intel_device_info *devinfo,int drm_fd) {drm_i915_getparam_t gp;int timestamp_frequency;
@@ -352,12 +352,44 @@ get_device_timestamp_frequency(const struct intel_device_info *devinfo, int drm_return 12000000 ; } +static uint64_t +get_device_oa_timestamp_frequency (const struct intel_device_info *devinfo,int drm_fd) +{ +drm_i915_getparam_t gp; +int timestamp_frequency; + + gp.param =I915_PARAM_OA_TIMESTAMP_FREQUENCY ; + gp.value = ×tamp_frequency; +if (perf_ioctl (drm_fd,DRM_IOCTL_I915_GETPARAM , &gp) ==0 ) +return timestamp_frequency; + + gp.param =I915_PARAM_CS_TIMESTAMP_FREQUENCY ; + gp.value = ×tamp_frequency; +if (perf_ioctl (drm_fd,DRM_IOCTL_I915_GETPARAM , &gp) ==0 ) +return timestamp_frequency; + + +if (devinfo->graphics_ver >9 ) { +fprintf (stderr,"Unable to query timestamp frequency from i915, please update kernel.\n" ); +return 0 ; + } + +fprintf (stderr,"Warning: unable to query timestamp frequency from i915, guessing values...\n" ); + +if (devinfo->graphics_ver <=8 ) +return 12500000 ; +if (devinfo->is_broxton) +return 19200000 ; +return 12000000 ; +} +struct recording_context {int drm_fd;int perf_fd;uint32_t devid; -uint64_t timestamp_frequency; +uint64_t oa_timestamp_frequency; +uint64_t cs_timestamp_frequency;const struct intel_device_info *devinfo;
@@ -459,7 +491,7 @@ static bool write_header(FILE *output, struct recording_context *ctx) {struct intel_perf_record_device_info info = { - .timestamp_frequency = ctx->timestamp_frequency, + .timestamp_frequency = ctx->oa_timestamp_frequency, .device_id = ctx->perf->devinfo.devid, .device_revision = ctx->perf->devinfo.revision, .gt_min_frequency = ctx->perf->devinfo.gt_min_freq,
@@ -964,7 +996,8 @@ main(int argc, char *argv[])intel_perf_load_perf_configs (ctx.perf, ctx.drm_fd); - ctx.timestamp_frequency =get_device_timestamp_frequency (ctx.devinfo, ctx.drm_fd); + ctx.oa_timestamp_frequency =get_device_oa_timestamp_frequency (ctx.devinfo, ctx.drm_fd); + ctx.cs_timestamp_frequency =get_device_cs_timestamp_frequency (ctx.devinfo, ctx.drm_fd);signal (SIGINT , sigint_handler);
@@ -1036,7 +1069,7 @@ main(int argc, char *argv[]) goto fail; } - ctx.oa_exponent =oa_exponent_for_period (ctx.timestamp_frequency, perf_period); + ctx.oa_exponent =oa_exponent_for_period (ctx.oa_timestamp_frequency, perf_period);fprintf (stdout,"Opening perf stream with metric_id=%" PRIu64" oa_exponent=%u oa_format=%u\n" , ctx.metric_set->perf_oa_metrics_set, ctx.oa_exponent, ctx.metric_set->perf_oa_format);
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 23/36] tools/i915-perf-recorder: capture OA & CS frequencies
Quoted 7 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
CS & OA timestamp don't have the same accuracy, so we need to adjuste the correlated timestamps (capture from the CS unit) in the recording to fit the data in the OA reports.
Signed-off-by Lionel Landwerlin <[email protected]>
lgtm,
Quoted 94 lines
---
tools/i915-perf/i915_perf_recorder.c| 43 ++++++++++++++++++++++++----1 file changed, 38 insertions(+), 5 deletions(-)tools/i915-perf/i915_perf_recorder.c +0
-0@@ -328,7 +328,7 @@ perf_ioctl(int fd, unsigned long request, void *arg) }static
uint64_ttools/i915-perf/i915_perf_recorder.c +1
-1-get_device_timestamp_frequency(const struct intel_device_info *devinfo, int drm_fd) +get_device_cs_timestamp_frequency(const struct intel_device_info *devinfo, int drm_fd){ drm_i915_getparam_t gp; int timestamp_frequency;tools/i915-perf/i915_perf_recorder.c +37
-4@@ -352,12 +352,44 @@ get_device_timestamp_frequency(const struct intel_device_info *devinfo, int drm_return 12000000 ; } +static uint64_t +get_device_oa_timestamp_frequency (const struct intel_device_info *devinfo,int drm_fd) +{ +drm_i915_getparam_t gp; +int timestamp_frequency; + + gp.param =I915_PARAM_OA_TIMESTAMP_FREQUENCY ; + gp.value = ×tamp_frequency; +if (perf_ioctl (drm_fd,DRM_IOCTL_I915_GETPARAM , &gp) ==0 ) +return timestamp_frequency; + + gp.param =I915_PARAM_CS_TIMESTAMP_FREQUENCY ; + gp.value = ×tamp_frequency; +if (perf_ioctl (drm_fd,DRM_IOCTL_I915_GETPARAM , &gp) ==0 ) +return timestamp_frequency; + + +if (devinfo->graphics_ver >9 ) { +fprintf (stderr,"Unable to query timestamp frequency from i915, please update kernel.\n" ); +return 0 ; + } + +fprintf (stderr,"Warning: unable to query timestamp frequency from i915, guessing values...\n" ); + +if (devinfo->graphics_ver <=8 ) +return 12500000 ; +if (devinfo->is_broxton) +return 19200000 ; +return 12000000 ; +} +struct recording_context {int drm_fd;int perf_fd;uint32_t devid; -uint64_t timestamp_frequency; +uint64_t oa_timestamp_frequency; +uint64_t cs_timestamp_frequency;const struct intel_device_info *devinfo;@@ -459,7 +491,7 @@ static bool write_header(FILE *output, struct recording_context *ctx) {struct intel_perf_record_device_info info = { - .timestamp_frequency = ctx->timestamp_frequency, + .timestamp_frequency = ctx->oa_timestamp_frequency, .device_id = ctx->perf->devinfo.devid, .device_revision = ctx->perf->devinfo.revision, .gt_min_frequency = ctx->perf->devinfo.gt_min_freq,@@ -964,7 +996,8 @@ main(int argc, char *argv[])intel_perf_load_perf_configs (ctx.perf, ctx.drm_fd); - ctx.timestamp_frequency =get_device_timestamp_frequency (ctx.devinfo, ctx.drm_fd); + ctx.oa_timestamp_frequency =get_device_oa_timestamp_frequency (ctx.devinfo, ctx.drm_fd); + ctx.cs_timestamp_frequency =get_device_cs_timestamp_frequency (ctx.devinfo, ctx.drm_fd);signal (SIGINT , sigint_handler);@@ -1036,7 +1069,7 @@ main(int argc, char *argv[]) goto fail; } - ctx.oa_exponent =oa_exponent_for_period (ctx.timestamp_frequency, perf_period); + ctx.oa_exponent =oa_exponent_for_period (ctx.oa_timestamp_frequency, perf_period);fprintf (stdout,"Opening perf stream with metric_id=%" PRIu64" oa_exponent=%u oa_format=%u\n" , ctx.metric_set->perf_oa_metrics_set, ctx.oa_exponent, ctx.metric_set->perf_oa_format);Signature · 1 line
2.25.1
PATCH v6 24/36 tools/i915-perf: make timestamp range easier to compare
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
With timestamp bit shifting stuff, aligning values in the console is useful to compare them.
--- tools/i915-perf/i915_perf_reader.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
tools/i915-perf/i915_perf_reader.c +6 -3
@@ -235,14 +235,17 @@ main(int argc, char *argv[])return EXIT_FAILURE ; } -fprintf (stdout,"Timestamp correlation CPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +fprintf (stdout,"Timestamp correlation CPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , reader.correlations[0 ]->cpu_timestamp, reader.correlations[reader.n_correlations -1 ]->cpu_timestamp); -fprintf (stdout,"Timestamp correlation GPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +fprintf (stdout,"Timestamp correlation GPU range (64b): 0x%016" PRIx64"-0x%016" PRIx64"\n" , reader.correlations[0 ]->gpu_timestamp, reader.correlations[reader.n_correlations -1 ]->gpu_timestamp); +fprintf (stdout,"Timestamp correlation GPU range (32b): 0x%016" PRIx64"-0x%016" PRIx64"\n" , + reader.correlations[0 ]->gpu_timestamp &0xffffffff , + reader.correlations[reader.n_correlations -1 ]->gpu_timestamp &0xffffffff ); - fprintf(stdout, "OA data timestamp range: 0x%016"PRIx64"-0x%016"PRIx64"\n", + fprintf(stdout, "OA data timestamp range: 0x%016"PRIx64"-0x%016"PRIx64"\n", intel_perf_read_record_timestamp(reader.perf, reader.metric_set, reader.records[0]),
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 24/36] tools/i915-perf: make timestamp range easier to compare
Quoted 6 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
With timestamp bit shifting stuff, aligning values in the console is useful to compare them.
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 32 lines
---
tools/i915-perf/i915_perf_reader.c| 9 ++++++---1 file changed, 6 insertions(+), 3 deletions(-)tools/i915-perf/i915_perf_reader.c +6
-3@@ -235,14 +235,17 @@ main(int argc, char *argv[])return EXIT_FAILURE ; } -fprintf (stdout,"Timestamp correlation CPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +fprintf (stdout,"Timestamp correlation CPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , reader.correlations[0 ]->cpu_timestamp, reader.correlations[reader.n_correlations -1 ]->cpu_timestamp); -fprintf (stdout,"Timestamp correlation GPU range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +fprintf (stdout,"Timestamp correlation GPU range (64b): 0x%016" PRIx64"-0x%016" PRIx64"\n" , reader.correlations[0 ]->gpu_timestamp, reader.correlations[reader.n_correlations -1 ]->gpu_timestamp); +fprintf (stdout,"Timestamp correlation GPU range (32b): 0x%016" PRIx64"-0x%016" PRIx64"\n" , + reader.correlations[0 ]->gpu_timestamp &0xffffffff , + reader.correlations[reader.n_correlations -1 ]->gpu_timestamp &0xffffffff ); - fprintf(stdout, "OA data timestamp range: 0x%016"PRIx64"-0x%016"PRIx64"\n", + fprintf(stdout, "OA data timestamp range: 0x%016"PRIx64"-0x%016"PRIx64"\n", intel_perf_read_record_timestamp(reader.perf, reader.metric_set, reader.records[0]),Signature · 1 line
2.25.1
PATCH v6 25/36 tools/i915-perf: printout CPU clock used
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
--- tools/i915-perf/i915_perf_recorder.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
tools/i915-perf/i915_perf_recorder.c +14 -0
@@ -588,6 +588,17 @@ static uint64_t timespec_diff(struct timespec *begin,static clock_t correlation_clock_id =CLOCK_MONOTONIC ; +static const char * +get_correlation_clock_name (clock_t clock_id) +{ +switch (clock_id) { +case CLOCK_BOOTTIME :return "bootime" ; +case CLOCK_MONOTONIC :return "monotonic" ; +case CLOCK_MONOTONIC_RAW :return "monotonic_raw" ; +default :return "*unknown*" ; + } +} +static bool get_correlation_timestamps (struct intel_perf_record_timestamp_correlation *corr,int drm_fd) {
@@ -1069,6 +1080,9 @@ main(int argc, char *argv[]) goto fail; } +fprintf (stdout,"Using correlation clock: %s\n" , +get_correlation_clock_name (correlation_clock_id)); + ctx.oa_exponent =oa_exponent_for_period (ctx.oa_timestamp_frequency, perf_period); fprintf(stdout, "Opening perf stream with metric_id=%"PRIu64" oa_exponent=%u oa_format=%u\n", ctx.metric_set->perf_oa_metrics_set, ctx.oa_exponent,
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 25/36] tools/i915-perf: printout CPU clock used
Quoted 3 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 39 lines
---
tools/i915-perf/i915_perf_recorder.c| 14 ++++++++++++++1 file changed, 14 insertions(+)tools/i915-perf/i915_perf_recorder.c +11
-0@@ -588,6 +588,17 @@ static uint64_t timespec_diff(struct timespec *begin,static clock_t correlation_clock_id =CLOCK_MONOTONIC ; +static const char * +get_correlation_clock_name (clock_t clock_id) +{ +switch (clock_id) { +case CLOCK_BOOTTIME :return "bootime" ; +case CLOCK_MONOTONIC :return "monotonic" ; +case CLOCK_MONOTONIC_RAW :return "monotonic_raw" ; +default :return "*unknown*" ; + } +} +static bool
get_correlation_timestamps(struct intel_perf_record_timestamp_correlation *corr, int drm_fd){tools/i915-perf/i915_perf_recorder.c +3
-0@@ -1069,6 +1080,9 @@ main(int argc, char *argv[]) goto fail; } +fprintf (stdout,"Using correlation clock: %s\n" , +get_correlation_clock_name (correlation_clock_id)); + ctx.oa_exponent =oa_exponent_for_period (ctx.oa_timestamp_frequency, perf_period); fprintf(stdout, "Opening perf stream with metric_id=%"PRIu64" oa_exponent=%u oa_format=%u\n", ctx.metric_set->perf_oa_metrics_set, ctx.oa_exponent,Signature · 1 line
2.25.1
PATCH v6 26/36 tools/i915-perf: record remaining perf data on exit
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
--- tools/i915-perf/i915_perf_recorder.c | 5 +++++ 1 file changed, 5 insertions(+)
tools/i915-perf/i915_perf_recorder.c +5 -0
@@ -1144,6 +1144,11 @@ main(int argc, char *argv[])fprintf (stdout,"Exiting...\n" ); +if (!write_i915_perf_data (ctx.output_stream, ctx.perf_fd)) { +fprintf (stderr,"Failed to write i915-perf data: %s\n" , +strerror (errno)); + } +if (!write_correlation_timestamps (ctx.output_stream, ctx.drm_fd)) { fprintf(stderr, "Failed to write final i915 timestamp correlation data: %s\n",
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 26/36] tools/i915-perf: record remaining perf data on exit
Quoted 3 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 23 lines
---
tools/i915-perf/i915_perf_recorder.c| 5 +++++1 file changed, 5 insertions(+)tools/i915-perf/i915_perf_recorder.c +5
-0@@ -1144,6 +1144,11 @@ main(int argc, char *argv[])fprintf (stdout,"Exiting...\n" ); +if (!write_i915_perf_data (ctx.output_stream, ctx.perf_fd)) { +fprintf (stderr,"Failed to write i915-perf data: %s\n" , +strerror (errno)); + } +if (!write_correlation_timestamps (ctx.output_stream, ctx.drm_fd)) { fprintf(stderr, "Failed to write final i915 timestamp correlation data: %s\n",Signature · 1 line
2.25.1
PATCH v6 27/36 lib/i915/perf: add support for new EuDualSubslicesTotalCount var
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
We already assume that the number of subslices is the same as the number of dualsubslices.
--- lib/i915/perf-configs/codegen.py | 1 + lib/i915/perf-configs/mdapi-xml-convert.py | 3 ++- 2 files changed, 3 insertions(+), 1 deletion(-)
lib/i915/perf-configs/codegen.py +1 -0
@@ -168,6 +168,7 @@ class Gen: "$EuCoresTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" },
lib/i915/perf-configs/mdapi-xml-convert.py +2 -1
@@ -167,7 +167,8 @@ sys_vars = { "EuCoresTotalCount", "GpuTimestampFrequency", "SliceMask", "SubsliceMask", - "EuSubslicesTotalCount" + "EuSubslicesTotalCount", + "EuDualSubslicesTotalCount", } def underscore(name):
Signature · 1 line
2.25.1
PATCH v6 28/36 lib/i915/perf-config: extend the device info
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
This will allow equations to check for finer information on the topology. Also add EuDualSubslicesSlice0123Count.
--- lib/i915/perf-configs/codegen.py | 1 + lib/i915/perf.c | 31 ++++++++++++ lib/i915/perf.h | 85 +++++++++++++++++++++++++++++++- lib/meson.build | 4 +- 4 files changed, 118 insertions(+), 3 deletions(-)
lib/i915/perf-configs/codegen.py +1 -0
@@ -169,6 +169,7 @@ class Gen: "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesSlice0123Count": { 'c': "perf->devinfo.n_eu_sub_slices_half_slices" }, "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" },
lib/i915/perf.c +31 -0
@@ -153,6 +153,10 @@ intel_perf_for_devinfo(uint32_t device_id, {const struct intel_device_info *devinfo =intel_get_device_info (device_id);struct intel_perf *perf; +uint32_t subslice_mask_len; +uint32_t eu_mask_len; +uint32_t half_max_subslices; +uint64_t half_subslices_mask;int bits_per_subslice;if (!devinfo)
@@ -180,6 +184,25 @@ intel_perf_for_devinfo(uint32_t device_id, "%s", devinfo->codename); } +/* Store i915 topology. */ + perf->devinfo.max_slices = topology->max_slices; + perf->devinfo.max_subslices_per_slice = topology->max_subslices; + perf->devinfo.max_eu_per_subslice = topology->max_eus_per_subslice; + + subslice_mask_len = + topology->max_slices * topology->subslice_stride; +assert (sizeof (perf->devinfo.subslice_masks) >= subslice_mask_len); +memcpy (perf->devinfo.subslice_masks, + &topology->data[topology->subslice_offset], + subslice_mask_len); + + eu_mask_len = topology->eu_stride * + topology->max_subslices * topology->max_slices; +assert (sizeof (perf->devinfo.eu_masks) >= eu_mask_len); +memcpy (perf->devinfo.eu_masks, + &topology->data[topology->eu_offset], + eu_mask_len); +/* On Gen11+ the equations from the xml files expect an 8bits * mask per subslice, versus only 3bits on prior Gens. */
@@ -205,6 +228,14 @@ intel_perf_for_devinfo(uint32_t device_id, perf->devinfo.n_eu_slices =__builtin_popcount (perf->devinfo.slice_mask); perf->devinfo.n_eu_sub_slices =__builtin_popcount (perf->devinfo.subslice_mask); +/* Compute number of subslices/dualsubslices in first half of +* the GPU. +*/ + half_max_subslices = topology->max_subslices /2 ; + half_subslices_mask = perf->devinfo.subslice_mask & + ((1 << half_max_subslices) -1 ); + perf->devinfo.n_eu_sub_slices_half_slices =__builtin_popcount (half_subslices_mask); +/* Valid on most generations except Gen9LP. */ perf->devinfo.eu_threads_count =7 ;
lib/i915/perf.h +84 -1
109 lines not on this page · raw message
lib/meson.build +2 -2
@@ -308,7 +308,7 @@ lib_igt_i915_perf_build = shared_library( dependencies: lib_igt_chipset, include_directories : inc, install: true, - soversion: '1.4') + soversion: '1.5') lib_igt_i915_perf = declare_dependency( link_with : lib_igt_i915_perf_build,
@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.4.0') +pkgconf.set('i915_perf_version', '1.5.0') configure_file( input : 'i915-perf.pc.in',
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 28/36] lib/i915/perf-config: extend the device info
Quoted 1 line · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Quoted 221 lines
This will allow equations to check for finer information on the topology. Also add EuDualSubslicesSlice0123Count.
