perf cpumap: Retrieve die id information
PATCH v2 1/5 perf cpumap: Retrieve die id information
From: Kan Liang <[email protected]>
There is no function to retrieve die id information of a given CPU.
Add cpu_map__get_die_id() to retrieve die id information.
---
New patch for V2.
tools/perf/util/cpumap.c | 7 +++++++ tools/perf/util/cpumap.h | 1 + 2 files changed, 8 insertions(+)
tools/perf/util/cpumap.c +7 -0
@@ -373,6 +373,13 @@ int cpu_map__build_map(struct cpu_map *cpus, struct cpu_map **res,return 0 ; } +int cpu_map__get_die_id (int cpu) +{ +int value, ret =cpu__get_topology_int (cpu,"die_id" , &value); + +return ret ?: value; +} +int cpu_map__get_core_id (int cpu) {int value, ret =cpu__get_topology_int (cpu,"core_id" , &value);
tools/perf/util/cpumap.h +1 -0
@@ -25,6 +25,7 @@ size_t cpu_map__snprint_mask(struct cpu_map *map, char *buf, size_t size);size_t cpu_map__fprintf (struct cpu_map *map,FILE *fp);int cpu_map__get_socket_id (int cpu);int cpu_map__get_socket (struct cpu_map *map,int idx,void *data); +int cpu_map__get_die_id (int cpu);int cpu_map__get_core_id (int cpu);int cpu_map__get_core (struct cpu_map *map,int idx,void *data);int cpu_map__build_socket_map (struct cpu_map *cpus,struct cpu_map **sockp);
Signature · 1 line
2.7.4
PATCH v2 2/5 perf header: Add die information in CPU topology
From: Kan Liang <[email protected]>
With the new CPUID.1F, a new level type of CPU topology, 'die', is introduced. The 'die' information in CPU topology should be added in perf header.
To be compatible with old perf.data, the patch checks the section size before reading the die information. The new info is added at the end of the cpu_topology section, the old perf tool ignores the extra data. It never reads data crossing the section boundary.
The new perf tool with the patch can be used on legacy kernel. Add a new function has_die_topology() to check if die topology information is supported by kernel. The function only check X86 and CPU 0. Assuming other CPUs have same topology.
Use similar method for core and socket to support die id and sibling dies string.
---
Changes since V1:
- Avoid has_die parameters for many functions
- Move cpu_map__get_die_id() to separate patch
- Rename check_x86_die_exists() to has_die_topology()
tools/perf/Documentation/perf.data-file-format.txt | 9 ++- tools/perf/util/cputopo.c | 76 +++++++++++++++-- tools/perf/util/cputopo.h | 2 + tools/perf/util/env.c | 1 + tools/perf/util/env.h | 3 + tools/perf/util/header.c | 94 ++++++++++++++++++++-- 6 files changed, 172 insertions(+), 13 deletions(-)
tools/perf/Documentation/perf.data-file-format.txt +7 -2
@@ -153,7 +153,7 @@ struct { String lists defining the core and CPU threads topology. The string lists are followed by a variable length array -which contains core_id and socket_id of each cpu. +which contains core_id, die_id (for x86) and socket_id of each cpu. The number of entries can be determined by the size of the section minus the sizes of both string lists.
@@ -162,14 +162,19 @@ struct { struct perf_header_string_list threads; /* Variable length */ struct { uint32_t core_id; + uint32_t die_id; uint32_t socket_id; } cpus[nr]; /* Variable length records */ }; Example: - sibling cores : 0-3 + sibling cores : 0-8 + sibling dies : 0-3 + sibling dies : 4-7 sibling threads : 0-1 sibling threads : 2-3 + sibling threads : 4-5 + sibling threads : 6-7 HEADER_NUMA_TOPOLOGY = 14,
tools/perf/util/cputopo.c +71 -5
137 lines not on this page · raw message
tools/perf/util/cputopo.h +2 -0
@@ -7,8 +7,10 @@struct cpu_topology {u32 core_sib; +u32 die_sib;u32 thread_sib;char **core_siblings; +char **die_siblings;char **thread_siblings; };
tools/perf/util/env.c +1 -0
@@ -246,6 +246,7 @@ int perf_env__read_cpu_topology_map(struct perf_env *env)for (cpu =0 ; cpu < nr_cpus; ++cpu) { env->cpu[cpu].core_id =cpu_map__get_core_id (cpu); env->cpu[cpu].socket_id =cpu_map__get_socket_id (cpu); + env->cpu[cpu].die_id =cpu_map__get_die_id (cpu); } env->nr_cpus_avail = nr_cpus;
tools/perf/util/env.h +3 -0
@@ -9,6 +9,7 @@struct cpu_topology_map {int socket_id; +int die_id;int core_id; };
@@ -49,6 +50,7 @@ struct perf_env {int nr_cmdline;int nr_sibling_cores; +int nr_sibling_dies;int nr_sibling_threads;int nr_numa_nodes;int nr_memory_nodes;
@@ -57,6 +59,7 @@ struct perf_env {char *cmdline;const char **cmdline_argv;char *sibling_cores; +char *sibling_dies;char *sibling_threads;char *pmu_mappings;struct cpu_topology_map *cpu;
tools/perf/util/header.c +88 -6
124 lines not on this page · raw message
Signature · 1 line
2.7.4
PATCH v2 3/5 perf stat: Support per-die aggregation
In reply to [email protected]From: Kan Liang <[email protected]>
It is useful to aggregate counts per die. E.g. Uncore becomes die-scope on Xeon Cascade Lake-AP.