Signed-off-by Lionel Landwerlin <[email protected]>---
lib/i915/perf-configs/codegen.py | 1 + lib/i915/perf.c | 31 ++++++++++++ lib/i915/perf.h | 85 +++++++++++++++++++++++++++++++- lib/meson.build | 4 +- 4 files changed, 118 insertions(+), 3 deletions(-)lib/i915/perf-configs/codegen.py +1
-0@@ -169,6 +169,7 @@ class Gen: "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesSlice0123Count": { 'c': "perf->devinfo.n_eu_sub_slices_half_slices" }, "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" },lib/i915/perf.c +31
-0@@ -153,6 +153,10 @@ intel_perf_for_devinfo(uint32_t device_id, {const struct intel_device_info *devinfo =intel_get_device_info (device_id);struct intel_perf *perf; +uint32_t subslice_mask_len; +uint32_t eu_mask_len; +uint32_t half_max_subslices; +uint64_t half_subslices_mask;int bits_per_subslice;if (!devinfo)@@ -180,6 +184,25 @@ intel_perf_for_devinfo(uint32_t device_id, "%s", devinfo->codename); } +/* Store i915 topology. */ + perf->devinfo.max_slices = topology->max_slices; + perf->devinfo.max_subslices_per_slice = topology->max_subslices; + perf->devinfo.max_eu_per_subslice = topology->max_eus_per_subslice; + + subslice_mask_len = + topology->max_slices * topology->subslice_stride; +assert (sizeof (perf->devinfo.subslice_masks) >= subslice_mask_len); +memcpy (perf->devinfo.subslice_masks, + &topology->data[topology->subslice_offset], + subslice_mask_len); + + eu_mask_len = topology->eu_stride * + topology->max_subslices * topology->max_slices; +assert (sizeof (perf->devinfo.eu_masks) >= eu_mask_len); +memcpy (perf->devinfo.eu_masks, + &topology->data[topology->eu_offset], + eu_mask_len); +/* On Gen11+ the equations from the xml files expect an 8bits * mask per subslice, versus only 3bits on prior Gens. */ @@ -205,6 +228,14 @@ intel_perf_for_devinfo(uint32_t device_id, perf->devinfo.n_eu_slices =__builtin_popcount (perf->devinfo.slice_mask); perf->devinfo.n_eu_sub_slices =__builtin_popcount (perf->devinfo.subslice_mask); +/* Compute number of subslices/dualsubslices in first half of +* the GPU. +*/ + half_max_subslices = topology->max_subslices /2 ; + half_subslices_mask = perf->devinfo.subslice_mask & + ((1 << half_max_subslices) -1 ); + perf->devinfo.n_eu_sub_slices_half_slices =__builtin_popcount (half_subslices_mask); +/* Valid on most generations except Gen9LP. */ perf->devinfo.eu_threads_count =7 ;lib/i915/perf.h +84
-1@@ -32,7 +32,11 @@ extern "C" {#include "igt_list.h" -struct intel_device_info ; +#define DIV_ROUND_UP (a, b) (((a) + (b) - 1) / (b)) + +#define INTEL_DEVICE_MAX_SLICES (6)/* Maximum on gfx10 */ +#define INTEL_DEVICE_MAX_SUBSLICES (8)/* Maximum on gfx11 */ +#define INTEL_DEVICE_MAX_EUS_PER_SUBSLICE (16)/* Maximum on gfx12 */ struct intel_perf_devinfo {char devname[20 ];@@ -77,12 +81,66 @@ struct intel_perf_devinfo {uint64_t n_eu_slices;/* Total number of subslices/dualsubslices */ uint64_t n_eu_sub_slices; +/* Number of subslices/dualsubslices in the first half of the +* slices. +*/ +uint64_t n_eu_sub_slices_half_slices;/* Mask of available subslices/dualsubslices */ uint64_t subslice_mask;/* Mask of available slices */ uint64_t slice_mask;/* Number of threads in one EU */ uint64_t eu_threads_count; + +/** +* Maximu number of slices present on this device (can be more than +* num_slices if some slices are fused). +*/ +uint16_t max_slices; + +/** +* Maximu number of subslices per slice present on this device (can be more +* than the maximum value in the num_subslices[] array if some subslices are +* fused). +*/ +uint16_t max_subslices_per_slice; + +/** +* Stride to access subslice_masks[]. +*/ +uint16_t subslice_slice_stride; + +/** +* Maximum number of EUs per subslice (can be more than +* num_eu_per_subslice if some EUs are fused off). +*/ +uint16_t max_eu_per_subslice; + +/** +* Strides to access eu_masks[]. +*/ +uint16_t eu_slice_stride; +uint16_t eu_subslice_stride; + +/** +* A bit mask of the slices available. +*/ +uint8_t slice_masks[DIV_ROUND_UP (INTEL_DEVICE_MAX_SLICES ,8 )]; + +/** +* An array of bit mask of the subslices available, use subslice_slice_stride +* to access this array. +*/ +uint8_t subslice_masks[INTEL_DEVICE_MAX_SLICES * +DIV_ROUND_UP (INTEL_DEVICE_MAX_SUBSLICES ,8 )]; + +/** +* An array of bit mask of EUs available, use eu_slice_stride & +* eu_subslice_stride to access this array. +*/ +uint8_t eu_masks[INTEL_DEVICE_MAX_SLICES * +INTEL_DEVICE_MAX_SUBSLICES * +DIV_ROUND_UP (INTEL_DEVICE_MAX_EUS_PER_SUBSLICE ,8 )]; }; typedef enum {@@ -232,6 +290,31 @@ struct intel_perf {struct drm_i915_perf_record_header ;struct drm_i915_query_topology_info ; +static inline bool +intel_perf_devinfo_slice_available (const struct intel_perf_devinfo *devinfo, +int slice) +{ +return (devinfo->slice_masks[slice /8 ] & (1U << (slice %8 ))) !=0 ; +} + +static inline bool +intel_perf_devinfo_subslice_available (const struct intel_perf_devinfo *devinfo, +int slice,int subslice) +{ +return (devinfo->subslice_masks[slice * devinfo->subslice_slice_stride + + subslice /8 ] & (1U << (subslice %8 ))) !=0 ; +} + +static inline bool +intel_perf_devinfo_eu_available (const struct intel_perf_devinfo *devinfo, +int slice,int subslice,int eu) +{ +unsigned subslice_offset = slice * devinfo->eu_slice_stride + + subslice * devinfo->eu_subslice_stride; + +return (devinfo->eu_masks[subslice_offset + eu /8 ] & (1U << eu %8 )) !=0 ; +} +struct intel_perf *intel_perf_for_fd (int drm_fd); struct intel_perf *intel_perf_for_devinfo(uint32_t device_id, uint32_t revision,lib/meson.build +1
-1@@ -308,7 +308,7 @@ lib_igt_i915_perf_build = shared_library( dependencies: lib_igt_chipset, include_directories : inc, install: true, - soversion: '1.4') + soversion: '1.5')
lib_igt_i915_perf= declare_dependency(
link_with:lib_igt_i915_perf_build,lib/meson.build +1
-1@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.4.0') +pkgconf.set('i915_perf_version', '1.5.0')configure_file(
input : 'i915-perf.pc.in',Signature · 1 line
2.25.1
PATCH v6 29/36 i915/perf: update import script
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
--- lib/i915/perf-configs/mdapi-xml-convert.py | 180 ++++++++++++--------- lib/i915/perf-configs/oa-bdw.xml | 2 +- lib/i915/perf-configs/oa_guid_registry.py | 15 +- lib/i915/perf-configs/update-guids.py | 23 +-- lib/i915/perf.h | 1 + 5 files changed, 131 insertions(+), 90 deletions(-)
lib/i915/perf-configs/mdapi-xml-convert.py +105 -75
@@ -45,6 +45,8 @@ import sys import time import uuid +import codegen + import xml.etree.ElementTree as et import xml.sax.saxutils as saxutils
@@ -54,6 +56,55 @@ import oa_guid_registry as oa_registry # MDAPI configs include writes to some non-config registers, # thus the blacklists... +hsw_chipset_params = { + 'a_offset': 12, + 'b_offset': 192, + 'c_offset': 224, + 'oa_report_size': 256, + 'registers': { + # TODO extend the symbol table for nicer output... + 0x2710: { 'name': 'OASTARTTRIG1' }, + 0x2714: { 'name': 'OASTARTTRIG1' }, + 0x2718: { 'name': 'OASTARTTRIG1' }, + 0x271c: { 'name': 'OASTARTTRIG1' }, + 0x2720: { 'name': 'OASTARTTRIG1' }, + 0x2724: { 'name': 'OASTARTTRIG6' }, + 0x2728: { 'name': 'OASTARTTRIG7' }, + 0x272c: { 'name': 'OASTARTTRIG8' }, + 0x2740: { 'name': 'OAREPORTTRIG1' }, + 0x2744: { 'name': 'OAREPORTTRIG2' }, + 0x2748: { 'name': 'OAREPORTTRIG3' }, + 0x274c: { 'name': 'OAREPORTTRIG4' }, + 0x2750: { 'name': 'OAREPORTTRIG5' }, + 0x2754: { 'name': 'OAREPORTTRIG6' }, + 0x2758: { 'name': 'OAREPORTTRIG7' }, + 0x275c: { 'name': 'OAREPORTTRIG8' }, + 0x2770: { 'name': 'OACEC0_0' }, + 0x2774: { 'name': 'OACEC0_1' }, + 0x2778: { 'name': 'OACEC1_0' }, + 0x277c: { 'name': 'OACEC1_1' }, + 0x2780: { 'name': 'OACEC2_0' }, + 0x2784: { 'name': 'OACEC2_1' }, + 0x2788: { 'name': 'OACEC3_0' }, + 0x278c: { 'name': 'OACEC3_1' }, + 0x2790: { 'name': 'OACEC4_0' }, + 0x2794: { 'name': 'OACEC4_1' }, + 0x2798: { 'name': 'OACEC5_0' }, + 0x279c: { 'name': 'OACEC5_1' }, + 0x27a0: { 'name': 'OACEC6_0' }, + 0x27a4: { 'name': 'OACEC6_1' }, + 0x27a8: { 'name': 'OACEC7_0' }, + 0x27ac: { 'name': 'OACEC7_1' }, + }, + 'config_reg_blacklist': { + 0x2364, # OASTATUS1 register + }, + 'register_offsets': { + 0x1f0: 'PERFCNT 0', + 0x1f8: 'PERFCNT 1', + }, +} + gen8_11_chipset_params = { 'a_offset': 16, 'b_offset': 192,
@@ -83,50 +134,7 @@ xehpsdv_chipset_params = { } chipsets = { - 'HSW': { - 'a_offset': 12, - 'b_offset': 192, - 'c_offset': 224, - 'oa_report_size': 256, - 'registers': { - # TODO extend the symbol table for nicer output... - 0x2710: { 'name': 'OASTARTTRIG1' }, - 0x2714: { 'name': 'OASTARTTRIG1' }, - 0x2718: { 'name': 'OASTARTTRIG1' }, - 0x271c: { 'name': 'OASTARTTRIG1' }, - 0x2720: { 'name': 'OASTARTTRIG1' }, - 0x2724: { 'name': 'OASTARTTRIG6' }, - 0x2728: { 'name': 'OASTARTTRIG7' }, - 0x272c: { 'name': 'OASTARTTRIG8' }, - 0x2740: { 'name': 'OAREPORTTRIG1' }, - 0x2744: { 'name': 'OAREPORTTRIG2' }, - 0x2748: { 'name': 'OAREPORTTRIG3' }, - 0x274c: { 'name': 'OAREPORTTRIG4' }, - 0x2750: { 'name': 'OAREPORTTRIG5' }, - 0x2754: { 'name': 'OAREPORTTRIG6' }, - 0x2758: { 'name': 'OAREPORTTRIG7' }, - 0x275c: { 'name': 'OAREPORTTRIG8' }, - 0x2770: { 'name': 'OACEC0_0' }, - 0x2774: { 'name': 'OACEC0_1' }, - 0x2778: { 'name': 'OACEC1_0' }, - 0x277c: { 'name': 'OACEC1_1' }, - 0x2780: { 'name': 'OACEC2_0' }, - 0x2784: { 'name': 'OACEC2_1' }, - 0x2788: { 'name': 'OACEC3_0' }, - 0x278c: { 'name': 'OACEC3_1' }, - 0x2790: { 'name': 'OACEC4_0' }, - 0x2794: { 'name': 'OACEC4_1' }, - 0x2798: { 'name': 'OACEC5_0' }, - 0x279c: { 'name': 'OACEC5_1' }, - 0x27a0: { 'name': 'OACEC6_0' }, - 0x27a4: { 'name': 'OACEC6_1' }, - 0x27a8: { 'name': 'OACEC7_0' }, - 0x27ac: { 'name': 'OACEC7_1' }, - }, - 'config_reg_blacklist': { - 0x2364, # OASTATUS1 register - }, - }, + 'HSW': hsw_chipset_params, 'BDW': gen8_11_chipset_params, 'CHV': gen8_11_chipset_params, 'SKL': gen8_11_chipset_params,
@@ -141,8 +149,11 @@ chipsets = { 'RKL': gen8_11_chipset_params, 'DG1': gen8_11_chipset_params, 'ADL': gen8_11_chipset_params, + 'ACM': xehpsdv_chipset_params, } +xehp_plus = ( 'ACM', ) + register_types = { 'OA', 'NOA', 'FLEX', 'PM' } default_set_blacklist = { "RenderDX1x", # TODO: rename to something non 'DX'
@@ -158,23 +169,11 @@ counter_blacklist = { # investiguate how to get this value. } -sys_vars = { "EuCoresTotalCount", - "EuSlicesTotalCount", - "SamplersTotalCount", - "EuThreadsCount", - "GpuMinFrequencyMHz", - "GpuMaxFrequencyMHz", - "GpuTimestampFrequency", - "SliceMask", - "SubsliceMask", - "EuSubslicesTotalCount", - "EuDualSubslicesTotalCount", - } - def underscore(name): s = re.sub('MHz', 'Mhz', name) s = re.sub('\.', '_', s) s = re.sub('(.)([A-Z][a-z]+)', r'\1_\2', s) + s = re.sub('#', '_', s) return re.sub('([a-z0-9])([A-Z])', r'\1_\2', s).lower() def print_err(*args):
@@ -215,7 +214,7 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): assert 0 elif offset < b_offset: a_cnt_offset = int((offset - a_offset) / 4) - if chipset == "XEHPSDV": + if chipset in xehp_plus: # Most A counters are in a contiguous array, except # this A37. if a_cnt_offset == 42:
@@ -246,7 +245,7 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): return "C " + str(idx - 54) + " READ" else: return "{0} READ".format(read_value(chipset, offset)) - elif chipset == "XEHPSDV": + elif chipset in xehp_plus: # For XEHPSDV the array of accumulated counters is # assumed to start with a GPU_TIME then GPU_CLOCK, # then 38 A counters, then 8 B counters and finally
@@ -354,10 +353,38 @@ def strip_dx_apis(text): return stripped.strip() +def add_gpu_core_clocks_if_missing(metric_set, counters, counter_deps): + if len(counters) < 1: + return + + for name,element in counters.items(): + if name == 'GpuCoreClocks': + return + + print_err("WARNING: add missing GpuCoreClocks counter for MetricSets=\"{0}\"".format(metric_set.get('ShortName'))) + counter = et.SubElement(metric_set, 'Metric') + counter.set("SymbolName", "GpuCoreClocks") + counter.set("SignalName", "oa.fixed") + counter.set("ShortName", "GPU Core Clocks") + counter.set("LongName", "The total number of GPU core clocks elapsed during the measurement.") + counter.set("Group", "GPU") + counter.set("UsageFlags", "Tier1 Frame Batch Draw") + counter.set("MetricType", "EVENT") + counter.set("ResultType", "UINT64") + counter.set("MetricUnits", "cycles") + counter.set("HWUnitType", "GPU") + counter.set("SnapshotReportReadEquation", "dw@0x0c") + counter.set("SnapshotReportDeltaFunction", "DELTA 32") + counter.set("DeltaReportReadEquation", "qw@0x08") + counter.set("NormalizationEquation", "") + + counters["GpuCoreClocks"] = counter + counter_deps["GpuCoreClocks"] = [] + # For recursively appending counters in order of dependencies... def append_deps_and_counter(mdapi_counter, mdapi_counters, deps, sorted_array, sorted_set): - symbol_name = mdapi_counter.get('SymbolName') + symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName')) if symbol_name in sorted_set: return
@@ -562,6 +589,7 @@ guids = {} guids_xml = et.parse(args.guids) for guid in guids_xml.findall(".//guid"): hashing_key = oa_registry.Registry.chipset_derive_hash(guid.get('chipset'), + guid.get('name'), guid.get('mdapi_config_hash')) guids[hashing_key] = guid.get('id')
@@ -569,6 +597,14 @@ for arg in args.xml: mdapi = et.parse(arg) concurrent_group = mdapi.find(".//ConcurrentGroup") + chipset = oa_registry.Registry.chipset_name(concurrent_group.get('SupportedHW')) + + chipset_fullname = chipset + if concurrent_group.get('SupportedGT') != None: + chipset_fullname = chipset_fullname + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')) + if chipset not in chipsets: + print_err("WARNING: unsupported chipset {0}, consider updating {1}".format(chipset, __file__)) + continue for mdapi_set in mdapi.findall(".//MetricSet"):
@@ -576,20 +612,12 @@ for arg in args.xml: if "OGL" not in apis and "OCL" not in apis and "MEDIA" not in apis: continue - set_symbol_name = mdapi_set.get('SymbolName') + set_symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_set.get('SymbolName')) if set_symbol_name in sets: print_err("WARNING: duplicate set named \"" + set_symbol_name + "\" (SKIPPING)") continue - chipset = oa_registry.Registry.chipset_name(mdapi_set.get('SupportedHW')) - chipset_fullname = chipset - if concurrent_group.get('SupportedGT') != None: - chipset_fullname = chipset_fullname + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')) - if chipset not in chipsets: - print_err("WARNING: unsupported chipset {0}, consider updating {1}".format(chipset, __file__)) - continue - if args.whitelist: set_whitelist = args.whitelist.split() if set_symbol_name not in set_whitelist:
@@ -666,6 +694,7 @@ for arg in args.xml: mdapi_hw_config_hash = oa_registry.Registry.mdapi_hw_config_hash(mdapi_set) guid_hash = oa_registry.Registry.chipset_derive_hash(chipset_fullname.lower(), + set_symbol_name, mdapi_hw_config_hash) hw_config_hash = oa_registry.Registry.hw_config_hash(set)
@@ -693,7 +722,7 @@ for arg in args.xml: mdapi_counter_deps = {} for mdapi_counter in mdapi_set.findall("Metrics/Metric"): - symbol_name = mdapi_counter.get('SymbolName') + symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName')) if symbol_name in counter_blacklist: continue;
@@ -712,12 +741,13 @@ for arg in args.xml: equations = expand_macros(equations) equations = equations.replace('$$', "$") for token in equations.split(): - if token[0] == '$' and token[1:] not in sys_vars and token[1:] != "Self": + if token[0] == '$' and not codegen.is_hw_var(token) and token[1:] != "Self": deps.append(token[1:]) mdapi_counters[symbol_name] = mdapi_counter mdapi_counter_deps[symbol_name] = deps + add_gpu_core_clocks_if_missing(mdapi_set, mdapi_counters, mdapi_counter_deps) sorted_mdapi_counters = sort_counters(mdapi_counters, mdapi_counter_deps) for mdapi_counter in sorted_mdapi_counters:
@@ -746,11 +776,11 @@ for arg in args.xml: #mdapi_counter.set('DeltaReportReadEquation', '$GpuCoreClocks $GpuTime UDIV') - symbol_name = mdapi_counter.get('SymbolName') + symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName')) counter = et.SubElement(set, 'counter') counter.set('name', apply_aliases(mdapi_counter.get('ShortName'), aliases)) - counter.set('symbol_name', mdapi_counter.get('SymbolName')) + counter.set('symbol_name', oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName'))) counter.set('underscore_name', underscore(mdapi_counter.get('SymbolName'))) counter.set('description', apply_aliases(mdapi_counter.get('LongName'), aliases)) counter.set('mdapi_group', apply_aliases(to_text(mdapi_counter.get('Group')), aliases))
@@ -943,7 +973,7 @@ for arg in args.xml: # print_err("WARNING: Counter equation (\"" + equation + "\") references un-kept raw equation of another counter : MetricSet=\"" + \ # mdapi_set.get('ShortName') + "\" Metric=\"" + mdapi_counter.get('ShortName') + "\"") - elif token[1:] not in raw_equations and token[1:] not in sys_vars: + elif token[1:] not in raw_equations and not codegen.is_hw_var(token): print_err("Unknown variable name: \"" + token + "\" in equation \"" + equation + "\"") symbol_name = counter.get('symbol_name')
@@ -952,7 +982,7 @@ for arg in args.xml: equation = equation.replace('$$', "$") for token in equation.split(): if token[0] == '$': - if token[1:] not in counters and token[1:] not in sys_vars: + if token[1:] not in counters and not codegen.is_hw_var(token): print_err("WARNING: Counter equation (\"" + equation + "\") with unknown variable " + \ token + " (maybe skipped counter): MetricSet=\"" + mdapi_set.get('ShortName') + \ "\" Metric=\"" + mdapi_counter.get('SymbolName') + "(" + mdapi_counter.get('ShortName') + \
lib/i915/perf-configs/oa-bdw.xml +1 -1
@@ -1,5 +1,5 @@ <?xml version="1.0"?> -<metrics version="1619078710" merge_md5=""> +<metrics version="1650881753" merge_md5=""> <set name="Render Metrics Basic set" chipset="BDW" symbol_name="RenderBasic"
lib/i915/perf-configs/oa_guid_registry.py +10 -5
@@ -72,16 +72,15 @@ class Registry: return hashlib.md5(registers_str).hexdigest() - @staticmethod - def chipset_derive_hash(chipset, hash): - """Derive a HW config hash for a given chipset. + def chipset_derive_hash(chipset, set_name, hash): + """Derive a HW config hash for a given chipset & set name. This helps us avoiding collisions with identical config across different Gen or GT. """ - return "%s-%s" % (chipset, hash) + return "%s-%s-%s" % (chipset, set_name, hash) @staticmethod
@@ -99,7 +98,9 @@ class Registry: 'EHL', 'TGL', 'RKL', - 'DG1' ) + 'DG1', + 'ACM', + 'PVC' ) if name in known_chipsets: return name
@@ -110,3 +111,7 @@ class Registry: @staticmethod def gt_name(name): return re.sub(' ', '', name) + + @staticmethod + def sanitize_symbol_name(text): + return text.replace('#', "_")
lib/i915/perf-configs/update-guids.py +14 -9
@@ -60,8 +60,10 @@ def print_err(*args): sys.stderr.write(' '.join(map(str,args)) + '\n') def guid_hashing_key(guid_obj): - return oa_registry.Registry.chipset_derive_hash(guid_obj['chipset'], - guid_obj['mdapi_config_hash']) + ret = oa_registry.Registry.chipset_derive_hash(guid_obj['chipset'], + guid_obj['name'], + guid_obj['mdapi_config_hash']) + return ret parser = argparse.ArgumentParser() parser.add_argument("xml", nargs="+", help="XML description of metrics")
@@ -101,6 +103,7 @@ for guid in guids_xml.findall(".//guid"): if 'mdapi_config_hash' in guid_obj: hashing_key = oa_registry.Registry.chipset_derive_hash(guid_obj['chipset'], + guid_obj['name'], guid_obj['mdapi_config_hash']) mdapi_config_hash_guid_table[hashing_key] = guid_obj
@@ -121,20 +124,20 @@ for arg in args.xml: internal = et.parse(arg) concurrent_group = internal.find(".//ConcurrentGroup") + chipset = oa_registry.Registry.chipset_name(concurrent_group.get('SupportedHW')).lower() + if concurrent_group.get('SupportedGT') != None: + chipset = chipset + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')).lower() + for mdapi_set in internal.findall(".//MetricSet"): mdapi_config_hash = oa_registry.Registry.mdapi_hw_config_hash(mdapi_set) - chipset = oa_registry.Registry.chipset_name(mdapi_set.get('SupportedHW')).lower() - if concurrent_group.get('SupportedGT') != None: - chipset = chipset + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')).lower() - - set_name = mdapi_set.get('SymbolName') + set_name = oa_registry.Registry.sanitize_symbol_name(mdapi_set.get('SymbolName')) name = chipset + "_" + set_name; - hashing_key = oa_registry.Registry.chipset_derive_hash(chipset, mdapi_config_hash) + hashing_key = oa_registry.Registry.chipset_derive_hash(chipset, set_name, mdapi_config_hash) if hashing_key in mdapi_config_hash_guid_table: guid_obj = mdapi_config_hash_guid_table[hashing_key]
@@ -174,7 +177,9 @@ chipsets = [ 'hsw', 'bxt', 'glk', 'cnl', 'icl', 'ehl', - 'tglgt1', 'tglgt2', 'rkl', 'dg1', 'adl' ] + 'tglgt1', 'tglgt2', 'rkl', 'dg1', 'adl', + 'acmgt1', 'acmgt2', 'acmgt3', +] for chipset in chipsets: filename = 'oa-' + chipset + '.xml'
lib/i915/perf.h +1 -0
@@ -189,6 +189,7 @@ typedef enum {INTEL_PERF_LOGICAL_COUNTER_UNIT_EU_ATOMIC_REQUESTS_TO_L3_CACHE_LINES ,INTEL_PERF_LOGICAL_COUNTER_UNIT_EU_REQUESTS_TO_L3_CACHE_LINES ,INTEL_PERF_LOGICAL_COUNTER_UNIT_EU_BYTES_PER_L3_CACHE_LINE , +INTEL_PERF_LOGICAL_COUNTER_UNIT_GBPS , INTEL_PERF_LOGICAL_COUNTER_UNIT_MAX } intel_perf_logical_counter_unit_t;
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 29/36] i915/perf: update import script
Quoted 3 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 465 lines
---
lib/i915/perf-configs/mdapi-xml-convert.py | 180 ++++++++++++--------- lib/i915/perf-configs/oa-bdw.xml | 2 +- lib/i915/perf-configs/oa_guid_registry.py | 15 +- lib/i915/perf-configs/update-guids.py | 23 +-- lib/i915/perf.h | 1 + 5 files changed, 131 insertions(+), 90 deletions(-)lib/i915/perf-configs/mdapi-xml-convert.py +0