Introduce a new option "--per-die" to support per-die aggregation.
The global id for each core has been changed to socket + die id + core id. The global id for each die is socket + die id.
Add die information for per-core aggregation. The output of per-core aggregation will be changed from "S0-C0" to "S0-D0-C0". Any scripts which rely on the output format of per-core aggregation probably be broken.
For perf stat record/report, there is no die information when processing the old perf.data. The per-die result will be the same as per-socket.
---
Changes since V1
- Fix the title shift for the case of "--per-die -a -I 1000"
- Add warn to check the size of socket id and die id
tools/perf/Documentation/perf-stat.txt | 10 ++++ tools/perf/builtin-stat.c | 85 +++++++++++++++++++++++++++++++--- tools/perf/util/cpumap.c | 60 ++++++++++++++++++++++-- tools/perf/util/cpumap.h | 9 +++- tools/perf/util/stat-display.c | 27 +++++++++-- tools/perf/util/stat-shadow.c | 1 + tools/perf/util/stat.c | 1 + tools/perf/util/stat.h | 1 + 8 files changed, 179 insertions(+), 15 deletions(-)
tools/perf/Documentation/perf-stat.txt +10 -0
@@ -200,6 +200,13 @@ use --per-socket in addition to -a. (system-wide). The output includes the socket number and the number of online processors on that socket. This is useful to gauge the amount of aggregation. +--per-die:: +Aggregate counts per processor die for system-wide mode measurements. This +is a useful mode to detect imbalance between dies. To enable this mode, +use --per-die in addition to -a. (system-wide). The output includes the +die number and the number of online processors on that die. This is +useful to gauge the amount of aggregation. + --per-core:: Aggregate counts per physical processor for system-wide mode measurements. This is a useful mode to detect imbalance between physical cores. To enable this mode,
@@ -239,6 +246,9 @@ Input file name. --per-socket:: Aggregate counts per processor socket for system-wide mode measurements. +--per-die:: +Aggregate counts per processor die for system-wide mode measurements. + --per-core:: Aggregate counts per physical processor for system-wide mode measurements.
tools/perf/builtin-stat.c +79 -6
148 lines not on this page · raw message
tools/perf/util/cpumap.c +55 -5
91 lines not on this page · raw message
tools/perf/util/cpumap.h +8 -1
@@ -26,9 +26,11 @@ size_t cpu_map__fprintf(struct cpu_map *map, FILE *fp);int cpu_map__get_socket_id (int cpu);int cpu_map__get_socket (struct cpu_map *map,int idx,void *data);int cpu_map__get_die_id (int cpu); +int cpu_map__get_die (struct cpu_map *map,int idx,void *data);int cpu_map__get_core_id (int cpu);int cpu_map__get_core (struct cpu_map *map,int idx,void *data);int cpu_map__build_socket_map (struct cpu_map *cpus,struct cpu_map **sockp); +int cpu_map__build_die_map (struct cpu_map *cpus,struct cpu_map **diep);int cpu_map__build_core_map (struct cpu_map *cpus,struct cpu_map **corep);const struct cpu_map *cpu_map__online (void );/* thread unsafe */
@@ -44,7 +46,12 @@ static inline int cpu_map__socket(struct cpu_map *sock, int s)static inline int cpu_map__id_to_socket (int id) { -return id >>16 ; +return id >>24 ; +} + +static inline int cpu_map__id_to_die (int id) +{ +return (id >>16 ) &0xff ; }static inline int cpu_map__id_to_cpu (int id)
tools/perf/util/stat-display.c +24 -3