-0@@ -45,6 +45,8 @@ import sysimport time
import uuidlib/i915/perf-configs/mdapi-xml-convert.py +2
-0+import codegen +import xml.etree.ElementTree as et
import xml.sax.saxutils as saxutilslib/i915/perf-configs/mdapi-xml-convert.py +53
-44@@ -54,6 +56,55 @@ import oa_guid_registry as oa_registry # MDAPI configs include writes to some non-config registers, # thus the blacklists... +hsw_chipset_params = { + 'a_offset': 12, + 'b_offset': 192, + 'c_offset': 224, + 'oa_report_size': 256, + 'registers': { + # TODO extend the symbol table for nicer output... + 0x2710: { 'name': 'OASTARTTRIG1' }, + 0x2714: { 'name': 'OASTARTTRIG1' }, + 0x2718: { 'name': 'OASTARTTRIG1' }, + 0x271c: { 'name': 'OASTARTTRIG1' }, + 0x2720: { 'name': 'OASTARTTRIG1' }, + 0x2724: { 'name': 'OASTARTTRIG6' }, + 0x2728: { 'name': 'OASTARTTRIG7' }, + 0x272c: { 'name': 'OASTARTTRIG8' }, + 0x2740: { 'name': 'OAREPORTTRIG1' }, + 0x2744: { 'name': 'OAREPORTTRIG2' }, + 0x2748: { 'name': 'OAREPORTTRIG3' }, + 0x274c: { 'name': 'OAREPORTTRIG4' }, + 0x2750: { 'name': 'OAREPORTTRIG5' }, + 0x2754: { 'name': 'OAREPORTTRIG6' }, + 0x2758: { 'name': 'OAREPORTTRIG7' }, + 0x275c: { 'name': 'OAREPORTTRIG8' }, + 0x2770: { 'name': 'OACEC0_0' }, + 0x2774: { 'name': 'OACEC0_1' }, + 0x2778: { 'name': 'OACEC1_0' }, + 0x277c: { 'name': 'OACEC1_1' }, + 0x2780: { 'name': 'OACEC2_0' }, + 0x2784: { 'name': 'OACEC2_1' }, + 0x2788: { 'name': 'OACEC3_0' }, + 0x278c: { 'name': 'OACEC3_1' }, + 0x2790: { 'name': 'OACEC4_0' }, + 0x2794: { 'name': 'OACEC4_1' }, + 0x2798: { 'name': 'OACEC5_0' }, + 0x279c: { 'name': 'OACEC5_1' }, + 0x27a0: { 'name': 'OACEC6_0' }, + 0x27a4: { 'name': 'OACEC6_1' }, + 0x27a8: { 'name': 'OACEC7_0' }, + 0x27ac: { 'name': 'OACEC7_1' }, + }, + 'config_reg_blacklist': { + 0x2364, # OASTATUS1 register + }, + 'register_offsets': { + 0x1f0: 'PERFCNT 0', + 0x1f8: 'PERFCNT 1', + }, +} + gen8_11_chipset_params = { 'a_offset': 16, 'b_offset': 192,@@ -83,50 +134,7 @@ xehpsdv_chipset_params = { } chipsets = { - 'HSW': { - 'a_offset': 12, - 'b_offset': 192, - 'c_offset': 224, - 'oa_report_size': 256, - 'registers': { - # TODO extend the symbol table for nicer output... - 0x2710: { 'name': 'OASTARTTRIG1' }, - 0x2714: { 'name': 'OASTARTTRIG1' }, - 0x2718: { 'name': 'OASTARTTRIG1' }, - 0x271c: { 'name': 'OASTARTTRIG1' }, - 0x2720: { 'name': 'OASTARTTRIG1' }, - 0x2724: { 'name': 'OASTARTTRIG6' }, - 0x2728: { 'name': 'OASTARTTRIG7' }, - 0x272c: { 'name': 'OASTARTTRIG8' }, - 0x2740: { 'name': 'OAREPORTTRIG1' }, - 0x2744: { 'name': 'OAREPORTTRIG2' }, - 0x2748: { 'name': 'OAREPORTTRIG3' }, - 0x274c: { 'name': 'OAREPORTTRIG4' }, - 0x2750: { 'name': 'OAREPORTTRIG5' }, - 0x2754: { 'name': 'OAREPORTTRIG6' }, - 0x2758: { 'name': 'OAREPORTTRIG7' }, - 0x275c: { 'name': 'OAREPORTTRIG8' }, - 0x2770: { 'name': 'OACEC0_0' }, - 0x2774: { 'name': 'OACEC0_1' }, - 0x2778: { 'name': 'OACEC1_0' }, - 0x277c: { 'name': 'OACEC1_1' }, - 0x2780: { 'name': 'OACEC2_0' }, - 0x2784: { 'name': 'OACEC2_1' }, - 0x2788: { 'name': 'OACEC3_0' }, - 0x278c: { 'name': 'OACEC3_1' }, - 0x2790: { 'name': 'OACEC4_0' }, - 0x2794: { 'name': 'OACEC4_1' }, - 0x2798: { 'name': 'OACEC5_0' }, - 0x279c: { 'name': 'OACEC5_1' }, - 0x27a0: { 'name': 'OACEC6_0' }, - 0x27a4: { 'name': 'OACEC6_1' }, - 0x27a8: { 'name': 'OACEC7_0' }, - 0x27ac: { 'name': 'OACEC7_1' }, - }, - 'config_reg_blacklist': { - 0x2364, # OASTATUS1 register - }, - }, + 'HSW': hsw_chipset_params, 'BDW': gen8_11_chipset_params, 'CHV': gen8_11_chipset_params, 'SKL': gen8_11_chipset_params,@@ -141,8 +149,11 @@ chipsets = { 'RKL': gen8_11_chipset_params, 'DG1': gen8_11_chipset_params, 'ADL': gen8_11_chipset_params, + 'ACM': xehpsdv_chipset_params, } +xehp_plus = ( 'ACM', ) + register_types = { 'OA', 'NOA', 'FLEX', 'PM' }
default_set_blacklist= { "RenderDX1x", # TODO: rename to something non 'DX'lib/i915/perf-configs/mdapi-xml-convert.py +0
-13@@ -158,23 +169,11 @@ counter_blacklist = { # investiguate how to get this value. } -sys_vars = { "EuCoresTotalCount", - "EuSlicesTotalCount", - "SamplersTotalCount", - "EuThreadsCount", - "GpuMinFrequencyMHz", - "GpuMaxFrequencyMHz", - "GpuTimestampFrequency", - "SliceMask", - "SubsliceMask", - "EuSubslicesTotalCount", - "EuDualSubslicesTotalCount", - } -def underscore(name):
s = re.sub('MHz', 'Mhz', name) s = re.sub('\.', '_', s) s = re.sub('(.)([A-Z][a-z]+)', r'\1_\2', s) + s = re.sub('#', '_', s) return re.sub('([a-z0-9])([A-Z])', r'\1_\2', s).lower()def
print_err(*args):lib/i915/perf-configs/mdapi-xml-convert.py +31
-3@@ -215,7 +214,7 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): assert 0 elif offset < b_offset: a_cnt_offset = int((offset - a_offset) / 4) - if chipset == "XEHPSDV": + if chipset in xehp_plus: # Most A counters are in a contiguous array, except # this A37. if a_cnt_offset == 42:@@ -246,7 +245,7 @@ def read_token_to_rpn_read(chipset, token, raw_offsets): return "C " + str(idx - 54) + " READ" else: return "{0} READ".format(read_value(chipset, offset)) - elif chipset == "XEHPSDV": + elif chipset in xehp_plus: # For XEHPSDV the array of accumulated counters is # assumed to start with a GPU_TIME then GPU_CLOCK, # then 38 A counters, then 8 B counters and finally@@ -354,10 +353,38 @@ def strip_dx_apis(text): return stripped.strip() +def add_gpu_core_clocks_if_missing(metric_set, counters, counter_deps): + if len(counters) < 1: + return + + for name,element in counters.items(): + if name == 'GpuCoreClocks': + return + + print_err("WARNING: add missing GpuCoreClocks counter for MetricSets=\"{0}\"".format(metric_set.get('ShortName'))) + counter = et.SubElement(metric_set, 'Metric') + counter.set("SymbolName", "GpuCoreClocks") + counter.set("SignalName", "oa.fixed") + counter.set("ShortName", "GPU Core Clocks") + counter.set("LongName", "The total number of GPU core clocks elapsed during the measurement.") + counter.set("Group", "GPU") + counter.set("UsageFlags", "Tier1 Frame Batch Draw") + counter.set("MetricType", "EVENT") + counter.set("ResultType", "UINT64") + counter.set("MetricUnits", "cycles") + counter.set("HWUnitType", "GPU") + counter.set("SnapshotReportReadEquation", "dw@0x0c") + counter.set("SnapshotReportDeltaFunction", "DELTA 32") + counter.set("DeltaReportReadEquation", "qw@0x08") + counter.set("NormalizationEquation", "") + + counters["GpuCoreClocks"] = counter + counter_deps["GpuCoreClocks"] = [] + # For recursively appending counters in order of dependencies... def append_deps_and_counter(mdapi_counter, mdapi_counters, deps, sorted_array, sorted_set): - symbol_name = mdapi_counter.get('SymbolName') + symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName')) if symbol_name in sorted_set: return@@ -562,6 +589,7 @@ guids = {} guids_xml = et.parse(args.guids)for guid in
guids_xml.findall(".//guid"):
hashing_key= oa_registry.Registry.chipset_derive_hash(guid.get('chipset'),lib/i915/perf-configs/mdapi-xml-convert.py +18
-15+ guid.get('name'), guid.get('mdapi_config_hash')) guids[hashing_key] = guid.get('id')@@ -569,6 +597,14 @@ for arg in args.xml: mdapi = et.parse(arg) concurrent_group = mdapi.find(".//ConcurrentGroup") + chipset = oa_registry.Registry.chipset_name(concurrent_group.get('SupportedHW')) + + chipset_fullname = chipset + if concurrent_group.get('SupportedGT') != None: + chipset_fullname = chipset_fullname + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')) + if chipset not in chipsets: + print_err("WARNING: unsupported chipset {0}, consider updating {1}".format(chipset, __file__)) + continue for mdapi_set in mdapi.findall(".//MetricSet"):@@ -576,20 +612,12 @@ for arg in args.xml: if "OGL" not in apis and "OCL" not in apis and "MEDIA" not in apis: continue - set_symbol_name = mdapi_set.get('SymbolName') + set_symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_set.get('SymbolName')) if set_symbol_name in sets: print_err("WARNING: duplicate set named \"" + set_symbol_name + "\" (SKIPPING)") continue - chipset = oa_registry.Registry.chipset_name(mdapi_set.get('SupportedHW')) - chipset_fullname = chipset - if concurrent_group.get('SupportedGT') != None: - chipset_fullname = chipset_fullname + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')) - if chipset not in chipsets: - print_err("WARNING: unsupported chipset {0}, consider updating {1}".format(chipset, __file__)) - continue - if args.whitelist: set_whitelist = args.whitelist.split() if set_symbol_name not in set_whitelist:@@ -666,6 +694,7 @@ for arg in args.xml: mdapi_hw_config_hash = oa_registry.Registry.mdapi_hw_config_hash(mdapi_set) guid_hash = oa_registry.Registry.chipset_derive_hash(chipset_fullname.lower(), + set_symbol_name, mdapi_hw_config_hash) hw_config_hash = oa_registry.Registry.hw_config_hash(set)@@ -693,7 +722,7 @@ for arg in args.xml: mdapi_counter_deps = {} for mdapi_counter in mdapi_set.findall("Metrics/Metric"): - symbol_name = mdapi_counter.get('SymbolName') + symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName')) if symbol_name in counter_blacklist: continue;@@ -712,12 +741,13 @@ for arg in args.xml: equations = expand_macros(equations) equations = equations.replace('$$', "$") for token in equations.split(): - if token[0] == '$' and token[1:] not in sys_vars and token[1:] != "Self": + if token[0] == '$' and not codegen.is_hw_var(token) and token[1:] != "Self": deps.append(token[1:]) mdapi_counters[symbol_name] = mdapi_counter mdapi_counter_deps[symbol_name] = deps + add_gpu_core_clocks_if_missing(mdapi_set, mdapi_counters, mdapi_counter_deps) sorted_mdapi_counters = sort_counters(mdapi_counters, mdapi_counter_deps) for mdapi_counter in sorted_mdapi_counters:@@ -746,11 +776,11 @@ for arg in args.xml: #mdapi_counter.set('DeltaReportReadEquation', '$GpuCoreClocks $GpuTime UDIV') - symbol_name = mdapi_counter.get('SymbolName') + symbol_name = oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName')) counter = et.SubElement(set, 'counter') counter.set('name', apply_aliases(mdapi_counter.get('ShortName'), aliases)) - counter.set('symbol_name', mdapi_counter.get('SymbolName')) + counter.set('symbol_name', oa_registry.Registry.sanitize_symbol_name(mdapi_counter.get('SymbolName'))) counter.set('underscore_name', underscore(mdapi_counter.get('SymbolName'))) counter.set('description', apply_aliases(mdapi_counter.get('LongName'), aliases)) counter.set('mdapi_group', apply_aliases(to_text(mdapi_counter.get('Group')), aliases))@@ -943,7 +973,7 @@ for arg in args.xml: # print_err("WARNING: Counter equation (\"" + equation + "\") references un-kept raw equation of another counter : MetricSet=\"" + \ # mdapi_set.get('ShortName') + "\" Metric=\"" + mdapi_counter.get('ShortName') + "\"") - elif token[1:] not in raw_equations and token[1:] not in sys_vars: + elif token[1:] not in raw_equations and not codegen.is_hw_var(token): print_err("Unknown variable name: \"" + token + "\" in equation \"" + equation + "\"") symbol_name = counter.get('symbol_name')@@ -952,7 +982,7 @@ for arg in args.xml: equation = equation.replace('$$', "$") for token in equation.split(): if token[0] == '$': - if token[1:] not in counters and token[1:] not in sys_vars: + if token[1:] not in counters and not codegen.is_hw_var(token): print_err("WARNING: Counter equation (\"" + equation + "\") with unknown variable " + \ token + " (maybe skipped counter): MetricSet=\"" + mdapi_set.get('ShortName') + \ "\" Metric=\"" + mdapi_counter.get('SymbolName') + "(" + mdapi_counter.get('ShortName') + \lib/i915/perf-configs/oa-bdw.xml +0
-0@@ -1,5 +1,5 @@<?xml version="1.0"?>
-<metrics version="1619078710" merge_md5="">
+<metrics version="1650881753" merge_md5="">
<set name="Render Metrics Basic set"
chipset="BDW"
symbol_name="RenderBasic"lib/i915/perf-configs/oa_guid_registry.py +10
-5@@ -72,16 +72,15 @@ class Registry: return hashlib.md5(registers_str).hexdigest() - @staticmethod - def chipset_derive_hash(chipset, hash): - """Derive a HW config hash for a given chipset. + def chipset_derive_hash(chipset, set_name, hash): + """Derive a HW config hash for a given chipset & set name. This helps us avoiding collisions with identical config across different Gen or GT. """ - return "%s-%s" % (chipset, hash) + return "%s-%s-%s" % (chipset, set_name, hash) @staticmethod@@ -99,7 +98,9 @@ class Registry: 'EHL', 'TGL', 'RKL', - 'DG1' ) + 'DG1', + 'ACM', + 'PVC' ) if name in known_chipsets: return name@@ -110,3 +111,7 @@ class Registry: @staticmethod def gt_name(name): return re.sub(' ', '', name) + + @staticmethod + def sanitize_symbol_name(text): + return text.replace('#', "_")lib/i915/perf-configs/update-guids.py +0
-0@@ -60,8 +60,10 @@ def print_err(*args): sys.stderr.write(' '.join(map(str,args)) + '\n')def
guid_hashing_key(guid_obj):lib/i915/perf-configs/update-guids.py +4
-2- return oa_registry.Registry.chipset_derive_hash(guid_obj['chipset'], - guid_obj['mdapi_config_hash']) + ret = oa_registry.Registry.chipset_derive_hash(guid_obj['chipset'], + guid_obj['name'], + guid_obj['mdapi_config_hash']) + return retparser =
argparse.ArgumentParser()parser.add_argument("xml", nargs="+", help="XML description of metrics")lib/i915/perf-configs/update-guids.py +10
-7@@ -101,6 +103,7 @@ for guid in guids_xml.findall(".//guid"): if 'mdapi_config_hash' in guid_obj: hashing_key = oa_registry.Registry.chipset_derive_hash(guid_obj['chipset'], + guid_obj['name'], guid_obj['mdapi_config_hash']) mdapi_config_hash_guid_table[hashing_key] = guid_obj@@ -121,20 +124,20 @@ for arg in args.xml: internal = et.parse(arg) concurrent_group = internal.find(".//ConcurrentGroup") + chipset = oa_registry.Registry.chipset_name(concurrent_group.get('SupportedHW')).lower() + if concurrent_group.get('SupportedGT') != None: + chipset = chipset + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')).lower() + for mdapi_set in internal.findall(".//MetricSet"): mdapi_config_hash = oa_registry.Registry.mdapi_hw_config_hash(mdapi_set) - chipset = oa_registry.Registry.chipset_name(mdapi_set.get('SupportedHW')).lower() - if concurrent_group.get('SupportedGT') != None: - chipset = chipset + oa_registry.Registry.gt_name(concurrent_group.get('SupportedGT')).lower() - - set_name = mdapi_set.get('SymbolName') + set_name = oa_registry.Registry.sanitize_symbol_name(mdapi_set.get('SymbolName')) name = chipset + "_" + set_name; - hashing_key = oa_registry.Registry.chipset_derive_hash(chipset, mdapi_config_hash) + hashing_key = oa_registry.Registry.chipset_derive_hash(chipset, set_name, mdapi_config_hash) if hashing_key in mdapi_config_hash_guid_table: guid_obj = mdapi_config_hash_guid_table[hashing_key]@@ -174,7 +177,9 @@ chipsets = [ 'hsw', 'bxt', 'glk', 'cnl', 'icl', 'ehl', - 'tglgt1', 'tglgt2', 'rkl', 'dg1', 'adl' ] + 'tglgt1', 'tglgt2', 'rkl', 'dg1', 'adl', + 'acmgt1', 'acmgt2', 'acmgt3', +]for chipset in chipsets:
filename = 'oa-' + chipset + '.xml'lib/i915/perf.h +1
-0@@ -189,6 +189,7 @@ typedef enum {INTEL_PERF_LOGICAL_COUNTER_UNIT_EU_ATOMIC_REQUESTS_TO_L3_CACHE_LINES ,INTEL_PERF_LOGICAL_COUNTER_UNIT_EU_REQUESTS_TO_L3_CACHE_LINES ,INTEL_PERF_LOGICAL_COUNTER_UNIT_EU_BYTES_PER_L3_CACHE_LINE , +INTEL_PERF_LOGICAL_COUNTER_UNIT_GBPS , INTEL_PERF_LOGICAL_COUNTER_UNIT_MAX } intel_perf_logical_counter_unit_t;Signature · 1 line
2.25.1
PATCH v6 30/36 lib/i915/perf: add a raw timestamp utility
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
Helpful to figure out the bitshifting needed in OA timestamp.
--- lib/i915/perf.c | 21 +++++++++++++++++++++ lib/i915/perf.h | 4 ++++ tools/i915-perf/i915_perf_reader.c | 7 +++++++ 3 files changed, 32 insertions(+)
lib/i915/perf.c +21 -0
@@ -810,3 +810,24 @@ uint64_t intel_perf_read_record_timestamp(const struct intel_perf *perf,return ts; } + +uint64_t intel_perf_read_record_timestamp_raw (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record) +{ +const uint32_t *report32 = (const uint32_t *)(record +1 ); +uint64_t ts; + +switch (metric_set->perf_oa_format) { +case I915_OA_FORMAT_A24u40_A14u32_B8_C8: +case I915_OA_FORMAT_A32u40_A4u32_B8_C8: +case I915_OA_FORMAT_A45_B8_C8 : + ts = report32[1 ]; +break ; + +default : +assert (0 ); + } + +return ts; +}
lib/i915/perf.h +4 -0
@@ -344,6 +344,10 @@ uint64_t intel_perf_read_record_timestamp(const struct intel_perf *perf,const struct intel_perf_metric_set *metric_set, const struct drm_i915_perf_record_header *record); +uint64_t intel_perf_read_record_timestamp_raw (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record); + #ifdef __cplusplus };#endif
tools/i915-perf/i915_perf_reader.c +7 -0
@@ -252,6 +252,13 @@ main(int argc, char *argv[])intel_perf_read_record_timestamp (reader.perf, reader.metric_set, reader.records[reader.n_records - 1])); +fprintf (stdout,"OA raw data timestamp range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +intel_perf_read_record_timestamp_raw (reader.perf, + reader.metric_set, + reader.records[0 ]), +intel_perf_read_record_timestamp_raw (reader.perf, + reader.metric_set, + reader.records[reader.n_records -1 ]));if (strcmp (reader.metric_set_uuid, reader.metric_set->hw_config_guid)) { fprintf(stdout,
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 30/36] lib/i915/perf: add a raw timestamp utility
Quoted 5 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Helpful to figure out the bitshifting needed in OA timestamp.
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 71 lines
---
lib/i915/perf.c | 21 +++++++++++++++++++++ lib/i915/perf.h | 4 ++++ tools/i915-perf/i915_perf_reader.c | 7 +++++++ 3 files changed, 32 insertions(+)lib/i915/perf.c +21
-0@@ -810,3 +810,24 @@ uint64_t intel_perf_read_record_timestamp(const struct intel_perf *perf,return ts; } + +uint64_t intel_perf_read_record_timestamp_raw (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record) +{ +const uint32_t *report32 = (const uint32_t *)(record +1 ); +uint64_t ts; + +switch (metric_set->perf_oa_format) { +case I915_OA_FORMAT_A24u40_A14u32_B8_C8: +case I915_OA_FORMAT_A32u40_A4u32_B8_C8: +case I915_OA_FORMAT_A45_B8_C8 : + ts = report32[1 ]; +break ; + +default : +assert (0 ); + } + +return ts; +}lib/i915/perf.h +4
-0@@ -344,6 +344,10 @@ uint64_t intel_perf_read_record_timestamp(const struct intel_perf *perf,const struct intel_perf_metric_set *metric_set, const struct drm_i915_perf_record_header *record); +uint64_t intel_perf_read_record_timestamp_raw (const struct intel_perf *perf, +const struct intel_perf_metric_set *metric_set, +const struct drm_i915_perf_record_header *record); + #ifdef __cplusplus };#endif tools/i915-perf/i915_perf_reader.c +7
-0@@ -252,6 +252,13 @@ main(int argc, char *argv[])intel_perf_read_record_timestamp (reader.perf, reader.metric_set, reader.records[reader.n_records - 1])); +fprintf (stdout,"OA raw data timestamp range: 0x%016" PRIx64"-0x%016" PRIx64"\n" , +intel_perf_read_record_timestamp_raw (reader.perf, + reader.metric_set, + reader.records[0 ]), +intel_perf_read_record_timestamp_raw (reader.perf, + reader.metric_set, + reader.records[reader.n_records -1 ]));if (strcmp (reader.metric_set_uuid, reader.metric_set->hw_config_guid)) { fprintf(stdout,Signature · 1 line
2.25.1
PATCH v6 31/36 lib/i915/perf: add helper function to get report reason
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
Useful to inspect reports.
--- lib/i915/perf.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++ lib/i915/perf.h | 3 +++ lib/meson.build | 2 +- 3 files changed, 58 insertions(+), 1 deletion(-)
lib/i915/perf.c +54 -0
@@ -831,3 +831,57 @@ uint64_t intel_perf_read_record_timestamp_raw(const struct intel_perf *perf,return ts; } + +const char *intel_perf_read_report_reason (const struct intel_perf *perf, +const struct drm_i915_perf_record_header *record) +{ +const uint32_t *report = (const uint32_t *) (record +1 ); + +/* Not really documented on Gfx7/7.5*/ +if (perf->devinfo.graphics_ver <8 ) +return "timer" ; + +/* Gfx8-11 */ +if (perf->devinfo.graphics_ver <12 ) { +uint32_t reason = report[0 ] >>19 ; +if (reason & (1u <<0 )) +return "timer" ; +if (reason & (1u <<1 )) +return "trigger1" ; +if (reason & (1u <<2 )) +return "trigger2" ; +if (reason & (1u <<3 )) +return "context-switch" ; +if (reason & (1u <<4 )) +return "go-transition" ; + +if (perf->devinfo.graphics_ver >=9 && + reason & (1u <<5 )) +return "clock-ratio-change" ; + +return "unknown" ; + } + +/* Gfx12 */ +if (perf->devinfo.graphics_ver <=12 ) { +uint32_t reason = report[0 ] >>19 ; +if (reason & (1u <<0 )) +return "timer" ; +if (reason & (1u <<1 )) +return "trigger1" ; +if (reason & (1u <<2 )) +return "trigger2" ; +if (reason & (1u <<3 )) +return "context-switch" ; +if (reason & (1u <<4 )) +return "go-transition" ; +if (reason & (1u <<5 )) +return "clock-ratio-change" ; +if (reason & (1u <<6 )) +return "mmio-trigger" ; + +return "unknown" ; + } + +return "unknown" ; +}
lib/i915/perf.h +3 -0
@@ -348,6 +348,9 @@ uint64_t intel_perf_read_record_timestamp_raw(const struct intel_perf *perf,const struct intel_perf_metric_set *metric_set, const struct drm_i915_perf_record_header *record); +const char *intel_perf_read_report_reason (const struct intel_perf *perf, +const struct drm_i915_perf_record_header *record); + #ifdef __cplusplus };#endif
lib/meson.build +1 -1
@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.5.0') +pkgconf.set('i915_perf_version', '1.5.1') configure_file( input : 'i915-perf.pc.in',
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 31/36] lib/i915/perf: add helper function to get report reason
Quoted 5 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Useful to inspect reports.
Signed-off-by Lionel Landwerlin <[email protected]>
Quoted 98 lines
---
lib/i915/perf.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++ lib/i915/perf.h | 3 +++ lib/meson.build | 2 +- 3 files changed, 58 insertions(+), 1 deletion(-)lib/i915/perf.c +54
-0@@ -831,3 +831,57 @@ uint64_t intel_perf_read_record_timestamp_raw(const struct intel_perf *perf,return ts; } + +const char *intel_perf_read_report_reason (const struct intel_perf *perf, +const struct drm_i915_perf_record_header *record) +{ +const uint32_t *report = (const uint32_t *) (record +1 ); + +/* Not really documented on Gfx7/7.5*/ +if (perf->devinfo.graphics_ver <8 ) +return "timer" ; + +/* Gfx8-11 */ +if (perf->devinfo.graphics_ver <12 ) { +uint32_t reason = report[0 ] >>19 ; +if (reason & (1u <<0 )) +return "timer" ; +if (reason & (1u <<1 )) +return "trigger1" ; +if (reason & (1u <<2 )) +return "trigger2" ; +if (reason & (1u <<3 )) +return "context-switch" ; +if (reason & (1u <<4 )) +return "go-transition" ; + +if (perf->devinfo.graphics_ver >=9 && + reason & (1u <<5 )) +return "clock-ratio-change" ; + +return "unknown" ; + } + +/* Gfx12 */ +if (perf->devinfo.graphics_ver <=12 ) { +uint32_t reason = report[0 ] >>19 ; +if (reason & (1u <<0 )) +return "timer" ; +if (reason & (1u <<1 )) +return "trigger1" ; +if (reason & (1u <<2 )) +return "trigger2" ; +if (reason & (1u <<3 )) +return "context-switch" ; +if (reason & (1u <<4 )) +return "go-transition" ; +if (reason & (1u <<5 )) +return "clock-ratio-change" ; +if (reason & (1u <<6 )) +return "mmio-trigger" ; + +return "unknown" ; + } + +return "unknown" ; +}lib/i915/perf.h +3
-0@@ -348,6 +348,9 @@ uint64_t intel_perf_read_record_timestamp_raw(const struct intel_perf *perf,const struct intel_perf_metric_set *metric_set, const struct drm_i915_perf_record_header *record); +const char *intel_perf_read_report_reason (const struct intel_perf *perf, +const struct drm_i915_perf_record_header *record); + #ifdef __cplusplus };#endif lib/meson.build +1
-1@@ -329,7 +329,7 @@ pkgconf.set('prefix', get_option('prefix')) pkgconf.set('exec_prefix', '${prefix}') pkgconf.set('libdir', '${prefix}/@0@'.format(get_option('libdir'))) pkgconf.set('includedir', '${prefix}/@0@'.format(get_option('includedir'))) -pkgconf.set('i915_perf_version', '1.5.0') +pkgconf.set('i915_perf_version', '1.5.1')configure_file(
input : 'i915-perf.pc.in',Signature · 1 line
2.25.1
PATCH v6 32/36 tools/i915-perf: add option to printout reports data
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
This is helpful to get an idea of how timelines of reports should be built.
--- tools/i915-perf/i915_perf_reader.c | 88 ++++++++++++++++++++---------- 1 file changed, 58 insertions(+), 30 deletions(-)
tools/i915-perf/i915_perf_reader.c +58 -30
@@ -53,7 +53,8 @@ usage(void)" --help, -h Print this screen\n" " --counters, -c c1,c2,... List of counters to display values for.\n" " Use 'all' to display all counters.\n" - " Use 'list' to list available counters.\n"); +" Use 'list' to list available counters.\n" + " --reports, -r Print out data per report.\n"); } static struct intel_perf_logical_counter *
@@ -165,12 +166,49 @@ get_logical_counters(struct intel_perf_metric_set *metric_set,return counters; } +static void +print_report_deltas (const struct intel_perf_data_reader *reader, +const struct drm_i915_perf_record_header *i915_report0, +const struct drm_i915_perf_record_header *i915_report1, +struct intel_perf_logical_counter **counters, +uint32_t n_counters) +{ +struct intel_perf_accumulator accu; + +intel_perf_accumulate_reports (&accu, + reader->perf, reader->metric_set, + i915_report0, i915_report1); + +for (uint32_t c =0 ; c < n_counters; c++) { +struct intel_perf_logical_counter *counter = counters[c]; + +switch (counter->storage) { +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT64 : +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT32 : +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_BOOL32 : +fprintf (stdout," %s: %" PRIu64"\n" , + counter->symbol_name, counter->read_uint64 (reader->perf, + reader->metric_set, + accu.deltas)); +break ; +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_DOUBLE : +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_FLOAT : +fprintf (stdout," %s: %f\n" , + counter->symbol_name, counter->read_float (reader->perf, + reader->metric_set, + accu.deltas)); +break ; + } + } +} +int main (int argc,char *argv[]) {const struct option long_options[] = { {"help" , no_argument,0 ,'h' }, {"counters" , required_argument,0 ,'c' }, + {"reports" , no_argument,0 ,'r' }, {0 ,0 ,0 ,0 } };struct intel_perf_data_reader reader;
@@ -179,8 +217,9 @@ main(int argc, char *argv[])const char *counter_names =NULL ;int32_t n_counters;int fd, opt; +bool print_reports = false; -while ((opt =getopt_long (argc, argv,"hc:" , long_options,NULL )) !=-1 ) { +while ((opt =getopt_long (argc, argv,"hc:r" , long_options,NULL )) !=-1 ) {switch (opt) {case 'h' :usage ();
@@ -188,6 +227,9 @@ main(int argc, char *argv[])case 'c' : counter_names = optarg;break ; +case 'r' : + print_reports = true; +break ;default :fprintf (stderr,"Internal error: " "unexpected getopt value: %d\n" , opt);
@@ -268,11 +310,6 @@ main(int argc, char *argv[])for (uint32_t i =0 ; i < reader.n_timelines; i++) {const struct intel_perf_timeline_item *item = &reader.timelines[i]; -const struct drm_i915_perf_record_header *i915_report0 = - reader.records[item->record_start]; -const struct drm_i915_perf_record_header *i915_report1 = - reader.records[item->record_end]; -struct intel_perf_accumulator accu; fprintf(stdout, "Time: CPU=0x%016" PRIx64 "-0x%016" PRIx64 " GPU=0x%016" PRIx64 "-0x%016" PRIx64"\n",
@@ -281,29 +318,20 @@ main(int argc, char *argv[])fprintf (stdout,"hw_id=0x%x %s\n" , item->hw_id, item->hw_id ==0xffffffff ?"(idle)" :"" ); -intel_perf_accumulate_reports (&accu, - reader.perf, reader.metric_set, - i915_report0, i915_report1); - -for (uint32_t c =0 ; c < n_counters; c++) { -struct intel_perf_logical_counter *counter = counters[c]; - -switch (counter->storage) { -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT64 : -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT32 : -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_BOOL32 : -fprintf (stdout," %s: %" PRIu64"\n" , - counter->symbol_name, counter->read_uint64 (reader.perf, - reader.metric_set, - accu.deltas)); -break ; -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_DOUBLE : -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_FLOAT : -fprintf (stdout," %s: %f\n" , - counter->symbol_name, counter->read_float (reader.perf, - reader.metric_set, - accu.deltas)); -break ; +print_report_deltas (&reader, + reader.records[item->record_start], + reader.records[item->record_end], + counters, n_counters); + +if (print_reports) { +for (uint32_t r = item->record_start; r < item->record_end; r++) { +fprintf (stdout," report%i = %s\n" , + r - item->record_start, +intel_perf_read_report_reason (reader.perf, reader.records[r])); +print_report_deltas (&reader, + reader.records[r], + reader.records[r +1 ], + counters, n_counters); } } }
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 32/36] tools/i915-perf: add option to printout reports data
Quoted 1 line · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Quoted 153 lines
This is helpful to get an idea of how timelines of reports should be built.
Signed-off-by Lionel Landwerlin <[email protected]>---
tools/i915-perf/i915_perf_reader.c| 88 ++++++++++++++++++++----------1 file changed, 58 insertions(+), 30 deletions(-)tools/i915-perf/i915_perf_reader.c +2
-1@@ -53,7 +53,8 @@ usage(void)" --help, -h Print this screen\n" " --counters, -c c1,c2,... List of counters to display values for.\n" " Use 'all' to display all counters.\n" - " Use 'list' to list available counters.\n"); +" Use 'list' to list available counters.\n" + " --reports, -r Print out data per report.\n"); }static struct
intel_perf_logical_counter*tools/i915-perf/i915_perf_reader.c +36
-0@@ -165,12 +166,49 @@ get_logical_counters(struct intel_perf_metric_set *metric_set,return counters; } +static void +print_report_deltas (const struct intel_perf_data_reader *reader, +const struct drm_i915_perf_record_header *i915_report0, +const struct drm_i915_perf_record_header *i915_report1, +struct intel_perf_logical_counter **counters, +uint32_t n_counters) +{ +struct intel_perf_accumulator accu; + +intel_perf_accumulate_reports (&accu, + reader->perf, reader->metric_set, + i915_report0, i915_report1); + +for (uint32_t c =0 ; c < n_counters; c++) { +struct intel_perf_logical_counter *counter = counters[c]; + +switch (counter->storage) { +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT64 : +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT32 : +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_BOOL32 : +fprintf (stdout," %s: %" PRIu64"\n" , + counter->symbol_name, counter->read_uint64 (reader->perf, + reader->metric_set, + accu.deltas)); +break ; +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_DOUBLE : +case INTEL_PERF_LOGICAL_COUNTER_STORAGE_FLOAT : +fprintf (stdout," %s: %f\n" , + counter->symbol_name, counter->read_float (reader->perf, + reader->metric_set, + accu.deltas)); +break ; + } + } +} +int
main(int argc, char *argv[]){ const struct option long_options[] = { {"help", no_argument, 0, 'h'}, {"counters", required_argument, 0, 'c'},tools/i915-perf/i915_perf_reader.c +1
-0+ {"reports", no_argument, 0, 'r'},{0, 0, 0, 0} }; struct intel_perf_data_reader reader;tools/i915-perf/i915_perf_reader.c +19
-29@@ -179,8 +217,9 @@ main(int argc, char *argv[])const char *counter_names =NULL ;int32_t n_counters;int fd, opt; +bool print_reports = false; -while ((opt =getopt_long (argc, argv,"hc:" , long_options,NULL )) !=-1 ) { +while ((opt =getopt_long (argc, argv,"hc:r" , long_options,NULL )) !=-1 ) {switch (opt) {case 'h' :usage ();@@ -188,6 +227,9 @@ main(int argc, char *argv[])case 'c' : counter_names = optarg;break ; +case 'r' : + print_reports = true; +break ;default :fprintf (stderr,"Internal error: " "unexpected getopt value: %d\n" , opt);@@ -268,11 +310,6 @@ main(int argc, char *argv[])for (uint32_t i =0 ; i < reader.n_timelines; i++) {const struct intel_perf_timeline_item *item = &reader.timelines[i]; -const struct drm_i915_perf_record_header *i915_report0 = - reader.records[item->record_start]; -const struct drm_i915_perf_record_header *i915_report1 = - reader.records[item->record_end]; -struct intel_perf_accumulator accu; fprintf(stdout, "Time: CPU=0x%016" PRIx64 "-0x%016" PRIx64 " GPU=0x%016" PRIx64 "-0x%016" PRIx64"\n",@@ -281,29 +318,20 @@ main(int argc, char *argv[])fprintf (stdout,"hw_id=0x%x %s\n" , item->hw_id, item->hw_id ==0xffffffff ?"(idle)" :"" ); -intel_perf_accumulate_reports (&accu, - reader.perf, reader.metric_set, - i915_report0, i915_report1); - -for (uint32_t c =0 ; c < n_counters; c++) { -struct intel_perf_logical_counter *counter = counters[c]; - -switch (counter->storage) { -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT64 : -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_UINT32 : -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_BOOL32 : -fprintf (stdout," %s: %" PRIu64"\n" , - counter->symbol_name, counter->read_uint64 (reader.perf, - reader.metric_set, - accu.deltas)); -break ; -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_DOUBLE : -case INTEL_PERF_LOGICAL_COUNTER_STORAGE_FLOAT : -fprintf (stdout," %s: %f\n" , - counter->symbol_name, counter->read_float (reader.perf, - reader.metric_set, - accu.deltas)); -break ; +print_report_deltas (&reader, + reader.records[item->record_start], + reader.records[item->record_end], + counters, n_counters); + +if (print_reports) { +for (uint32_t r = item->record_start; r < item->record_end; r++) { +fprintf (stdout," report%i = %s\n" , + r - item->record_start, +intel_perf_read_report_reason (reader.perf, reader.records[r])); +print_report_deltas (&reader, + reader.records[r], + reader.records[r +1 ], + counters, n_counters); } } }Signature · 1 line
2.25.1
PATCH v6 33/36 lib/i915: prepare codegen for new ACM/DG2 variables
In reply to Umesh Nerlige RamappaFrom: Lionel Landwerlin <[email protected]>
--- lib/i915/perf-configs/codegen.py | 84 +++++++++++++++---- .../perf-configs/perf-metricset-codegen.py | 13 +++ 2 files changed, 81 insertions(+), 16 deletions(-)
lib/i915/perf-configs/codegen.py +68 -16
@@ -1,3 +1,4 @@ +import re import xml.etree.ElementTree as et class Codegen:
@@ -123,6 +124,51 @@ class Set: return self.xml.find(path) +hw_vars_mapping = { + "$EuCoresTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, + "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, + "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesSlice0123Count": { 'c': "perf->devinfo.n_eu_sub_slices_half_slices" }, + "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, + + "$VectorEngineTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, + "$VectorEnginePerXeCoreCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$VectorEngineThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, + + "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, + "$SliceTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, + + "$SubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + + "$GtSliceMask": { 'c': "perf->devinfo.slice_mask" }, + "$GtSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + "$GtDualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + + "$GtXeCoreMask": { 'c': "perf->devinfo.slice_mask" }, + "$XeCoreMask": { 'c': "perf->devinfo.slice_mask" }, + "$XeCoreTotalCount": { 'c': 'perf->devinfo.n_eu_sub_slices' }, + + "$GpuTimestampFrequency": { 'c': "perf->devinfo.timestamp_frequency" }, + "$GpuMinFrequency": { 'c': "perf->devinfo.gt_min_freq" }, + "$GpuMaxFrequency": { 'c': "perf->devinfo.gt_max_freq" }, + "$SkuRevisionId": { 'c': "perf->devinfo.revision" }, + "$QueryMode": { 'c': "perf->devinfo.query_mode" }, +} + +def is_hw_var(name): + m = re.search('\$GtSlice([0-9]+)XeCore([0-9]+)$', name) + if m: + return True + m = re.search('\$GtSlice([0-9]+)$', name) + if m: + return True + m = re.search('\$GtSlice([0-9]+)DualSubslice([0-9]+)$', name) + if m: + return True + return name in hw_vars_mapping + class Gen: def __init__(self, filename, c): self.filename = filename
@@ -163,23 +209,11 @@ class Gen: self.exp_ops["ULTE"] = (2, self.splice_ulte) self.exp_ops["ULT"] = (2, self.splice_ult) self.exp_ops["&&"] = (2, self.splice_logical_and) + self.exp_ops["<<"] = (2, self.splice_lshft) + self.exp_ops[">>"] = (2, self.splice_rshft) + self.exp_ops["UMUL"] = (2, self.splice_uml) - self.hw_vars = { - "$EuCoresTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, - "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, - "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, - "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, - "$EuDualSubslicesSlice0123Count": { 'c': "perf->devinfo.n_eu_sub_slices_half_slices" }, - "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, - "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, - "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, - "$SubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, - "$GpuTimestampFrequency": { 'c': "perf->devinfo.timestamp_frequency" }, - "$GpuMinFrequency": { 'c': "perf->devinfo.gt_min_freq" }, - "$GpuMaxFrequency": { 'c': "perf->devinfo.gt_max_freq" }, - "$SkuRevisionId": { 'c': "perf->devinfo.revision" }, - "$QueryMode": { 'c': "perf->devinfo.query_mode" }, - } + self.hw_vars = hw_vars_mapping def emit_fadd(self, tmp_id, args): self.c("double tmp{0} = {1} + {2};".format(tmp_id, args[1], args[0]))
@@ -286,11 +320,29 @@ class Gen: def splice_ugt(self, args): return self.brkt(args[1]) + " > " + self.brkt(args[0]) + def splice_lshft(self, args): + return '(' + self.brkt(args[1]) + " << " + self.brkt(args[0]) + ')' + + def splice_rshft(self, args): + return '(' + self.brkt(args[1]) + " >> " + self.brkt(args[0]) + ')' + + def splice_uml(self, args): + return self.brkt(args[1]) + " * " + self.brkt(args[0]) + def resolve_variable(self, name, set): if name in self.hw_vars: return self.hw_vars[name]['c'] if name in set.counter_vars: return set.read_funcs[name] + "(perf, metric_set, accumulator)" + m = re.search('\$GtSlice([0-9]+)$', name) + if m: + return 'intel_perf_devinfo_slice_available(&perf->devinfo, {0})'.format(m.group(1)) + m = re.search('\$GtSlice([0-9]+)DualSubslice([0-9]+)$', name) + if m: + return 'intel_perf_devinfo_subslice_available(&perf->devinfo, {0}, {1})'.format(m.group(1), m.group(2)) + m = re.search('\$GtSlice([0-9]+)XeCore([0-9]+)$', name) + if m: + return 'intel_perf_devinfo_subslice_available(&perf->devinfo, {0}, {1})'.format(m.group(1), m.group(2)) return None def output_rpn_equation_code(self, set, counter, equation):
lib/i915/perf-configs/perf-metricset-codegen.py +13 -0
@@ -126,6 +126,19 @@ def generate_metric_sets(args, gen): metric_set->c_offset = metric_set->b_offset + 8; metric_set->perfcnt_offset = metric_set->c_offset + 8; + """)) + elif gen.chipset.startswith("acm"): + c(textwrap.dedent("""\ + metric_set->perf_oa_format = I915_OA_FORMAT_A24u40_A14u32_B8_C8; + + metric_set->perf_raw_size = 256; + metric_set->gpu_time_offset = 0; + metric_set->gpu_clock_offset = 1; + metric_set->a_offset = 2; + metric_set->b_offset = metric_set->a_offset + 38; + metric_set->c_offset = metric_set->b_offset + 8; + metric_set->perfcnt_offset = metric_set->c_offset + 8; + """)) else: c(textwrap.dedent("""\
Signature · 1 line
2.25.1
Re: [igt-dev] [PATCH i-g-t v6 33/36] lib/i915: prepare codegen for new ACM/DG2 variables
Quoted 3 lines · Umesh Nerlige Ramappa
From: Lionel Landwerlin <[email protected]>
Signed-off-by Lionel Landwerlin <[email protected]>
lgtm,
Quoted 151 lines
---
lib/i915/perf-configs/codegen.py| 84 +++++++++++++++----.../perf-configs/perf-metricset-codegen.py| 13 +++2 files changed, 81 insertions(+), 16 deletions(-)lib/i915/perf-configs/codegen.py +1
-0@@ -1,3 +1,4 @@ +import reimport xml.etree.ElementTree as et
class Codegen:
lib/i915/perf-configs/codegen.py +45
-0@@ -123,6 +124,51 @@ class Set: return self.xml.find(path) +hw_vars_mapping = { + "$EuCoresTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, + "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, + "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$EuDualSubslicesSlice0123Count": { 'c': "perf->devinfo.n_eu_sub_slices_half_slices" }, + "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, + + "$VectorEngineTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, + "$VectorEnginePerXeCoreCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, + "$VectorEngineThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, + + "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, + "$SliceTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, + + "$SubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + + "$GtSliceMask": { 'c': "perf->devinfo.slice_mask" }, + "$GtSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + "$GtDualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, + + "$GtXeCoreMask": { 'c': "perf->devinfo.slice_mask" }, + "$XeCoreMask": { 'c': "perf->devinfo.slice_mask" }, + "$XeCoreTotalCount": { 'c': 'perf->devinfo.n_eu_sub_slices' }, + + "$GpuTimestampFrequency": { 'c': "perf->devinfo.timestamp_frequency" }, + "$GpuMinFrequency": { 'c': "perf->devinfo.gt_min_freq" }, + "$GpuMaxFrequency": { 'c': "perf->devinfo.gt_max_freq" }, + "$SkuRevisionId": { 'c': "perf->devinfo.revision" }, + "$QueryMode": { 'c': "perf->devinfo.query_mode" }, +} + +def is_hw_var(name): + m = re.search('\$GtSlice([0-9]+)XeCore([0-9]+)$', name) + if m: + return True + m = re.search('\$GtSlice([0-9]+)$', name) + if m: + return True + m = re.search('\$GtSlice([0-9]+)DualSubslice([0-9]+)$', name) + if m: + return True + return name in hw_vars_mapping +class Gen:
def__init__(self, filename, c):
self.filename = filenamelib/i915/perf-configs/codegen.py +22
-16@@ -163,23 +209,11 @@ class Gen: self.exp_ops["ULTE"] = (2, self.splice_ulte) self.exp_ops["ULT"] = (2, self.splice_ult) self.exp_ops["&&"] = (2, self.splice_logical_and) + self.exp_ops["<<"] = (2, self.splice_lshft) + self.exp_ops[">>"] = (2, self.splice_rshft) + self.exp_ops["UMUL"] = (2, self.splice_uml) - self.hw_vars = { - "$EuCoresTotalCount": { 'c': "perf->devinfo.n_eus", 'desc': "The total number of execution units" }, - "$EuSlicesTotalCount": { 'c': "perf->devinfo.n_eu_slices" }, - "$EuSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, - "$EuDualSubslicesTotalCount": { 'c': "perf->devinfo.n_eu_sub_slices" }, - "$EuDualSubslicesSlice0123Count": { 'c': "perf->devinfo.n_eu_sub_slices_half_slices" }, - "$EuThreadsCount": { 'c': "perf->devinfo.eu_threads_count" }, - "$SliceMask": { 'c': "perf->devinfo.slice_mask" }, - "$DualSubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, - "$SubsliceMask": { 'c': "perf->devinfo.subslice_mask" }, - "$GpuTimestampFrequency": { 'c': "perf->devinfo.timestamp_frequency" }, - "$GpuMinFrequency": { 'c': "perf->devinfo.gt_min_freq" }, - "$GpuMaxFrequency": { 'c': "perf->devinfo.gt_max_freq" }, - "$SkuRevisionId": { 'c': "perf->devinfo.revision" }, - "$QueryMode": { 'c': "perf->devinfo.query_mode" }, - } + self.hw_vars = hw_vars_mapping def emit_fadd(self, tmp_id, args): self.c("double tmp{0} = {1} + {2};".format(tmp_id, args[1], args[0]))@@ -286,11 +320,29 @@ class Gen: def splice_ugt(self, args): return self.brkt(args[1]) + " > " + self.brkt(args[0]) + def splice_lshft(self, args): + return '(' + self.brkt(args[1]) + " << " + self.brkt(args[0]) + ')' + + def splice_rshft(self, args): + return '(' + self.brkt(args[1]) + " >> " + self.brkt(args[0]) + ')' + + def splice_uml(self, args): + return self.brkt(args[1]) + " * " + self.brkt(args[0]) + def resolve_variable(self, name, set): if name in self.hw_vars: return self.hw_vars[name]['c'] if name in set.counter_vars: return set.read_funcs[name] + "(perf, metric_set, accumulator)" + m = re.search('\$GtSlice([0-9]+)$', name) + if m: + return 'intel_perf_devinfo_slice_available(&perf->devinfo, {0})'.format(m.group(1)) + m = re.search('\$GtSlice([0-9]+)DualSubslice([0-9]+)$', name) + if m: + return 'intel_perf_devinfo_subslice_available(&perf->devinfo, {0}, {1})'.format(m.group(1), m.group(2)) + m = re.search('\$GtSlice([0-9]+)XeCore([0-9]+)$', name) + if m: + return 'intel_perf_devinfo_subslice_available(&perf->devinfo, {0}, {1})'.format(m.group(1), m.group(2)) return None def output_rpn_equation_code(self, set, counter, equation):lib/i915/perf-configs/perf-metricset-codegen.py +13
-0@@ -126,6 +126,19 @@ def generate_metric_sets(args, gen): metric_set->c_offset = metric_set->b_offset + 8; metric_set->perfcnt_offset = metric_set->c_offset + 8; + """)) + elif gen.chipset.startswith("acm"): + c(textwrap.dedent("""\ + metric_set->perf_oa_format = I915_OA_FORMAT_A24u40_A14u32_B8_C8; + + metric_set->perf_raw_size = 256; + metric_set->gpu_time_offset = 0; + metric_set->gpu_clock_offset = 1; + metric_set->a_offset = 2; + metric_set->b_offset = metric_set->a_offset + 38; + metric_set->c_offset = metric_set->b_offset + 8; + metric_set->perfcnt_offset = metric_set->c_offset + 8; + """)) else: c(textwrap.dedent("""\Signature · 1 line
2.25.1
PATCH v6 34/36 lib/i915/perf: Add ACM GT1 metrics
In reply to Umesh Nerlige RamappaAdd ACM GT1 metrics XML files for perf OA tests.
--- lib/i915/perf-configs/guids.xml | 345 + lib/i915/perf-configs/oa-acmgt1.xml | 76252 ++++++++++++++++++++++++++ lib/i915/perf.c | 35 + lib/meson.build | 1 + 4 files changed, 76633 insertions(+) create mode 100644 lib/i915/perf-configs/oa-acmgt1.xml
lib/i915/perf-configs/guids.xml +345 -0
@@ -431,4 +431,349 @@ <guid config_hash="b420d7b956fb1b097a7bb812b6c87dc0" mdapi_config_hash="d81f15a77d18a099782c18acaa3132f7" id="c3ea2a23-f1c7-4a19-9da2-b569226fb6de" chipset="adl" name="EuActivity7" /> <guid config_hash="5face026dac6a0549f8a28cf226dfbc8" mdapi_config_hash="de7326838bcc8abf024bac9c771e22e4" id="a5e2f79b-cecb-4eff-8f29-cda8e2a58749" chipset="adl" name="EuActivity8" /> <guid config_hash="1affe7abeb3739d2c845cc6fa9aacf88" mdapi_config_hash="d071f4933574945979595246be61d3f0" id="89173c19-fcfe-48da-ac26-fb64425f141b" chipset="adl" name="TestOa" /> + <guid config_hash="03581feed70897846839305bef294970" mdapi_config_hash="580f399d67c355cd996df94bf6592ff1" id="f5b8f05e-c84c-4f1c-bb05-68fbea73879b" chipset="acmgt1" name="RenderBasic" /> + <guid config_hash="1909915afc07af2faffd448b12b5fa70" mdapi_config_hash="a1dbe33e0f731b73942f4d7a11885780" id="53ee6271-905c-4b2a-b449-9117ffd10a48" chipset="acmgt1" name="ComputeBasic" /> + <guid config_hash="a60e5cb58af7c7fe4c772ad3c0d723d1" mdapi_config_hash="dfebdf7eb743a99cda700c3211311acc" id="952f5f55-449d-43db-87d7-6113e9c3b89a" chipset="acmgt1" name="ThreadsAndRast1" /> + <guid config_hash="256875afc20fecd1dfb75e193b9a2a74" mdapi_config_hash="e27a13f37cffd8e650000197adc14a65" id="797ae4a3-6d36-4be1-9064-a9bc3263957f" chipset="acmgt1" name="ThreadsAndRast2" /> + <guid config_hash="749c3bc9b8f01d8f89248610e4ce0d63" mdapi_config_hash="67783e424eb168c03b3e1dc1e3a5a152" id="abb4b8fd-b9cd-4b11-a610-1c0707af14cf" chipset="acmgt1" name="ThreadsAndRast3" /> + <guid config_hash="acff235c23ea5530d4256044fab9e59d" mdapi_config_hash="e67a2ab3f35019287d15c999d0faedb7" id="2da159d5-8ace-42bb-9d81-26f2b162d8d8" chipset="acmgt1" name="ThreadsAndRast4" /> + <guid config_hash="da49b5c7e497eab30522295f416dd84a" mdapi_config_hash="80f802eb9434e078d1c158be06940d88" id="91273130-42fe-4c5e-ba34-6f9140b612c6" chipset="acmgt1" name="AsyncCompute" /> + <guid config_hash="1a28d05d945714b099a9417181b72962" mdapi_config_hash="0cdf3168ffd88a84c2a42d4c62567189" id="fd9d1ad4-13b3-4583-9f2c-998f627e370d" chipset="acmgt1" name="XveActivity1" /> + <guid config_hash="d86a9b202b07a4443c05054f997e2203" mdapi_config_hash="ce7908d800c405d187a088af61ee7c7b" id="4b785529-8efb-471e-93c5-481dae7b3b8b" chipset="acmgt1" name="GpuBusyness" /> + <guid config_hash="f83112e995826a1135aad56c0648d216" mdapi_config_hash="6bd0c6d87cd9dbe104d9368f73ebb82e" id="8a0c5d31-e0cb-4d2c-8056-f3824c57ad9a" chipset="acmgt1" name="RenderPipeProfile1" /> + <guid config_hash="33bca4b0b4ee572a26c55a4531a1f6dd" mdapi_config_hash="45a93937cef959756b3d318718d9df9d" id="0f40fba0-129a-4a19-87de-5fdbe111b915" chipset="acmgt1" name="RenderPipeProfile2" /> + <guid config_hash="96eb95ce7630b395a44860c40947a1db" mdapi_config_hash="fc4172abe1625c861ba9cd81bdb8f41e" id="24695fb7-a404-4cd6-9aad-d23738b6d499" chipset="acmgt1" name="L3" /> + <guid config_hash="a08acb81d60803a93751565c31854e77" mdapi_config_hash="c4d973b15c80bfbcb661c1ee1d76eb09" id="4486113b-a8a7-4f75-a1c2-7d307fa0e00e" chipset="acmgt1" name="Sampler" /> + <guid config_hash="b7863c3b8519d3ad409eaae2c1521861" mdapi_config_hash="d2097d3ca767e653a616a187b5acd609" id="75fc20f8-8694-47f2-b67b-cbb35f697a8e" chipset="acmgt1" name="HDCAndSF1" /> + <guid config_hash="20152c87392ec9354a3e8e2817740361" mdapi_config_hash="6bea5af3aac1c7dee17b3aaeb7f7758c" id="d5e02a31-fc87-4cdc-8136-4b19e00b7e94" chipset="acmgt1" name="HDCAndSF2" /> + <guid config_hash="66c3c5b54fdf2b4a565f7bfd26b84c63" mdapi_config_hash="69b97529d9b16cefbee90f60ac3b34e6" id="08858210-0dc3-452b-afee-e5c7811c01cb" chipset="acmgt1" name="RasterizerAndPixelBackend1" /> + <guid config_hash="706cf923d603aaefef70ae92e213817a" mdapi_config_hash="e150586afbdcf120d1bcc195a1c8fe5f" id="30b71e0c-55eb-49ed-8863-ee7db68ff1b1" chipset="acmgt1" name="RasterizerAndPixelBackend2" /> + <guid config_hash="5ab62a7b918b70b04699b77a78107f50" mdapi_config_hash="5ecd892e51e757a976cd0cd4cae3f274" id="3620f0df-42c6-4904-b395-5497f5a2ea4b" chipset="acmgt1" name="TDL1" /> + <guid config_hash="dfda4f47ade0b496dd2d500e293537f9" mdapi_config_hash="0adf78de606d7b56bca529d410dd0cd5" id="28357a3a-a53b-4da4-a5b7-395dbfa501f1" chipset="acmgt1" name="TDL2" /> + <guid config_hash="2059b848ec44ec172df992f22113debc" mdapi_config_hash="fc35170d1e41ff4e236f27578534b8fd" id="262cf8cb-824e-4f48-bac4-6be3d8dda6a5" chipset="acmgt1" name="TDL3" /> + <guid config_hash="798658d1cefbb2a0cceef3c524506243" mdapi_config_hash="69970e440aaa413a66d8ec8d349105d5" id="1e629e21-abd2-418e-a184-753977bb529b" chipset="acmgt1" name="TDL4" /> + <guid config_hash="0e12a6e7ae5c8ca70c55df37051f3144" mdapi_config_hash="521df5debce4cf5f78aba4df5d7adc16" id="57f210de-b537-464a-af7b-7dfe2f3780c1" chipset="acmgt1" name="TestOa" /> + <guid config_hash="fa26c66ee01a7fee7926f74df828ae34" mdapi_config_hash="0d796271d2d971b6aed0152e491c14e6" id="37b5b93b-6acf-4726-9b57-1a65fab22bb2" chipset="acmgt1" name="ComputeOverview" /> + <guid config_hash="51cad4f5fea07a4ce99a5bba032b18d2" mdapi_config_hash="07bcc84a37b7b7fe4c00595074abd1c9" id="a2fe4388-0540-4550-b57e-6948cd7ab47a" chipset="acmgt1" name="L1ProfileReads_XeCore0" /> + <guid config_hash="e40e35af9a062176288ab33e751bb3ae" mdapi_config_hash="7c924206ab3846cd6aa3e1a90364e6b1" id="16bc63c3-d365-4bee-baf7-bafbc1aa7c2d" chipset="acmgt1" name="L1ProfileReads_XeCore3" /> + <guid config_hash="827fe956c669c3743bd7d6c17b3e4455" mdapi_config_hash="13df90f63d812fc8aaaf1b109ea0b8c3" id="28da4c17-4629-4111-a575-787a85100db9" chipset="acmgt1" name="L1ProfileSlmBankConflicts_XeCore0" /> + <guid config_hash="e36b0dc82578c8dafa4df9e6fbbb63e0" mdapi_config_hash="1f968ce59957872fcaef21069cee3b64" id="825f82b5-486a-4ca2-96fa-f358871984b0" chipset="acmgt1" name="L1ProfileSlmBankConflicts_XeCore3" /> + <guid config_hash="2d3df306a101c592228f870c7e986005" mdapi_config_hash="031cfdabf0beb78492f0ed2f56f8594f" id="5d58094c-105e-4562-b53d-68185319c794" chipset="acmgt1" name="L1ProfileWrites_XeCore0" /> + <guid config_hash="2210e989a4f69eabe5f2ce3ecd2d06dd" mdapi_config_hash="39cb4e3f7aa692b7f6fb3e5fe4a4261e" id="3ce21cd6-fe3e-468f-ae53-07dbf73c45f9" chipset="acmgt1" name="L1ProfileWrites_XeCore3" /> + <guid config_hash="92c8cdbb2ab634f6ad62d9eea317a29c" mdapi_config_hash="dc698ef9cec93afbea6ec9278fa1d67a" id="54d0193f-f8e0-4b81-99ba-e73a23fd9418" chipset="acmgt1" name="DataportReads" /> + <guid config_hash="583d1255b4b66971940337f23b1ab7ef" mdapi_config_hash="e837db0e46a70327237c2a9ce06c0794" id="17982512-46b0-4b58-9e8d-d59b5bf8b3f5" chipset="acmgt1" name="DataportWrites" /> + <guid config_hash="1095da332ad471328e7d24fb9c791970" mdapi_config_hash="5412924efbb3e88e10cd2f5507123d38" id="480895b0-8013-4be1-8bbf-fc3f1ad2e97f" chipset="acmgt1" name="Ext1" /> + <guid config_hash="2596c836815c65c7d2a0be12e227d0fe" mdapi_config_hash="47bdd89d7224117f7027dc0f1c57dbc6" id="b899d843-d8ce-4803-8174-f2bf1107de00" chipset="acmgt1" name="Ext2" /> + <guid config_hash="fae7ab4aabc5813bbb455662395519bf" mdapi_config_hash="f284fb226d661979e3efe04b6139a43b" id="59e0ed29-cd2d-43ee-bf49-e4637e8ef9ee" chipset="acmgt1" name="Ext3" /> + <guid config_hash="7f3a057a2c7a2195a548706eb70cd06d" mdapi_config_hash="187da7ed3c88699d277fd80456dba4b0" id="475c650e-1045-45dc-a280-3365f7af8f8c" chipset="acmgt1" name="Ext4" /> + <guid config_hash="06f122a3183d4bb5dcc88742d13461ce" mdapi_config_hash="ddda913231390dc12f81aaf19d56f362" id="b7d7a874-19d8-48e6-bf54-f859a994b0a4" chipset="acmgt1" name="Ext5" /> + <guid config_hash="6676f55334cf33b7bcf9b0a6fcb0e0fc" mdapi_config_hash="96031d1b1d16608c499f8f02cb8f90cf" id="f1a94d28-672b-41a6-9ab9-e2040200d28a" chipset="acmgt1" name="Ext6" /> + <guid config_hash="24824a663575f4a31fffc1f9a128c15d" mdapi_config_hash="f16d5878aab4b516bb083dc0c23e6ddd" id="31d119ed-a150-4b52-9206-3f9832e275ea" chipset="acmgt1" name="Ext7" /> + <guid config_hash="a609fc1dd0f960c0c9982d108c8f08f3" mdapi_config_hash="547552a736af00217cc74423a283f6f8" id="4eb14f16-d2ac-46d3-b974-e413430fddec" chipset="acmgt1" name="Ext8" /> + <guid config_hash="1fa6c42a02c55de9d3303b85d4d645ed" mdapi_config_hash="caf913a6e7a4526cea3edbf3d7feb102" id="b875dedd-f121-4f89-b02f-ff2c873e2543" chipset="acmgt1" name="Ext9" /> + <guid config_hash="06bc9a6e909f6fde296f8c6e1a23e3b9" mdapi_config_hash="48d1f44984f040e58c8aca02f5dfebec" id="691b6a0d-d3d7-4b5e-a70e-d8232722cddf" chipset="acmgt1" name="Ext10" /> + <guid config_hash="8769662891c52753c4b030a98c7f0863" mdapi_config_hash="42bb12d42d843a329652e01fa54be6eb" id="3c626d93-4368-4115-a21d-8fd8cd10e8fa" chipset="acmgt1" name="Ext11" /> + <guid config_hash="96eb95ce7630b395a44860c40947a1db" mdapi_config_hash="fc4172abe1625c861ba9cd81bdb8f41e" id="4678846c-99cd-4a3f-acb4-8f055f48da56" chipset="acmgt1" name="Ext12" /> + <guid config_hash="98df80f68ee1bcadbdc63c12b4f8c2c8" mdapi_config_hash="2fb63f5ac3ae58063304652f77c876c5" id="04d1f1c5-17e5-4104-b5a1-d62337b681b9" chipset="acmgt1" name="Ext13" /> + <guid config_hash="b82c25b890c976604baf594ae84f9c62" mdapi_config_hash="09d629c6e33a79a9b8ade7d925132968" id="9312e21f-5d97-4058-a2fc-c7497c18f0a1" chipset="acmgt1" name="Ext14" /> + <guid config_hash="2dfeb8a292a30913ee66936de3d649ca" mdapi_config_hash="6d652105301aa9166c75106f193ac9e7" id="fd355e93-4edf-425c-99e6-8693a22ae3cd" chipset="acmgt1" name="Ext15" /> + <guid config_hash="b6eee444b5af376bdee240747ce82c01" mdapi_config_hash="fc869df2123540ca67502b8177addf14" id="844efb3d-c453-4dd2-b556-7bac8e111b46" chipset="acmgt1" name="Ext16" /> + <guid config_hash="217beaaaaead570019303250a6019427" mdapi_config_hash="eeaa331a6d07320fedcbf0d6fcdf163b" id="fadda67d-1b94-40d1-be89-759f78641ab2" chipset="acmgt1" name="Ext17" /> + <guid config_hash="a16442d7bd977878f20311cf3d23ecfd" mdapi_config_hash="cb53931cc9c09815599e095ab3220fd5" id="bdd021fa-a841-4f92-826f-c3ae681c68e6" chipset="acmgt1" name="Ext18" /> + <guid config_hash="6de41a724e88178fe7b0b6292e3ee82e" mdapi_config_hash="fde1cd12a41fb58e8f4f8a0f50dfd64c" id="1181d35b-bf9a-4640-82b8-e31e031ea8d3" chipset="acmgt1" name="Ext19" /> + <guid config_hash="aea935cb55c9d33934dcd20816cd2622" mdapi_config_hash="4247bf9bf066566c01d932c98533b5e8" id="aa39f75d-0a8a-43d1-9c51-bcd8c62cdfe6" chipset="acmgt1" name="Ext20" /> + <guid config_hash="31e1878ef818b7210297a2844ba3cfdf" mdapi_config_hash="649675b01a26c17a0fb3d31350eaa4e5" id="ae58d1ad-8aed-4104-a8e2-443d7e800378" chipset="acmgt1" name="Ext21" /> + <guid config_hash="1a8aca34497e1017c16aede6c0413fc3" mdapi_config_hash="7f1a3badabd8a173acf0c8b48cab030d" id="c27ebd95-f242-4f4b-969a-9850f1081c30" chipset="acmgt1" name="Ext22" /> + <guid config_hash="d52260dc03030dbbe1a61094843f8837" mdapi_config_hash="6083cc80198df201c7250466a1e08b04" id="881a8616-b63f-4bb1-970a-73b40b762434" chipset="acmgt1" name="Ext23" /> + <guid config_hash="68d2055233fb6df4092862c8883f1b66" mdapi_config_hash="091894211e9ec6634527f9b90eb07cfb" id="9fe09ad3-d296-4d41-ab19-270431f7ab29" chipset="acmgt1" name="Ext24" /> + <guid config_hash="1088f1d5ebfe937878b6dda5edd59423" mdapi_config_hash="47fdb93257e96d1bc897e5155253cd39" id="e2016425-8755-4ca0-bc93-354ba7a6ca7a" chipset="acmgt1" name="Ext25" /> + <guid config_hash="d71cd0d8288d1fbfea844f95c773437c" mdapi_config_hash="c3b269cff5fa487c395637ebafba0349" id="9fc2d1f6-7504-41f2-9f4b-96ff3307f6cb" chipset="acmgt1" name="Ext26" /> + <guid config_hash="7ca4c48540372dfb3dfe582d24c71c76" mdapi_config_hash="32ab1bf15096f632e2bc5bebbb6883de" id="7bc1c162-2b4b-4cb8-b351-65be069d3f10" chipset="acmgt1" name="Ext27" /> + <guid config_hash="7b615b3384df6fe083ca50b10170e7b7" mdapi_config_hash="479eb5015dac41432a2df2c0a60d0d24" id="afdf809e-962c-4af5-8e08-cd55b455dee9" chipset="acmgt1" name="Ext28" /> + <guid config_hash="38b59093074ec3998cbdceabaceb661e" mdapi_config_hash="4122f0759ac7ca14f56dde50f691f080" id="4f22c4f0-68ce-46ad-91a7-272b517bfa2e" chipset="acmgt1" name="Ext29" /> + <guid config_hash="500fdd072cd83b69c087b917b212c106" mdapi_config_hash="033f37a2db3d39b76f5f25618620a791" id="376f6937-f69d-41b0-a1ae-402dee9d42cb" chipset="acmgt1" name="Ext30" /> + <guid config_hash="00523c71250bb077c645041bbba13faf" mdapi_config_hash="273454896b03cac032e94be61447d1b2" id="28e81769-8b4e-4a6f-a658-19901c71f724" chipset="acmgt1" name="Ext31" /> + <guid config_hash="a363178889e787c8efe4f53666e9d2c3" mdapi_config_hash="aa3f66e56575c46fd7fdcbfade4db5b8" id="0efcc5a6-bbc5-4af3-bf07-db3597f314ab" chipset="acmgt1" name="Ext32" /> + <guid config_hash="cf7e0869f4e6dcf61cb448c1dd6245ca" mdapi_config_hash="3bd004f89ef56fb2d6264ae6df8555d7" id="038bcd96-55da-41c0-b9ad-6d71e943f1b9" chipset="acmgt1" name="Ext33" /> + <guid config_hash="bbb1e565cd4bc7eb56cf1950f9eb1920" mdapi_config_hash="3b0f2a3fa2d9556a317d43cd3392f617" id="48d0ac9b-3e47-43c9-9779-3e2cdf315635" chipset="acmgt1" name="Ext34" /> + <guid config_hash="b5113330470868e8ce163c0044c181aa" mdapi_config_hash="c168ea89ec8f771ee652138027b0f55b" id="ebe68cc1-ddda-412d-b9de-4616c4421712" chipset="acmgt1" name="Ext35" /> + <guid config_hash="86fa67e7bbb608382fb68932ea7e5b15" mdapi_config_hash="5e5984fa2c63db8cc0f6754f0095ff77" id="b305a4ce-2466-4172-97f1-a09a874f377c" chipset="acmgt1" name="Ext36" /> + <guid config_hash="3b9805da69dd4c732e5ff34b5cdf04fa" mdapi_config_hash="82081df90adf3609bc82bd7d39efcf36" id="8184361a-7e98-4f4d-b047-42c767d4c863" chipset="acmgt1" name="Ext37" /> + <guid config_hash="8bff66c1ee65e2f1333d9691a8d4d4d1" mdapi_config_hash="77bebdc4e3236605767fac9d76222cb1" id="0298d65d-a984-4f19-b968-1e65f072189d" chipset="acmgt1" name="Ext38" /> + <guid config_hash="39a24fbf6c40f5bf46b58ef936dca0e7" mdapi_config_hash="0d09039ba9745a61ac91bef176ebc4df" id="602b3d8e-9299-4615-9739-fe44dd9b7102" chipset="acmgt1" name="Ext39" /> + <guid config_hash="8c71d34d2a9ccbdea1a44a9194adce66" mdapi_config_hash="b2f4e2f74dfa5922197b5933ddd19179" id="7595f984-7cd3-4b1c-8332-bacfd2b3a359" chipset="acmgt1" name="Ext40" /> + <guid config_hash="ef0912210dbbd1ab995c0b5474fed41b" mdapi_config_hash="fef9ebb4cade91da7cdabe595b6e4820" id="61bf91ad-0dfa-421c-b503-0eed63b64ef5" chipset="acmgt1" name="Ext41" /> + <guid config_hash="399cd8c94c8266ccfea473b16e84464c" mdapi_config_hash="20dd2965f3c6287abd81b699a86d0db8" id="a2f3417e-fbe1-4c26-a20c-509c7205445f" chipset="acmgt1" name="Ext42" /> + <guid config_hash="317a16ed5d58d97e9254827eaba93660" mdapi_config_hash="3e50a4a360a311eaefacdc521d3d1136" id="343460a0-c54f-4fe5-b7ac-6d9870a56aaf" chipset="acmgt1" name="Ext43" /> + <guid config_hash="a5b3e2556c23e85d9e0ef2dc91ba4a96" mdapi_config_hash="05d582e1eceacaa854aa7e1ba5078d0a" id="cfaeaf8e-4888-459f-8019-c3080595c528" chipset="acmgt1" name="Ext44" /> + <guid config_hash="7b4384392adfb001c80f12e712f5710c" mdapi_config_hash="e733e238bdd06e217753f452d5eafc19" id="c50cb371-e940-4499-aa74-4e75bc131329" chipset="acmgt1" name="Ext45" /> + <guid config_hash="aa9887dedacf803b0dbcc3e1dccb6fb1" mdapi_config_hash="13ae8a7e98206164bed268a592ef48a7" id="b56a3837-111d-4ec1-8d3f-42e9e304d8f5" chipset="acmgt1" name="Ext46" /> + <guid config_hash="ad4dda46ff84a5dbff1d42e90d0beaba" mdapi_config_hash="1459d002a686719edfca252a2510031b" id="3725926f-9a84-4cc2-963c-2294f47c017c" chipset="acmgt1" name="Ext47" /> + <guid config_hash="c9e1ea60e961eb2d62cd63e0335baa6e" mdapi_config_hash="31b6df06a991835e8111e8f3fd29c28c" id="357a3801-3f4e-4106-8f1e-c0bfe1c5f419" chipset="acmgt1" name="Ext48" /> + <guid config_hash="31ef72c1c9b5d63caab81e46dd8b965b" mdapi_config_hash="1b0b219d27f5d4d58d308294db29b1d9" id="8b324152-65d7-4b13-8fa9-cab57cd68471" chipset="acmgt1" name="Ext49" /> + <guid config_hash="b83478645a26c217b57cff42f3b03bf1" mdapi_config_hash="b98313c3d757eefb8d74bc2de9064791" id="db09b08d-a300-4f2a-aeb0-b797e7073ef3" chipset="acmgt1" name="Ext50" /> + <guid config_hash="027b27ac0e5a013b7d80974267bd01ba" mdapi_config_hash="84f75739a20eba4f730ec12d29cf3955" id="fc148084-3936-4dab-9b42-d7ba6ae1952d" chipset="acmgt1" name="Ext51" /> + <guid config_hash="c65948af6faaca6eaaa4b5a4333163c3" mdapi_config_hash="0a3053a9bf34e4c812e8eeec0e407ae4" id="f6a3c36f-9fd8-49c4-b25a-dbf26256bdc6" chipset="acmgt1" name="Ext52" /> + <guid config_hash="df10961b75bce890f27197ff4754a3a7" mdapi_config_hash="a7be445a851647c45e0697981c9b0ca4" id="a9a91699-a2b3-4841-8641-4f09f04be7e1" chipset="acmgt1" name="Ext53" /> + <guid config_hash="626b637a0a15877846696c77d5ef4d5a" mdapi_config_hash="3767426b64103ba3602d847fbb94e1dc" id="c44659b1-937a-44bc-8819-9c1ae31d9bdd" chipset="acmgt1" name="Ext54" /> + <guid config_hash="7c2880e5d049ddbd405fa599d85a6aaa" mdapi_config_hash="fc5263303b63feffda9d05addeaa4348" id="5bfd7ae1-27ea-4211-beb2-64565a49dfbe" chipset="acmgt1" name="Ext55" /> + <guid config_hash="84fd99cc784524421ddec2058f933ebc" mdapi_config_hash="d03ad457a06fd3c56b1c288f5218706e" id="7aefc824-149a-4f75-abed-f6430f8d4eee" chipset="acmgt1" name="Ext56" /> + <guid config_hash="1b71b566f061d64396f13d8763598031" mdapi_config_hash="1d779e9af2c4dfdc28c49734ca37ec05" id="0ccb67b5-7f31-4885-9973-d97e318a6f90" chipset="acmgt1" name="Ext57" /> + <guid config_hash="8ffbe4ee825fb9f9b0663a2e31dd4584" mdapi_config_hash="cbe482acfc6210f2c6e83b67e1dd79f4" id="af47d7b7-691b-4ec5-abfd-67ce50938fe4" chipset="acmgt1" name="Ext58" /> + <guid config_hash="5fd1e9d4b3477530518c4aa19e4f2e88" mdapi_config_hash="24f94b2640d05a5a78cfc381b06bb40b" id="0df81caa-d128-4f0e-9eff-69e0359fedf1" chipset="acmgt1" name="Ext59" /> + <guid config_hash="ac981182d4d54e8bed42914e98b134fe" mdapi_config_hash="1c183faebb4efd797a9f3d660531515a" id="09e71710-e49d-4d9e-b7f8-0e597aabf25b" chipset="acmgt1" name="Ext60" /> + <guid config_hash="1970fb6b4acf5e8651edba5fd0eb5898" mdapi_config_hash="f7353e7659737a8560dc6c88f2dd396d" id="374c4608-89a8-46d3-92fa-d4848e4a7a95" chipset="acmgt1" name="Ext61" /> + <guid config_hash="a610ce08cfff3c054f7579e71b50f5f6" mdapi_config_hash="a26be652e14ae14776fd120dbd0e27a5" id="1013ed95-4958-4cfe-8fbc-1482c07046cb" chipset="acmgt1" name="Ext62" /> + <guid config_hash="c550a26391338e8a2449cd190e577ae2" mdapi_config_hash="1692fe62dfcc589dd422e73375fbd4f7" id="7a3c4561-56d3-4b61-a619-ed8c1887e034" chipset="acmgt1" name="Ext63" /> + <guid config_hash="29ed261052036cf0fbd58503d7bc419c" mdapi_config_hash="ec50472b4169f291bcd9207459320d18" id="7d50a975-2d8d-48c2-ae61-86decd3aba1d" chipset="acmgt1" name="Ext64" /> + <guid config_hash="77de77319621ab975a70e13421a28364" mdapi_config_hash="7821e62cab1fc8ee6ef2f6ce696a0c77" id="c036e0c0-c23a-45ec-8b24-7ab7577b18d1" chipset="acmgt1" name="Ext65" /> + <guid config_hash="e44c2c6fa127e5968d2601c3831704a0" mdapi_config_hash="6c8126d0bb42dd17ea781ad2760d1456" id="df8127a7-0efc-487e-a819-a20c2471e30b" chipset="acmgt1" name="Ext66" /> + <guid config_hash="d136f506aac0742dc70ea2e73b671573" mdapi_config_hash="375ce65ce2a4ec475618717d66eadb00" id="f9d2c7da-14c9-45a4-97f5-a0d58dd3a1aa" chipset="acmgt1" name="Ext67" /> + <guid config_hash="12870e7e4ce04e3c685784cf7d1738b9" mdapi_config_hash="3ae6645297d97040354209d5f9df5543" id="210f665b-abaa-4389-95e3-7cad6f71db8b" chipset="acmgt1" name="Ext68" /> + <guid config_hash="9b3499f739fbe3ebc9f98bc1c2d5c1bf" mdapi_config_hash="c82d4511180c738cb3eac8e36c7e8d97" id="26eeaeca-61f0-46ba-b6ad-effb4ea076ae" chipset="acmgt1" name="Ext69" /> + <guid config_hash="5078259374d5a73b09d6f6a2c688b762" mdapi_config_hash="7be12445347ddcd53cde7da4ccf78325" id="18a0a208-526a-4705-9b06-9145f03433ea" chipset="acmgt1" name="Ext70" /> + <guid config_hash="cd9a83b23280bda0e6d460496d17b828" mdapi_config_hash="551313ba64fe8341b486f3c1cc64a57c" id="0677d73a-a563-47ac-8b11-9c7c3371fc04" chipset="acmgt1" name="Ext71" /> + <guid config_hash="a9239fa89f1f0cf7c6620dc2dcad4f8f" mdapi_config_hash="cb12023502953f503bf17dae2147616d" id="0a10b421-f9ee-4147-81e2-b2ceeb74697f" chipset="acmgt1" name="Ext72" /> + <guid config_hash="cb203c23ea26050dd482bd103414e45e" mdapi_config_hash="1673c8003c4ead48e73b044cb1214823" id="4051668d-2020-4f10-a9e7-e840455b125b" chipset="acmgt1" name="Ext73" /> + <guid config_hash="c03aa6470417b5be7d4f5487ef062a75" mdapi_config_hash="b16cc43512e2e40c9740123f3a2eb4d0" id="0548ee0d-a4b2-42a7-86c6-4171ed191bec" chipset="acmgt1" name="Ext74" /> + <guid config_hash="d77dc4e4035980a22d07bf5bf874448e" mdapi_config_hash="340dd4213c75862848df188c56d5f4d8" id="4604dea5-4e86-4713-b82d-fce110cb82f8" chipset="acmgt1" name="Ext75" /> + <guid config_hash="08e1815765794370d5e42dd67c95994f" mdapi_config_hash="363dc02eb84640ac6b3d311893f3abd0" id="d4171f1f-6e14-41e1-8b9d-0e28077aca25" chipset="acmgt1" name="Ext76" /> + <guid config_hash="5b0938a6bcc2dac189957c04420329bf" mdapi_config_hash="2eca3ee5793b0eeefb7c7bd89dedfe4a" id="8e91f318-149b-4ffe-9b51-42dd9b2c80a5" chipset="acmgt1" name="Ext77" /> + <guid config_hash="0190f37da63186cd4c61d6126cca0cfd" mdapi_config_hash="f80a3f9424d0b2500890d305811a6d2e" id="bf68f6bc-fdcb-4dbe-a923-abaeb72e8af6" chipset="acmgt1" name="Ext78" /> + <guid config_hash="c1550707504a93e323813e212372cb61" mdapi_config_hash="4788c65f10a4a2ef675cb5cc65feaf01" id="76b52ef9-7a1b-4475-b3bc-3a035acd32e7" chipset="acmgt1" name="Ext79" /> + <guid config_hash="b8d4127f9c668873b67ac274dea15397" mdapi_config_hash="c2d4aeefee3086dab23215c6ce17069b" id="171d99e0-0ace-4af7-8702-19e9fbb04a46" chipset="acmgt1" name="Ext80" /> + <guid config_hash="6de77558df213cb47f51df8c8c893828" mdapi_config_hash="67295b5dbc9f1eb7b11a297382c7cba1" id="f399ee12-9656-47fe-8566-416a8e4e4c07" chipset="acmgt1" name="Ext81" /> + <guid config_hash="7ba7f1fe44d3d883699261d9964b4dd5" mdapi_config_hash="151393cc3171de02e5b53ef5b056664c" id="766bbe37-ffda-41f0-92da-e31184140e43" chipset="acmgt1" name="Ext82" /> + <guid config_hash="d3d68205a754764911176ba6762372a0" mdapi_config_hash="e6ad03f903f93f9c50c09c3b1952d030" id="4d5d9009-c75e-4655-8374-503fd68ac0c7" chipset="acmgt1" name="Ext83" /> + <guid config_hash="7f78d20b5fbb0b58d128b642abec3533" mdapi_config_hash="e2cfb5bf6cdac08453235b342a8a784b" id="fd091d19-b48d-49d2-b99e-c5505624daa3" chipset="acmgt1" name="Ext84" /> + <guid config_hash="1cedd0dd5436db0a8a9f131dcffc247f" mdapi_config_hash="7bb34e96be686ba5d8d6bc643abf992c" id="d45a3f38-372d-4757-88cf-5589035652d5" chipset="acmgt1" name="Ext85" /> + <guid config_hash="1cdfa73f50573c5a12ab8518fca2ecb2" mdapi_config_hash="03a97ed99c0d771ddb9df899bd97cfb3" id="46f2e8eb-1a50-499e-9096-66f79932703e" chipset="acmgt1" name="Ext86" /> + <guid config_hash="71de590b75a77db655a0f06209c0151d" mdapi_config_hash="3319bcc11b652d32373fac2f49217648" id="35a680fa-e0c7-4e60-9a42-4e5c67d90b4d" chipset="acmgt1" name="Ext87" /> + <guid config_hash="fe82143b346d08a105650a5fe66ee823" mdapi_config_hash="2f05b7ad1f0b5ddafc7cebba157af45a" id="e1d0a1c4-41c7-4d82-a6f0-42ea537e5664" chipset="acmgt1" name="Ext88" /> + <guid config_hash="01c9fd39213390b28effeec74ebe4b1d" mdapi_config_hash="a1a6e3f5a5b271b64881fafc18eb75c5" id="c26ba023-58d3-4ec7-9282-f2638a292912" chipset="acmgt1" name="Ext91" /> + <guid config_hash="461caab506bc1142902f9cf82408280b" mdapi_config_hash="0a1b817f8d78cbc130cdc52d2da72858" id="7e5a8ef5-b434-4bf1-81f0-eeddaa072d37" chipset="acmgt1" name="Ext92" /> + <guid config_hash="c1550707504a93e323813e212372cb61" mdapi_config_hash="4788c65f10a4a2ef675cb5cc65feaf01" id="917a8688-0aa2-4b9a-a1a1-60dc016b17e6" chipset="acmgt1" name="Ext93" /> + <guid config_hash="b8d4127f9c668873b67ac274dea15397" mdapi_config_hash="c2d4aeefee3086dab23215c6ce17069b" id="dad96cfd-65c0-470c-a9d2-33d7881234c0" chipset="acmgt1" name="Ext94" /> + <guid config_hash="6de77558df213cb47f51df8c8c893828" mdapi_config_hash="67295b5dbc9f1eb7b11a297382c7cba1" id="811672a2-3bd6-473e-bd4b-dae0a5786a00" chipset="acmgt1" name="Ext95" /> + <guid config_hash="7ba7f1fe44d3d883699261d9964b4dd5" mdapi_config_hash="151393cc3171de02e5b53ef5b056664c" id="91ebb988-2cc5-4cb9-b071-3a9d76f42b55" chipset="acmgt1" name="Ext96" /> + <guid config_hash="a3be46f0437c5d4d8f99367d7bb15a4f" mdapi_config_hash="a2da0011f5563b65b6e58ad2067ef9f9" id="3e446a1f-d62b-4ea4-bbf3-845dc7f437e4" chipset="acmgt1" name="Ext97" /> + <guid config_hash="a63f8f95431c62f66d78154d919879a3" mdapi_config_hash="d911d2365e1cc9376ce297e97683d681" id="5ff858aa-50d0-44b9-bbf7-d07881ecdd9c" chipset="acmgt1" name="Ext98" /> + <guid config_hash="334de7afbe196e32ccba3f91a440ffbb" mdapi_config_hash="19a9b2203b3dfce6561f98e1d533ecc1" id="ba34845b-431b-41c8-9d51-5c8be2f0ac4c" chipset="acmgt1" name="Ext99" /> + <guid config_hash="0dab9bd08f0d241f1b3b7bff15a7b3e6" mdapi_config_hash="a4be5821e680778669686d2e1b69d18b" id="978e5720-eb5a-46fc-aa24-7a34a853130b" chipset="acmgt1" name="Ext100" /> + <guid config_hash="cdf411e1398919009c793edb3c6a1ef5" mdapi_config_hash="d0b3b253fc92a97d1d1e7d40a1e688f9" id="9320b9d5-43c9-435d-b0ed-ec8a565d742d" chipset="acmgt1" name="Ext101" /> + <guid config_hash="00f65434ddfd767057046f4a8529109c" mdapi_config_hash="58360716f9563bf0527eab58a9477b00" id="ff32c8bc-39ed-4dbe-9d88-2d2cdb973872" chipset="acmgt1" name="Ext102" /> + <guid config_hash="76f88688dc2cbc79b414c4f87de1ffd1" mdapi_config_hash="57663957a360068e72c665f510c5e03d" id="73a834ac-afbe-4ca1-af18-af9aabc0280a" chipset="acmgt1" name="Ext103" /> + <guid config_hash="5bd9e967fa142c9cd99ce18942cce146" mdapi_config_hash="f4cbe312ad55a843507568f33ebf4948" id="eb13a257-4c07-469e-8bef-55da7dff1b41" chipset="acmgt1" name="Ext104" /> + <guid config_hash="6faa9519a6a3ea870fd0455d1d0b1c42" mdapi_config_hash="df94ec94cd4f8b182bb2e11117efe708" id="0343e7a1-8f7d-44b8-a75c-73baef6cc22b" chipset="acmgt1" name="Ext105" /> + <guid config_hash="c5b60a8b8d9b458cf541728c63fc1401" mdapi_config_hash="b3e49a50efa2110de26dce1314c6d3fb" id="048a2bcb-b786-4b51-b350-8908f50e69d7" chipset="acmgt1" name="Ext106" /> + <guid config_hash="af725fdf810c81708ed915adc4958b31" mdapi_config_hash="d55a8dcf74374a916e50477874564051" id="e393956e-78b0-4f7e-8482-cb13c1706ecd" chipset="acmgt1" name="Ext107" /> + <guid config_hash="b3d0672f1145734142cd778b220e6d5c" mdapi_config_hash="b8a1b217157338f594c735597903e064" id="2ea41c1b-4751-446f-b178-9aa17cf85d72" chipset="acmgt1" name="Ext108" /> + <guid config_hash="aaf1aba5206945089786b80ee5dbbce3" mdapi_config_hash="665caf49abebd4fa0fd7db4dc10b8c0f" id="32b61c78-b7d2-4162-8964-dbf47f187290" chipset="acmgt1" name="Ext109" /> + <guid config_hash="7154499c71bfd907f108bb2027c766b7" mdapi_config_hash="9a3e1efa778cc11535c5e236498ffaba" id="536ed3af-e1e6-4f5a-be8b-3eabfa9487fa" chipset="acmgt1" name="Ext110" /> + <guid config_hash="547f2d8ffd47d9568936e7195ebef561" mdapi_config_hash="4dc7582d62a68ce09049cbab94dd8f1e" id="11eedd05-3490-4a57-8ea7-56e1d2dd066a" chipset="acmgt1" name="Ext111" /> + <guid config_hash="dbb982cd29b346a7fa3640d9c2db8adc" mdapi_config_hash="963f6749d596acd802831ee8284e92e7" id="2324dd37-ac7e-4a5e-8d0c-f7b94fd36961" chipset="acmgt1" name="Ext112" /> + <guid config_hash="dfda4f47ade0b496dd2d500e293537f9" mdapi_config_hash="0adf78de606d7b56bca529d410dd0cd5" id="03551880-b834-45f0-8b71-e61f548e5d83" chipset="acmgt1" name="Ext113" /> + <guid config_hash="812fd46833424d9eb0d05ee28f3e9365" mdapi_config_hash="91bd557a7b5f5fb05169b2a434c14ace" id="8d368e2f-7c30-4a9e-a20f-b098105300b7" chipset="acmgt1" name="Ext114" /> + <guid config_hash="c5442482b5d76fbecca8136a426596b9" mdapi_config_hash="17e398bab2a10acc1afa048d8b17b622" id="cb96d66c-bfc4-4de8-a3dc-5d67c5acda38" chipset="acmgt1" name="Ext115" /> + <guid config_hash="33b58de69e947a671ad57448198b5b46" mdapi_config_hash="e6a643e3eee08059b653de6565148a70" id="2355cc8b-c596-479d-ab76-16c2d3644b5b" chipset="acmgt1" name="Ext116" /> + <guid config_hash="c3cb81f0b071c85b13c2cf08ae14bbb3" mdapi_config_hash="e96ff5b4ea7dfa7eba7e649d2f177043" id="006584d5-ebb1-4d21-a8ab-cfa66e75dae4" chipset="acmgt1" name="Ext117" /> + <guid config_hash="14b20811c2c2318e80b542ff3879654f" mdapi_config_hash="05ec953ec963b69a62ec53c5b8ec6972" id="9ae38060-8785-44ec-a758-4082b94af42d" chipset="acmgt1" name="Ext118" /> + <guid config_hash="bb81ce9686f82ee27a3e71e99f144c1c" mdapi_config_hash="6d808ffc8031ed3235b2fb62d5138ad3" id="8ef448b2-776c-4d73-8a44-eabfb2fcc888" chipset="acmgt1" name="Ext119" /> + <guid config_hash="08f78d136c39dab6db05a652d0608dae" mdapi_config_hash="ad30e32ab47c0b4bee479e434443d1a8" id="3b5b473c-259e-4eb7-8643-a88b53aa2e1c" chipset="acmgt1" name="Ext120" /> + <guid config_hash="646acc562ed82884a3461f07162ec912" mdapi_config_hash="963f4359857199bbfc3fb3e339806659" id="f2045966-fcf0-479b-aef4-0358574f67ee" chipset="acmgt1" name="Ext121" /> + <guid config_hash="c57b8fa92ab135ad9aa90ed90993938f" mdapi_config_hash="077f4d70e66e7ad22189f8e91eacb79a" id="e175d8c3-7672-4e57-83a0-87caf56a5440" chipset="acmgt1" name="Ext122" /> + <guid config_hash="541656fadae59b87e7767e418e2fb496" mdapi_config_hash="5a5456483d752bff51426fca7dbc3a80" id="7f7218b1-462c-4f08-8032-86e426dbc83d" chipset="acmgt1" name="Ext123" /> + <guid config_hash="b5de9a4dd9da9c6502f6cd3c5fd212c0" mdapi_config_hash="20bc797bcce3ba76e4085454a0585bf6" id="bd7a8af3-2d23-4c95-9cb1-429840dfc38f" chipset="acmgt1" name="Ext124" /> + <guid config_hash="1c44b1bcadec2656fbffb4ad467c7000" mdapi_config_hash="8983c6d0d569869031b460755f781856" id="2e16a8a7-613a-4df1-91bf-eec73d281822" chipset="acmgt1" name="Ext125" /> + <guid config_hash="33d81eee011fe0eb1e07dca082e37e9d" mdapi_config_hash="b5eead3b9158531950af4fda3f2fd895" id="7cddf3f3-4ed8-484d-b8c7-82a4857ad030" chipset="acmgt1" name="Ext126" /> + <guid config_hash="2541a923559b89e096312f25291f3a56" mdapi_config_hash="d3611aec69ee7d4be3ae9b4f61a30fd0" id="cc78dbe3-bf19-4e4f-8957-c75488d8a47b" chipset="acmgt1" name="Ext127" /> + <guid config_hash="52228f4f134a3d67aa2c47b5d90e41b7" mdapi_config_hash="c3444a5f01a4b0f1b5b3593c9864914c" id="9722a829-b7e5-4716-b62e-ff6cb930232b" chipset="acmgt1" name="Ext128" /> + <guid config_hash="7770de77f7a96738f93607fad710f052" mdapi_config_hash="079deb439f2c075c8f7bfd83a7c20f1a" id="4d7bb9a8-4b02-435a-b2be-16a2314c421d" chipset="acmgt1" name="Ext129" /> + <guid config_hash="28b1d112fca73254fe836d767bc09cc5" mdapi_config_hash="b20dc315eaec1e299473f9afe76b3fee" id="307951f8-0159-4d0c-bc8f-04fee7247e0a" chipset="acmgt1" name="Ext130" /> + <guid config_hash="f28bc24b5f2f0af90d1c7ecbe81cf63c" mdapi_config_hash="d3f0e2966b0ecea7dd97ac151a7872ee" id="aa04f1fa-826e-4b25-a7a6-4fbb4a7d0dd2" chipset="acmgt1" name="Ext131" /> + <guid config_hash="55940375eada7d9e77c0d6c20af8c841" mdapi_config_hash="8d8b1ebc95f5448cf6a234ac02cfae7e" id="9121bb2c-a28d-4796-a60b-52905e317576" chipset="acmgt1" name="Ext132" /> + <guid config_hash="82dbf6cb5eb61b5bfaad996e73f51f95" mdapi_config_hash="673f1150779b67b72218e18d29d1c50b" id="14a8d00b-a361-4bc5-94fe-4664b837cc56" chipset="acmgt1" name="Ext133" /> + <guid config_hash="676a59a9f81d7ba65913df56d804733a" mdapi_config_hash="3d5f622d4122f3536ba117713ae0d394" id="74fda750-6436-4a97-897b-8eff0a63de70" chipset="acmgt1" name="Ext134" /> + <guid config_hash="ddc017469ab0800b55c472101abf2a7c" mdapi_config_hash="83d7b8f2ff34803dd62c7cd242be4bbc" id="1712c473-6635-49af-a6da-43c724c96dcd" chipset="acmgt1" name="Ext135" /> + <guid config_hash="e66c1a3078d45e4eba55ff67672595d2" mdapi_config_hash="e93269c4c6b92f99cb87ce1c291604b4" id="5c1a57d6-2dff-48ca-98bf-02c71a22057a" chipset="acmgt1" name="Ext136" /> + <guid config_hash="d889f4982005a8be6a9032b55a53f661" mdapi_config_hash="eb63aef957a615915f35a63942071700" id="6ede1a62-069a-4b3d-8b17-000005541bfd" chipset="acmgt1" name="Ext137" /> + <guid config_hash="7569421c86416bdfef5ea87766553659" mdapi_config_hash="f986e11d7c0bd27e898c1a335c7c05da" id="4a31afda-7a08-4d3d-91fd-138bae3eb836" chipset="acmgt1" name="Ext138" /> + <guid config_hash="76431bdb2c840d296d08fb8490fa412e" mdapi_config_hash="5916bd135a5e11f46cdfff26e5df894f" id="c800871b-2505-40f3-99e2-6c0ccb9102f2" chipset="acmgt1" name="Ext139" /> + <guid config_hash="59bb5b34670411dba6ec563832b0194c" mdapi_config_hash="38ac5318030f247a1d1f71b6ffc79730" id="067cbc21-6b6c-48c6-b47c-922a27b60b30" chipset="acmgt1" name="Ext140" /> + <guid config_hash="23300ecf639bf4d6f03ce687883ad6ef" mdapi_config_hash="f01f79b830965539a25808f87b675224" id="e8ee8621-7110-4d70-9651-846fd4fa3c3c" chipset="acmgt1" name="Ext141" /> + <guid config_hash="6be90ac264f6686075e7b0983cbafb45" mdapi_config_hash="5940a41ff1b27c7f40dece9288fec68c" id="87a9d5f2-2029-48b9-bbd3-78867e022166" chipset="acmgt1" name="Ext142" /> + <guid config_hash="97014e93fa896b41c6a6249b290a4810" mdapi_config_hash="79935c695e94c27bd1c5aa7d5f8fb9b5" id="bffa2382-3aff-471a-92af-5dd571c7868f" chipset="acmgt1" name="Ext143" /> + <guid config_hash="5834074eda185c8cc325c2b7f132e3d7" mdapi_config_hash="cfa85b8484c57014397c4874d0f54891" id="951f3ee4-6a3c-470c-b403-c01ee6a7e97d" chipset="acmgt1" name="Ext144" /> + <guid config_hash="a8b569a52638fef946a6cb5410c99d6c" mdapi_config_hash="2a98991afe86f48f7e776d105f412d32" id="4d34c592-973b-4e75-a451-cf4ad90f79e1" chipset="acmgt1" name="Ext145" /> + <guid config_hash="e6d0d033ae35a3e8300b39a13c4ea1cc" mdapi_config_hash="7bb9c8d361090eb4a40bf2e4de80d45e" id="4a881a06-44b8-4199-a609-e0e4e22631c4" chipset="acmgt1" name="Ext146" /> + <guid config_hash="fa5beb51e65b39aaaff4e502d0e6a06e" mdapi_config_hash="f1916b9e3cd30c4346be0539fa09c792" id="20b339b5-abb6-48fc-8a09-bb4cafacbb72" chipset="acmgt1" name="Ext147" /> + <guid config_hash="a5e21e9d2f3b381db9fce8d35cf19952" mdapi_config_hash="c50ef475e25bc2d4c3f4493015e9dc90" id="8b62b22b-a467-4638-a113-81a5c59d64e8" chipset="acmgt1" name="Ext148" /> + <guid config_hash="b968adeb9021a13e0f2263500f6a3833" mdapi_config_hash="c70a6d563e125c3d94425ee1ad7f0678" id="a563237d-76f3-45bc-82d7-c4a441c9230c" chipset="acmgt1" name="Ext149" /> + <guid config_hash="5c0aca3eefedb553f92bcd63514e29fe" mdapi_config_hash="a49da2e3406c3e44cc79dab1bd62b286" id="dee6b408-d907-4eae-a961-7262816be56c" chipset="acmgt1" name="Ext150" /> + <guid config_hash="d658163dc7d15f51bcfa16b94eb00bf1" mdapi_config_hash="a9602665e74fd06fce336790d4969b28" id="c2bf9d7d-0998-483b-bede-7ced3f8cf423" chipset="acmgt1" name="Ext151" /> + <guid config_hash="b52c6f39fe2cb4069835665013ae1067" mdapi_config_hash="7506b0638b426552ff36724edc616b02" id="06c6ecdb-9b66-400b-83ca-484309fba532" chipset="acmgt1" name="Ext152" /> + <guid config_hash="07f1fa5eab01e0979008b318bb18e0b3" mdapi_config_hash="a879c73b8ffa76232398a36e29e025c5" id="af12ca37-4f11-4c7a-b6b3-695210972c92" chipset="acmgt1" name="Ext153" /> + <guid config_hash="9cb71e8d432d5abeadf3a8f3dedc9300" mdapi_config_hash="09bc198f7a44f1505d34e8aa7145d20d" id="f72930f8-79c3-49f3-9fa3-07a05070d1fd" chipset="acmgt1" name="Ext154" /> + <guid config_hash="888c5441777eb18dfafb04ccbbabc63e" mdapi_config_hash="4684fcb9f09f632579567c77d3e11a12" id="ac5346ae-2be4-47d1-8300-6e25c445297d" chipset="acmgt1" name="Ext155" /> + <guid config_hash="dcffe4412b7fe076c6b2512e3795bd39" mdapi_config_hash="1840f16963fe0890e9830e4cc50cf7b9" id="f8f7091d-7f71-42cf-84de-8ad23fe45432" chipset="acmgt1" name="Ext156" /> + <guid config_hash="c557ae19d5c006a14f9d599e7d93182e" mdapi_config_hash="5940392ed7093a62f549687907af0bad" id="ac6eecdc-0074-4a94-9269-48b5e3833439" chipset="acmgt1" name="Ext157" /> + <guid config_hash="8c230a0109d0d2352a5ee9557ce72ca6" mdapi_config_hash="b7b0d316e74f43e17e7e50d34fb34014" id="4c27f089-06ab-4c33-b084-fc7cc9cc78ea" chipset="acmgt1" name="Ext158" /> + <guid config_hash="7c79bd0534b65f2550b179e1f454d38a" mdapi_config_hash="c6718e4bb039e1da93839b0461171ce2" id="1918a5b9-f991-4451-afff-565a739dbfb9" chipset="acmgt1" name="Ext159" /> + <guid config_hash="35c217ed780ba0cb9f4b8fd8862725aa" mdapi_config_hash="8b4c9654098dcbeb200d1d912ff6dba2" id="6cf8ce0d-617c-4ed8-962c-6b750c0cf012" chipset="acmgt1" name="Ext160" /> + <guid config_hash="604a47fc01da97814df40e74c343b6af" mdapi_config_hash="e68dce7ab35522ba00b6bd7b0a03fa48" id="453d205a-8bd1-4ebe-b617-1168cbfda3ca" chipset="acmgt1" name="Ext161" /> + <guid config_hash="9913f4fa2c601a9e7bb052e57cdfa7d1" mdapi_config_hash="26db04f8e7f6b37025289329820079dc" id="9d94eefc-0c25-4725-a6af-3eb468cae42d" chipset="acmgt1" name="Ext162" /> + <guid config_hash="07805a92788aa3b698fcb69715e9d077" mdapi_config_hash="c08395849f5c0c80b27aca3eeafea470" id="fbba6a19-98df-4a94-9a42-2660674d8e25" chipset="acmgt1" name="Ext163" /> + <guid config_hash="51855c8bac86872917a66d0187f7c5b6" mdapi_config_hash="c5bb813efe5c112728ffd4ffcc5e54f8" id="ec971ed3-3fb2-4d4c-bf27-08eeb418dc68" chipset="acmgt1" name="Ext164" /> + <guid config_hash="04545ecc50b3f013a688e4ad9c463907" mdapi_config_hash="e7c76293ab50c4f6221e8e1e150bf5d9" id="a1d6b535-1d00-4951-bdcc-8586b9ba7c0a" chipset="acmgt1" name="Ext165" /> + <guid config_hash="e395d03773d9eab337244898db0fb89e" mdapi_config_hash="2ea8208d937ed0cc22b25c84949a4ec6" id="5f78bfe9-b17b-4df2-b1de-54e7c84d93ab" chipset="acmgt1" name="Ext166" /> + <guid config_hash="3a442c930a8f7ca95ee6dc1620c2a11b" mdapi_config_hash="e2426faae7e57dcc2ce905aa984b10f1" id="c76128ab-fe65-4ccc-94ea-f351ccbd5f16" chipset="acmgt1" name="Ext167" /> + <guid config_hash="7948e4a15362974affa25138054d6c6a" mdapi_config_hash="d43502fe41d3881002b288a2460606ab" id="fac0445a-ae46-412e-bf26-0b4dddd64485" chipset="acmgt1" name="Ext168" /> + <guid config_hash="857e476f2817d9ef41011cf7b68af59b" mdapi_config_hash="1deabc74574a3fca3acb201789fd546d" id="386322a7-4cf5-4bf9-bbd9-35a6477a4479" chipset="acmgt1" name="Ext169" /> + <guid config_hash="21388f53af5966ab4a742b05690659ea" mdapi_config_hash="01ff8f7baddcd345049ec8af458b50de" id="b175ecb9-b8cd-4f9c-97d4-73bf5cf12fa7" chipset="acmgt1" name="Ext170" /> + <guid config_hash="b1683389c5870c195ef875f67da2eba7" mdapi_config_hash="4b69e7390c790e0e859f6b3bfefe040e" id="f2b8f05e-1936-47b6-b1ac-99a441971cea" chipset="acmgt1" name="Ext171" /> + <guid config_hash="e5dbd4c2f74d568fa61a926a0c26768c" mdapi_config_hash="ca3e762d09e1ba42a80275552827ced2" id="72ae9945-8887-4c8b-acd6-559e4b92203e" chipset="acmgt1" name="Ext172" /> + <guid config_hash="d60274e2d6d33f0aac5fe66c88ae7745" mdapi_config_hash="3b75890a46add6824dee3a488b932c7c" id="1861b4c1-3950-4855-a8db-6e0b90fc44aa" chipset="acmgt1" name="Ext173" /> + <guid config_hash="c4cbd0d5f79d939528bf99f3c9d19f06" mdapi_config_hash="830845e5c1a7eaac5c279444ca53a68e" id="b29accb6-ff91-45d5-953b-2ca3c0983f66" chipset="acmgt1" name="Ext174" /> + <guid config_hash="b5c3801c4d7f0805bdadd73395de63b3" mdapi_config_hash="b156b7caa8403ea387aac418dd000681" id="218c081a-6fe6-4a26-afb6-ca6bd1bf78eb" chipset="acmgt1" name="Ext175" /> + <guid config_hash="92965f8f98c2ac74cb086d8e0a5f8ec6" mdapi_config_hash="84f869fc7c855520e201c3743d5a249c" id="f7d99e5e-4cb7-47da-835f-ec31a8838dc5" chipset="acmgt1" name="Ext176" /> + <guid config_hash="f53b3396bf7a5a0583f9d799788ec87a" mdapi_config_hash="0603e1b4d6ae13d1ceb7d7e7f101ff6e" id="da0fd1f5-25bd-44d7-aced-87d78f3f0a7d" chipset="acmgt1" name="Ext177" /> + <guid config_hash="a848277d97c7cf23195bae0613ce5afb" mdapi_config_hash="2db7e834c32ac0b499c8db71423e7b05" id="e1495a70-d6ea-4867-aa48-4135567bf19b" chipset="acmgt1" name="Ext178" /> + <guid config_hash="23480ec698fed6b016fe7e2f24427e4d" mdapi_config_hash="705755afc8ecaea45b2ec83a4f437291" id="d2b946b8-ddbc-47bb-b02d-f89652163b6a" chipset="acmgt1" name="Ext179" /> + <guid config_hash="94ef64f24ad8c70496934cb35ad7904f" mdapi_config_hash="c8226547ef53e1ed62bb0601038c0c89" id="44915f22-aaca-4871-b8ca-7a83aa83d605" chipset="acmgt1" name="Ext180" /> + <guid config_hash="c8da848485fb2a10ad3160d208fe5d3b" mdapi_config_hash="302146d82d8f774d8ff506b7bad22fcc" id="fe0eda19-22d8-410e-b61d-24bce297d8ae" chipset="acmgt1" name="Ext181" /> + <guid config_hash="2d8e82c9343238ecb411ae60acabb82e" mdapi_config_hash="9650fe0004b7911ec21ab17a7be7a0c4" id="aef5adcb-e3ef-4839-aea5-a6bbb9bb5faa" chipset="acmgt1" name="Ext182" /> + <guid config_hash="8b9f3bf34f2d3c62b8b6e8a25f345e6c" mdapi_config_hash="477a24237a8ae6609ce26951d9906482" id="1cf7ed05-8260-4abe-b00c-7f3ae3441d4b" chipset="acmgt1" name="Ext183" /> + <guid config_hash="65823700cb27a3511db3893df627340c" mdapi_config_hash="6fd71469af2adf045b72f40701246365" id="39badcb4-5410-4218-b272-1e5edc606b5c" chipset="acmgt1" name="Ext184" /> + <guid config_hash="1d1adf473dede14cac82d966e834c82d" mdapi_config_hash="842415ea688559f6802382682672f2e5" id="e1e40930-36f2-480c-b15d-cf2fb07f18e1" chipset="acmgt1" name="Ext185" /> + <guid config_hash="2ba9800bf3cb57a4da14f82928c67398" mdapi_config_hash="cf0037ecc186702f368df7b980269fa1" id="4edf8144-d144-4110-b34c-0f6133409bae" chipset="acmgt1" name="Ext186" /> + <guid config_hash="ece2e98ef546e8fdc50d12eab7d71bed" mdapi_config_hash="21271b27f8a41301e8fa9bd4ab328dae" id="a035c742-180e-4bf0-8009-5668cb08981f" chipset="acmgt1" name="Ext187" /> + <guid config_hash="cdd16828dff28943f225ff159c0f5a52" mdapi_config_hash="c76a7c4a68e1c738aebb6d832bc88589" id="1236f8d6-fa44-40eb-addc-8e4695b02ab3" chipset="acmgt1" name="Ext188" /> + <guid config_hash="85e29c24456fa871dcb87f305e0140c0" mdapi_config_hash="b32c81a656e0b0692e1985c223afb0df" id="d37f51e7-16dd-4616-88a0-cfc54b90a43f" chipset="acmgt1" name="Ext189" /> + <guid config_hash="0ad3bbfa2aeb4d2b449d6f43101931cd" mdapi_config_hash="70c6c88ad073893c8624ed20673ed0fd" id="c9ac09ca-af08-4072-8e8b-dbc81cd95ada" chipset="acmgt1" name="Ext190" /> + <guid config_hash="e64a4fa60526e0d951731a93329ff1d6" mdapi_config_hash="48bd683c1de30f027d832c89df7841cb" id="9e1699e6-cea6-46de-bb89-050415f0a8a2" chipset="acmgt1" name="Ext191" /> + <guid config_hash="a91cccc12d8966a3ff456c56599cc805" mdapi_config_hash="ce4527482c9b6eb92192c0a8e93f52c3" id="e671dd75-968d-49e4-9c03-5269d95de66e" chipset="acmgt1" name="Ext192" /> + <guid config_hash="55a91f48dddc3417ceeddd2c1db1ebdd" mdapi_config_hash="d6406f5a85d661045b3b840d475680f4" id="9a53799a-f746-4d6d-a723-57af53243454" chipset="acmgt1" name="Ext193" /> + <guid config_hash="afab6f08fa93d48cc8c86dd6ef451b8c" mdapi_config_hash="3d7b6804ef19dee55a7ec8895bfb3f3b" id="3b4157d9-9905-48c5-a9fa-3f53f94b9b5d" chipset="acmgt1" name="Ext194" /> + <guid config_hash="571e08e397d67e114b9ae50f18dbaab9" mdapi_config_hash="580a0056d940c4e47fa87a1a4f134106" id="b7468d66-8862-4deb-aeb7-d2cefa81ea0e" chipset="acmgt1" name="Ext195" /> + <guid config_hash="963bf529feff600aa3a4094b02016c11" mdapi_config_hash="21cefdaf4f5b32f25e7c9ba83c2ae6fb" id="a09b78bb-b10e-4250-84c5-336f839c4e0c" chipset="acmgt1" name="Ext196" /> + <guid config_hash="e8525f80ae6aea52ca46a1ad0613b7ed" mdapi_config_hash="55fca8f4273b872b71e9b0dcfab070cf" id="8b7e5a75-7734-4d48-bdd7-0adf976b8a89" chipset="acmgt1" name="Ext197" /> + <guid config_hash="8e61af04e6cc7b31995cfa7eaab214f9" mdapi_config_hash="367308324b88151877ec160474dc4313" id="8157717a-d7e3-465e-a6b9-a5a5d4902b49" chipset="acmgt1" name="Ext198" /> + <guid config_hash="c2dfaa6c92411bf8d12b7de300db927e" mdapi_config_hash="1ca97769ab5b0f1369fafa639b3b63a9" id="fc122f7d-c66a-458b-a06c-cee29569262d" chipset="acmgt1" name="Ext199" /> + <guid config_hash="0ead438a310183129e09f8f44ab75620" mdapi_config_hash="6555926f971e2592381817db5e2146f1" id="80da4065-1cc6-4e17-8594-731e1acc01ef" chipset="acmgt1" name="Ext200" /> + <guid config_hash="f11f6e82febb1c53137bee7e1550caeb" mdapi_config_hash="9b2b0b068414507e89fd35009d30d55a" id="5c664618-1692-4ded-b6bb-a9e17f37250d" chipset="acmgt1" name="Ext201" /> + <guid config_hash="aacd63fb18fb085688a206352d3d76bf" mdapi_config_hash="3370c56f2f4bf2d6499a8c422f40f630" id="671c0d10-0f02-490f-acf6-2d2a88c0bc62" chipset="acmgt1" name="Ext202" /> + <guid config_hash="860d7ddea477c3016b49bfebd5800154" mdapi_config_hash="a5552fc70c572fd1dacc6081502ea76a" id="d1c46c3e-fd18-4ca4-905b-64b95a6f10e8" chipset="acmgt1" name="Ext203" /> + <guid config_hash="beb1b18cc16f797a2e82c02869235d25" mdapi_config_hash="1de83bc512187b0e4c80028d28a0616d" id="4e426fcb-9591-42a4-9b9a-7207d548bc0b" chipset="acmgt1" name="Ext204" /> + <guid config_hash="82a2f46c4eef6d9cddc0b121fdf25171" mdapi_config_hash="f806294b27a9f4bc0743049749279c4f" id="12aa9868-f7c6-4c8d-80fe-cc1605608ef1" chipset="acmgt1" name="Ext205" /> + <guid config_hash="2e41aecbbb175c3cc8ce6c71fc6a2456" mdapi_config_hash="15f2b8c96079fb0d22f8a70a7b82a81d" id="7a64382c-70d1-41c4-a1e9-28a8bcb4785a" chipset="acmgt1" name="Ext206" /> + <guid config_hash="6ee3ff83d61794660af8edefdeb312ca" mdapi_config_hash="f4435ca88925816e7b329f36fc365223" id="2fa63ea7-c2d9-40fc-b9d5-74cf4a7a3761" chipset="acmgt1" name="Ext207" /> + <guid config_hash="ed66f125d1a6288f9e744a3a012b9eab" mdapi_config_hash="2deb80294fb1d10c393dbeb19cc9a272" id="c26befd3-0199-4b5a-9768-963eded47ddc" chipset="acmgt1" name="Ext208" /> + <guid config_hash="d070985e3873f5c426ab81284beae9bc" mdapi_config_hash="c3dd7e3225334fe0fbada0150e1835d8" id="6ac62a5e-3593-4f68-90c1-452cbc83af7b" chipset="acmgt1" name="Ext209" /> + <guid config_hash="e350f7c72e22d5f9e16ceb58a5176ce7" mdapi_config_hash="b27726f0242de4625cbfa2dba61e450e" id="869238af-0548-47ba-9703-fb4d12825f89" chipset="acmgt1" name="Ext210" /> + <guid config_hash="b7a2bad2e3db75626d0c1ea6ab2fda04" mdapi_config_hash="1a1710d5c47135cb3c4e5d438ab93cec" id="1986e401-2a47-4ced-be1e-1412f795388f" chipset="acmgt1" name="Ext211" /> + <guid config_hash="4c013e7592931a1c49cff2a9b5008ccd" mdapi_config_hash="d5e721c891920ce6edad86dbf29d3eec" id="61f0e78f-0095-4595-993c-0d2a7b474f05" chipset="acmgt1" name="Ext212" /> + <guid config_hash="2ff9fdd0f973d1b2c9981a73e4c11a57" mdapi_config_hash="6037ab583666d0712c47d02cb906a976" id="aac878c4-c131-4e46-a022-673b0c0f6adf" chipset="acmgt1" name="Ext213" /> + <guid config_hash="2204c90670f09655ea8a4aefbacfdfbc" mdapi_config_hash="7718c7f404e43bd1aef3d373ee1032a6" id="02fc686b-6061-4224-957b-0c80f4fb5973" chipset="acmgt1" name="Ext214" /> + <guid config_hash="0fa38b874093d098ea91a3bd99af92db" mdapi_config_hash="0c4a2e0e4b1c911ae03c911904a5675b" id="6965f87c-43a9-437b-9bfb-78e30f504806" chipset="acmgt1" name="Ext215" /> + <guid config_hash="0dbca4480da832f2b9793898ca72cb4d" mdapi_config_hash="5854c837a22ed98fd5eff9fd6df26f79" id="834f9ddb-0ef2-47a8-bcf3-871cf2b5634e" chipset="acmgt1" name="Ext216" /> + <guid config_hash="81f19bd1f4670e075909fedacd30b88b" mdapi_config_hash="b2a9f3ed862189e64a45f8c08add29d2" id="dd1c62fd-8d41-4f1d-a980-95eba5b62d3d" chipset="acmgt1" name="Ext217" /> + <guid config_hash="a282e7cbb57df81b6a723aeb2e205092" mdapi_config_hash="77b385307a36d49b59450fee078d4272" id="1a1e9195-858b-47aa-8f1c-c66eea6efbb9" chipset="acmgt1" name="Ext218" /> + <guid config_hash="a75d849fbca5e34d2993b4cf7e41aa9a" mdapi_config_hash="cd8713a3c6807080e662f4f9a9d73a46" id="a75e98da-d723-42d0-8cd2-c4dc704e6aa1" chipset="acmgt1" name="Ext219" /> + <guid config_hash="70261a93ce4650d846fe430ab7fcda40" mdapi_config_hash="3bda96e4f7e882ac7dfb065e9a8b41c1" id="440ba901-38f7-495d-986b-486534274741" chipset="acmgt1" name="Ext220" /> + <guid config_hash="56a4bd04f44424a407d92c3c75044fb8" mdapi_config_hash="13206261e313debd6d6c2ef1b8df6176" id="6d047d05-95ca-46b4-aac3-a7b055b8fd15" chipset="acmgt1" name="Ext221" /> + <guid config_hash="8f6c91a5e970a5048e55a4517b650dec" mdapi_config_hash="4c0a273f63e225fd4fec4b22741318a3" id="a338ef12-c1bf-45f8-8c0b-084fa99376c8" chipset="acmgt1" name="Ext222" /> + <guid config_hash="e160804e09a5607f5e1b226cb4935926" mdapi_config_hash="e84d31b2c04f8e8bd179ba877325b6cc" id="2684ee0f-2950-4445-939a-549f07d3370a" chipset="acmgt1" name="Ext223" /> + <guid config_hash="af65d1ae7b8dc14e99a9cf4f58cc9e8d" mdapi_config_hash="9a8ab35720351d714b3d6e9aae109f95" id="e1de83dc-4fb2-4e1a-852d-9c5e49bf45db" chipset="acmgt1" name="Ext224" /> + <guid config_hash="6c07fa24d7b4defe09008f9ffecab192" mdapi_config_hash="9ec8053955ed7565061a683ad3c8f883" id="81beafd2-8d1e-46bc-96c0-7f914522a760" chipset="acmgt1" name="Ext225" /> + <guid config_hash="8c0460b9c0e48f4a64020d0dd37c8f7e" mdapi_config_hash="b21942eb6a9ff37c65d3e210ff331980" id="e299ef7e-7be1-4554-a5c7-e64d36be98e6" chipset="acmgt1" name="Ext226" /> + <guid config_hash="73a49c0333e18513b6aadc429d072d67" mdapi_config_hash="3cd5f41724914fa12a8e0b2372db6400" id="e9d1019b-205c-4d1e-a2a7-fd7708a6ad91" chipset="acmgt1" name="Ext227" /> + <guid config_hash="2e4d3d499567062a97d9aa13809d3533" mdapi_config_hash="aa100730afb50729fa562dafa28689a7" id="ccf8706c-40ea-40a6-ab4c-47e5b32a1485" chipset="acmgt1" name="Ext228" /> + <guid config_hash="6a19e9e3497eaff3ce66618dbde20ef0" mdapi_config_hash="6a0112d99e95a9d5426155190ce64c1b" id="2ebb4747-7eb7-4a65-96f8-6a57131a5e82" chipset="acmgt1" name="Ext229" /> + <guid config_hash="195df4681d7198f5631062d25f2edd0c" mdapi_config_hash="0a5b381836c45fc8a14737b06877372b" id="baaa8d7e-40b7-4abd-991a-d71265437230" chipset="acmgt1" name="Ext230" /> + <guid config_hash="5d2eefa979d4c0db01bbcfdcf3ca245a" mdapi_config_hash="3633f6117ad1f4bdd29e2bd2c77cddd6" id="a57ee736-e6ab-4e8b-a719-3a63d7d20f34" chipset="acmgt1" name="Ext231" /> + <guid config_hash="25d0f6725e5b3c88f738e2492c063363" mdapi_config_hash="8cadc932b5ef29d27d16b5009e9ccff9" id="c33082a0-ae0d-4679-adbc-9566c75ffe6a" chipset="acmgt1" name="Ext232" /> + <guid config_hash="af5f05fa3cbf4463fd5efaf4435632a9" mdapi_config_hash="62b378e721b105a41e85292337e7d7d6" id="3381cf54-6e4f-48b1-81c9-d5cfd542d316" chipset="acmgt1" name="Ext233" /> + <guid config_hash="5f0ff9ce44509e07359137b83c40a050" mdapi_config_hash="dd0195354e6f119221975d5dcbe16acf" id="b7f1e445-7aaa-4fe3-a7f1-53fe57e1a7c8" chipset="acmgt1" name="Ext234" /> + <guid config_hash="c57d96d553d437bf0c0ba4b7d48541ba" mdapi_config_hash="0c084157d183b56b77156c47e5b53cdd" id="7e29d0d3-2ec5-4716-90d4-917c7b3cbdca" chipset="acmgt1" name="Ext235" /> + <guid config_hash="83d6433f9db93399da432bc21e274d08" mdapi_config_hash="1678730ecfb5b5382eeff06f0423d719" id="ae42189a-f2a1-4494-84b0-f2d506decbc1" chipset="acmgt1" name="Ext236" /> + <guid config_hash="8a2219263d91fa782b154fdcc20c8246" mdapi_config_hash="e53c035736288a34d9689782f5d7411f" id="6f65878e-1b6b-4a3b-8107-837073903643" chipset="acmgt1" name="Ext237" /> + <guid config_hash="75645472bf7411c9ec4592746ac24bc4" mdapi_config_hash="1b655ed590a8b791f963d785157f5bdf" id="a1c8b5b0-7b8c-4dbc-a4dd-eb29f8055ab0" chipset="acmgt1" name="Ext238" /> + <guid config_hash="e6e1bf7b7add5f4a06e6efc76ddc1a91" mdapi_config_hash="a5d4b65a3d2eaaf68cf9f6a5f7b06d0f" id="cc574967-0ba6-4c22-801f-881b3974395a" chipset="acmgt1" name="Ext239" /> + <guid config_hash="f5fff228cac4a33a2ad3fa6cfad15324" mdapi_config_hash="3ec7dc1e09de82b25cb41587a64f6226" id="d96495d3-d734-4b58-bf54-ece1a4a1ab09" chipset="acmgt1" name="Ext240" /> + <guid config_hash="292301c8cd06bb2ace701f88fcb74444" mdapi_config_hash="0447383cbb462c27879caaf77e6b4c63" id="1c157760-179c-40d1-aa3c-b419f609ea79" chipset="acmgt1" name="Ext500" /> + <guid config_hash="c88166576d30bc0d817949d89eb4afdb" mdapi_config_hash="a195cf96c2537f74090ef890c69c0026" id="06451560-fa04-4f38-b502-e0a096965a64" chipset="acmgt1" name="Ext501" /> + <guid config_hash="73cf9d34be55b9fe9352ed160e17c4ca" mdapi_config_hash="dfc86f31ed91149ef9329bbda6a06612" id="9d21c47c-92e0-4454-b3f1-484f27763e83" chipset="acmgt1" name="Ext502" /> + <guid config_hash="f1968a06f8ac53a3866e020ce93c04dd" mdapi_config_hash="cddac67e3eee29285177057d4f40f929" id="3cbb77d0-15f3-4bd0-8db1-5644266d78fc" chipset="acmgt1" name="Ext503" /> + <guid config_hash="bf319d9b55185073e31d28d67bb5a29f" mdapi_config_hash="29db5209c3ead6c4e5c8fb0e65fe5e5a" id="e81aaa56-a54b-491f-9803-ed19d2762e13" chipset="acmgt1" name="Ext508" /> + <guid config_hash="1f75dc15ec32ae3bc3fef249a3815542" mdapi_config_hash="f5e6861401967100bdff48c423e1e688" id="3c10260c-cb15-4474-ab84-c226d54f2974" chipset="acmgt1" name="Ext509" /> + <guid config_hash="3ebbed84f247db40096c3ca56d89e40e" mdapi_config_hash="1bbe2727551b21d94cc77727f2da16b5" id="95455b22-bda5-4f6e-86ae-0e500f0b419e" chipset="acmgt1" name="Ext510" /> + <guid config_hash="a5c49c3ebdb51b9c7c3a9596e78f1dbd" mdapi_config_hash="8cc5a168503a2fac73ca37fc14e5f0e1" id="95137f7b-f35e-4bef-9949-3c6681813bb9" chipset="acmgt1" name="Ext511" /> + <guid config_hash="0dc11e8b51b55f875d6a910a62528b36" mdapi_config_hash="527497315e9ce87f6b27256ffbf7ba50" id="3effc828-e8ca-4aab-a1a6-c478a41d55a7" chipset="acmgt1" name="Ext512" /> + <guid config_hash="f5fc6bb49cd289c389bbcdd97f8f3e94" mdapi_config_hash="4e3656a8c49229794ba3c2a39025e1ad" id="42d607af-0d8c-4d80-8443-73f5952060bb" chipset="acmgt1" name="Ext513" /> + <guid config_hash="cc372fbb060ad700fb983bfa05cf81b0" mdapi_config_hash="36153177f4d2a1a4370a166316fba516" id="bc6702b9-80d9-463e-9699-334c9543e076" chipset="acmgt1" name="Ext514" /> + <guid config_hash="f39e8d6c423d66200d984a53f027227f" mdapi_config_hash="ff56cd2d3621b6a5d90681e26662ecbb" id="5a8af975-8ceb-4111-b13a-3584b5485afb" chipset="acmgt1" name="Ext515" /> + <guid config_hash="a1fafcde16ae6adf015cb61f5461e0e1" mdapi_config_hash="e781bf31f152aad4938ceb3aa787fca1" id="a0bd2654-8654-4aec-b9c3-609b7cfc7c29" chipset="acmgt1" name="Ext516" /> + <guid config_hash="545871b03bf96eab2b5d68025ef4666b" mdapi_config_hash="07797e0c662798a7521edcad32a7ae5f" id="56c67cc8-4bbb-44bc-8552-1d095f1800bc" chipset="acmgt1" name="Ext517" /> + <guid config_hash="a145c7b67316ffe803202c00937e3180" mdapi_config_hash="1e00466df9653fd5b1e59a5bc0a0c490" id="f4f45c2b-4810-43db-a47d-cb854884f724" chipset="acmgt1" name="Ext518" /> + <guid config_hash="acf3f4ab24f3181ada2596c8c4b33323" mdapi_config_hash="71b7599209e7f623ffbf77e6e97dc404" id="58283b64-3a38-4985-b19c-db349947311a" chipset="acmgt1" name="Ext519" /> + <guid config_hash="c9262ebe1a9085ba8e17f0b56df167eb" mdapi_config_hash="9a86263ab0c5344708b56523db530e59" id="144cc1f2-ff34-4ce7-8bb9-7cb6cee76403" chipset="acmgt1" name="Ext520" /> + <guid config_hash="cc0cbd19cd8396daf31a4c43f0c9b72c" mdapi_config_hash="1c0c9dd022f0a8ba459a765b2a84ebce" id="9b5b6101-2131-4069-ace5-bcb3300991cb" chipset="acmgt1" name="Ext521" /> + <guid config_hash="d3ac23113c6c78bfa1e4c9ba272a2509" mdapi_config_hash="4a63c3f2bd5e5fb13d2800f11dc7740f" id="b3cfc81d-5297-483d-8e45-b77e4935fbe9" chipset="acmgt1" name="Ext522" /> + <guid config_hash="6bb65219794dfb6a1f1e2d6d7fc878ac" mdapi_config_hash="3e1b4888a83beba85489969de7cf1763" id="1975278e-2c3f-49a3-891a-fc2ed8f22eec" chipset="acmgt1" name="Ext523" /> + <guid config_hash="30d289b4c979c42d54a45a4ff1af0519" mdapi_config_hash="f5bd3bc0bca95bfce98199f8a5cf90ba" id="06d3fa26-617a-4443-a552-17937fc2bc83" chipset="acmgt1" name="Ext524" /> + <guid config_hash="9b25a3c3dcd18d39c811738c9bb80a5b" mdapi_config_hash="47c0e13155017f90a141cad97dce055b" id="5115960e-f55c-4854-bac4-b771c4185270" chipset="acmgt1" name="Ext525" /> + <guid config_hash="92109d4b551d47f1ebddb92af9577905" mdapi_config_hash="5ed607edcf09e43720510b366d5164ae" id="4dc30759-6119-4bf0-85b7-4a7cd5e79956" chipset="acmgt1" name="Ext526" /> + <guid config_hash="d348c04a69800278e93ce696589a8ef5" mdapi_config_hash="a5751a259780e1c128f81fb58d73b454" id="2060bec5-9739-4970-ab77-09dd7c107044" chipset="acmgt1" name="Ext527" /> + <guid config_hash="7272e30b65a50a446bba9e9d4e6c38e2" mdapi_config_hash="0ede2625ff325abc9556c2f79f135df2" id="91384c6e-5087-43d6-a32f-699690ca8917" chipset="acmgt1" name="Ext528" /> + <guid config_hash="e7ebfa8f96d5e8d3fe943fa862a475ed" mdapi_config_hash="f4d468897cdd9d7618c9c5c890864787" id="cb4f7554-9a12-4f83-a7c6-0ff8911f9e73" chipset="acmgt1" name="Ext529" /> + <guid config_hash="e2c841acf140e73947b4fde1b3637340" mdapi_config_hash="6a9d00819e14fb3178878b66a648ad92" id="80502c88-9c1e-4804-8d54-53b3d784a540" chipset="acmgt1" name="Ext530" /> + <guid config_hash="65e598c36f24fbbed61b9f29f22906b6" mdapi_config_hash="4318133904297748c76a038f742268ff" id="eccbb94d-f7ed-4d45-89bb-d17a0e28e97d" chipset="acmgt1" name="Ext531" /> + <guid config_hash="31a9994c9523d67e612a89860956a748" mdapi_config_hash="562ded5c070fd64bd3254155d782e228" id="09cb4790-899a-456f-9c12-f0b5ae36bf39" chipset="acmgt1" name="Ext532" /> + <guid config_hash="06b0b6e00cadfe410e4a16ac268dc4d1" mdapi_config_hash="26cffa7b241feebff27888e6228687be" id="e257595b-2e38-4da5-b375-bc494b4d0ee8" chipset="acmgt1" name="Ext533" /> + <guid config_hash="ef8296994d00e8de953b430dc2336936" mdapi_config_hash="b46f8af2ff32b1cb0a148ab25607dcc1" id="f2477c56-db39-43cf-bd4c-42782dd02e27" chipset="acmgt1" name="Ext534" /> + <guid config_hash="597d72f1b88eb636d7e4947f55126a7c" mdapi_config_hash="6da98998f96cc35043ecda01a28bf3eb" id="b6417928-9c95-4df7-88a2-e6b8de52abec" chipset="acmgt1" name="Ext535" /> + <guid config_hash="ea58095d9149eefb3b46afd33b27c94c" mdapi_config_hash="b7662b81a69c2318aa199eef8947eae5" id="ac88f2a7-2f38-4e61-a6d5-9da2d50edd38" chipset="acmgt1" name="Ext536" /> + <guid config_hash="38307a56bf07276dd8d3d07eccdaadff" mdapi_config_hash="beae875603308037baef26cb4e56ea41" id="87640cd9-04b0-4b76-a016-308ab7975f8d" chipset="acmgt1" name="Ext537" /> + <guid config_hash="8c246b3c51760bc56021dcca4d98b64f" mdapi_config_hash="e9ffa7ab3397a2078ea72ca72b648042" id="2a2f8a81-b046-4a67-92c3-2ded36755f76" chipset="acmgt1" name="Ext538" /> + <guid config_hash="cc5c545e4579215cd5e1daf4f8ccf598" mdapi_config_hash="c06b0cfe2ef7f7d0ee15d5d9d59f04af" id="89f03450-0e5f-4370-b7cf-cea4e627d4e8" chipset="acmgt1" name="Ext539" /> + <guid config_hash="9cd0d55e6f24cccaf43362e66846110c" mdapi_config_hash="8bbe7c98aee8c4e13cd4bcfc88356089" id="85c7638d-976e-47e5-8f52-2079a0ea147f" chipset="acmgt1" name="Ext540" /> + <guid config_hash="4cffd8f8e3ce9cf708b9930bfefa6d64" mdapi_config_hash="51c0d733753392b68f9f05eb6eed2ad6" id="69645853-2e7f-4a77-83fc-fa03f43b5d35" chipset="acmgt1" name="Ext541" /> + <guid config_hash="86fe7c5b0f79e2f195b8b07b60f2edae" mdapi_config_hash="579fe3e074f48c52eac60403cf5a747e" id="657218aa-a73a-4127-b081-07ece46ff786" chipset="acmgt1" name="Ext542" /> + <guid config_hash="5cc9510caf7275ff5cc7e61a319234a4" mdapi_config_hash="ae5c2ef4361cc50202b1f4d3283ac607" id="1a51853a-37a2-4020-8ccb-650604ea4f64" chipset="acmgt1" name="Ext543" /> + <guid config_hash="5afad4a36d651c8ed2d343a4ee64656a" mdapi_config_hash="969e77a6b883f3e4bd11597003c0b3e7" id="9aa2e068-b13a-44b2-be93-a4cf68924cdd" chipset="acmgt1" name="Ext544" /> + <guid config_hash="5b9bf405632b8d14a5fd2de009a6617f" mdapi_config_hash="770396733fa2c5f888dedc1ab959dfa9" id="5e8c65a2-97ac-4dea-af1b-27b9f12dafe6" chipset="acmgt1" name="Ext545" /> + <guid config_hash="648a597b7183dd3b493c029db7e9cb30" mdapi_config_hash="cf2390eef15a57224d0678f25b975c7d" id="192ec14b-c329-4e81-bd2d-02f7cae24579" chipset="acmgt1" name="Ext546" /> + <guid config_hash="95fbff94cc3a238e65c0443caad2f6a0" mdapi_config_hash="4aa2ab70084a111471c3f5427ebbe18a" id="2e5ac978-b4e4-4693-8f88-d90d0353dd77" chipset="acmgt1" name="Ext547" /> + <guid config_hash="40963244f7b26a91769f85cb64b16851" mdapi_config_hash="515ab3d31821af95b14941f048add55e" id="7c95ba0f-cd85-4eb2-ab43-987f84214820" chipset="acmgt1" name="Ext548" /> + <guid config_hash="472bcdb97493000caf3f593414f08b48" mdapi_config_hash="00d5e25a79e41ada6a22f2ec515c45c1" id="8fa03693-d2d5-42eb-a382-e2cf05bf8197" chipset="acmgt1" name="Ext549" /> + <guid config_hash="d837f394515cef5ab41e20506bbd5896" mdapi_config_hash="937ccad99681cb98faf0112bb88e5be3" id="3714122b-03ee-499b-bc5f-31ed5400924e" chipset="acmgt1" name="Ext550" /> + <guid config_hash="4e359a4c6c5bfbb90c11c8385f70f38e" mdapi_config_hash="69d945906a10272834a2df243e358111" id="31ed9c46-b219-4bcc-8c20-22586fb077f8" chipset="acmgt1" name="Ext551" /> + <guid config_hash="7be9f83be5e14c74f84886dd895fdd9b" mdapi_config_hash="677ec46b2d65b89ca5bf4a34724bdff7" id="f639c566-949f-4484-8c12-5127fb771b4c" chipset="acmgt1" name="Ext552" /> + <guid config_hash="acbb06143814fa672b47b2e9af68bb15" mdapi_config_hash="7fcaf7e5cf736ffb51ab7e28beba8e83" id="40b20c9d-0935-4000-91d7-f312a515bb4b" chipset="acmgt1" name="Ext553" /> + <guid config_hash="dc483448783ecee75c50c4c35c20c759" mdapi_config_hash="de26db366e2acbbe24add896075500f3" id="bcff8729-44e4-4efa-8436-f1775fcb204a" chipset="acmgt1" name="Ext554" /> + <guid config_hash="ffecd22693e6aa74e47e1e09c65f697a" mdapi_config_hash="5d7af72d16b3bf039bbcd2776e6636cd" id="6bf30c0e-a169-4ba3-8527-047d4f691aa3" chipset="acmgt1" name="Ext555" /> + <guid config_hash="57e90f0cc7ef3b9c7ca4f75c4803672c" mdapi_config_hash="c22c5d4f2faa7d9da87b62ad67f616fe" id="bc1cda24-8e1b-4000-9443-1135be1b976e" chipset="acmgt1" name="Ext556" /> + <guid config_hash="e682eeada8bbcb893050e4f2c2e1bada" mdapi_config_hash="51185678325b66968b2652bdffd0bb3d" id="c8a172f8-4a06-4931-b877-bc8aa687a9b8" chipset="acmgt1" name="Ext557" /> + <guid config_hash="ec4ed96d22de43c663680a06f4f43ea3" mdapi_config_hash="1ab2110a870c4a69e2b81d9215735bd5" id="a5a9b5ed-c78c-4e3d-a808-22f1debd6b43" chipset="acmgt1" name="Ext558" /> + <guid config_hash="145db1c6e054541b6aee3bf6fd06137f" mdapi_config_hash="19bab8aa5d3397e7ddfcf2e9f3fd6227" id="93bba39f-4d65-434b-9251-37f2badf256e" chipset="acmgt1" name="Ext559" /> + <guid config_hash="344f255d653846067d34bf4e4530ff27" mdapi_config_hash="d6d5153a13d55895f89b5fb0a064d035" id="a16d5961-ab03-4193-87d9-394ad9c2da80" chipset="acmgt1" name="Ext560" /> + <guid config_hash="df565eb53e7967860a4cc406bbb974a8" mdapi_config_hash="9861b6b99c31de2d8d511fc2b7fc173e" id="8dfd3f4a-0686-48bb-94cf-76a3bc5386ec" chipset="acmgt1" name="Ext561" /> + <guid config_hash="9c8a7c975bda5e68d705bd62afe9acff" mdapi_config_hash="041925e9e42ea2aadee504a8d552d1f2" id="f41e3a7c-de83-46af-8be2-b26c605d1317" chipset="acmgt1" name="Ext562" /> + <guid config_hash="d3c5828116e850f0b8ca67a0d37c108a" mdapi_config_hash="fa5f272f45c3353007b20560eb4b8d8e" id="c599736d-1b09-4ae6-9e82-c98d17c3582a" chipset="acmgt1" name="Ext563" /> + <guid config_hash="ddb0cf4a3ebc5bad4975b834ce0141f5" mdapi_config_hash="73e86e4e676b7fb8fc68ae1f48b060c0" id="0f9c5fb0-e5bc-4119-a11f-d32d0304bca8" chipset="acmgt1" name="Ext564" /> + <guid config_hash="7b22bd29eb16da8a28daa526677ea4f8" mdapi_config_hash="c3a83d8d29ff0be5d617f62eae2983bf" id="def2b365-d8d1-4e4b-b832-39afd5e214d9" chipset="acmgt1" name="Ext565" /> + <guid config_hash="e6b25c38dbdafbf3d8cba9d47b96758e" mdapi_config_hash="0c3221607ea6779e1e1366677702d2eb" id="a53e91c2-03b4-488a-ab7e-293f4466e00f" chipset="acmgt1" name="Ext566" /> + <guid config_hash="4dac49ea134b9be63fadad389dd16395" mdapi_config_hash="5166a562ec90ffe5b13f8cae2b5b6ab2" id="0cdefa4f-f4ac-4072-a021-be48ed5b0ee7" chipset="acmgt1" name="Ext567" /> + <guid config_hash="f25e194664e6448e7fbd558b5477fbb2" mdapi_config_hash="38d493f8bfdaf461db090fb5a1d3f6a9" id="7404a462-f9c0-4fb6-9ec9-b03bbd5e4572" chipset="acmgt1" name="Ext568" /> + <guid config_hash="2e3dae9aea38ec547c6afafcca82dc2c" mdapi_config_hash="6d392143f3ec1e8940b9c5c676261968" id="b3515130-788c-4d36-b20d-e8ca02e11cf1" chipset="acmgt1" name="Ext569" /> + <guid config_hash="7c0b2e864e654ca268d58ba7a4e36e0c" mdapi_config_hash="fde6da1eeefa44f3bd95b0c8876b2903" id="73e5a9f9-ca88-4467-b407-737e7c7080fe" chipset="acmgt1" name="Ext570" /> + <guid config_hash="baad6f80facbd4d135de5c9393d9c4f6" mdapi_config_hash="ad060f6502ef3021d523aff45a4952cd" id="ae2fa4f6-2afb-4ac9-866e-bfd943f91a54" chipset="acmgt1" name="Ext571" /> + <guid config_hash="6389c2dcb6a2b100db74a782d4f7006f" mdapi_config_hash="d8d0fcb20772d25731d0ac8a7d587cb9" id="cea714fc-17a9-4dae-b3e8-8be0f3149918" chipset="acmgt1" name="Ext1000" /> + <guid config_hash="e358e6156c7379f8064b679128fe0747" mdapi_config_hash="a9988177c98737d0b52bcf215079e95b" id="3ec074a8-e3cb-4090-86c0-0d1426ac9a63" chipset="acmgt1" name="Ext1001" /> + <guid config_hash="c4809b4f9dbd405691ebd29dea153aa2" mdapi_config_hash="8d1002c55a03d16517de67b4b8970d12" id="17275fa7-be46-48a6-a654-6931ecbfed77" chipset="acmgt1" name="Ext1002" /> + <guid config_hash="8bb1139fff1ee995a5ade92734e11d8a" mdapi_config_hash="fbbbc2eaec0f182e838cbdfa1c8372c2" id="a56f028c-27c7-4e9b-bc19-04bbf67444b0" chipset="acmgt1" name="Ext1004" /> + <guid config_hash="83d7b8e952e3598565267c65964b2771" mdapi_config_hash="7a5482c523806abe25c8c4dbcf5273e7" id="f43c884f-817d-4037-bc96-7bed64906683" chipset="acmgt1" name="Ext1005" /> + <guid config_hash="e7efe91c7aeb591043358c1b5e3c019d" mdapi_config_hash="94ecec147a25de742814fbb01fd4e9c3" id="6a6e8e1d-6148-4316-95f9-2a8b5236520f" chipset="acmgt1" name="Ext1006" /> + <guid config_hash="ae85ee9664f10b42cc854c26878d6cba" mdapi_config_hash="de92fab5a4df472219094e6a3b635632" id="b6b53888-0a0b-4fc6-b77e-c604875495a5" chipset="acmgt1" name="Ext1007" /> + <guid config_hash="310dddcb04d1088a3865f698d758cbc2" mdapi_config_hash="c6a3be7d49b5697074021ca673a39a57" id="6e06382a-3c98-464a-9811-cdfb956e4e8e" chipset="acmgt1" name="Ext1010" /> </guids>
lib/i915/perf-configs/oa-acmgt1.xml new file +76252 -0
2,651 lines not on this page · raw message
The rest of this diff is not shown. View the raw message.
Signature · 1 line
2.25.1
Part of the diff is not shown here. View the raw message.