@@ -69,8 +69,9 @@ static void aggr_printout(struct perf_stat_config *config, {switch (config->aggr_mode) {case AGGR_CORE : - fprintf(config->output, "S%d-C%*d%s%*d%s", + fprintf(config->output, "S%d-D%d-C%*d%s%*d%s", cpu_map__id_to_socket(id), + cpu_map__id_to_die(id), config->csv_output ? 0 : -8, cpu_map__id_to_cpu(id), config->csv_sep,
@@ -78,6 +79,16 @@ static void aggr_printout(struct perf_stat_config *config, nr, config->csv_sep);break ; +case AGGR_DIE : +fprintf (config->output,"S%d-D%*d%s%*d%s" , +cpu_map__id_to_socket (id <<16 ), + config->csv_output ?0 :-8 , +cpu_map__id_to_die (id <<16 ), + config->csv_sep, + config->csv_output ?0 :4 , + nr, + config->csv_sep); +break ;case AGGR_SOCKET : fprintf(config->output, "S%*d%s%*d%s", config->csv_output ? 0 : -5,
@@ -89,8 +100,9 @@ static void aggr_printout(struct perf_stat_config *config,break ;case AGGR_NONE :if (evsel->percore) { -fprintf (config->output,"S%d-C%*d%s" , +fprintf (config->output,"S%d-D%d-C%*d%s" ,cpu_map__id_to_socket (id), +cpu_map__id_to_die (id), config->csv_output ?0 :-5 ,cpu_map__id_to_cpu (id), config->csv_sep); }else {
@@ -407,6 +419,7 @@ static void printout(struct perf_stat_config *config, int id, int nr, [AGGR_THREAD] = 1, [AGGR_NONE] = 1, [AGGR_SOCKET] = 2, + [AGGR_DIE] = 2, [AGGR_CORE] = 2, };
@@ -879,7 +892,8 @@ static void print_no_aggr_metric(struct perf_stat_config *config, }static int aggr_header_lens[] = { - [AGGR_CORE ] =18 , + [AGGR_CORE ] =24 , + [AGGR_DIE ] =18 , [AGGR_SOCKET ] =12 , [AGGR_NONE ] =6 , [AGGR_THREAD ] =24 ,
@@ -888,6 +902,7 @@ static int aggr_header_lens[] = {static const char *aggr_header_csv[] = { [AGGR_CORE ] ="core,cpus," , + [AGGR_DIE ] ="die,cpus" , [AGGR_SOCKET ] ="socket,cpus" , [AGGR_NONE ] ="cpu," , [AGGR_THREAD ] ="comm-pid," ,
@@ -954,6 +969,11 @@ static void print_interval(struct perf_stat_config *config,if (!metric_only)fprintf (output," counts %*s events\n" , unit_width,"unit" );break ; +case AGGR_DIE : +fprintf (output,"# time die cpus" ); +if (!metric_only) +fprintf (output," counts %*s events\n" , unit_width,"unit" ); +break ;case AGGR_CORE :fprintf (output,"# time core cpus" );if (!metric_only)
@@ -1165,6 +1185,7 @@ perf_evlist__print_counters(struct perf_evlist *evlist,switch (config->aggr_mode) {case AGGR_CORE : +case AGGR_DIE :case AGGR_SOCKET :print_aggr (config, evlist, prefix);break ;
tools/perf/util/stat-shadow.c +1 -0
@@ -12,6 +12,7 @@/* * AGGR_GLOBAL: Use CPU 0 * AGGR_SOCKET: Use first CPU of socket +* AGGR_DIE: Use first CPU of die * AGGR_CORE: Use first CPU of core * AGGR_NONE: Use matching CPU * AGGR_THREAD: Not supported?
tools/perf/util/stat.c +1 -0
@@ -272,6 +272,7 @@ process_counter_values(struct perf_stat_config *config, struct perf_evsel *evselswitch (config->aggr_mode) {case AGGR_THREAD :case AGGR_CORE : +case AGGR_DIE :case AGGR_SOCKET :case AGGR_NONE :if (!evsel->snapshot)
tools/perf/util/stat.h +1 -0
@@ -44,6 +44,7 @@ enum aggr_mode {AGGR_NONE ,AGGR_GLOBAL ,AGGR_SOCKET , +AGGR_DIE ,AGGR_CORE ,AGGR_THREAD , AGGR_UNSET,
Signature · 1 line
2.7.4
Re: [PATCH V2 3/5] perf stat: Support per-die aggregation
On Thu, May 30, 2019 at 07:53:47AM -0700, [email protected] wrote:
SNIP
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 990
Quoted 13 lines
tools/perf/builtin-stat.c +2
-4+static int perf_env__get_core (struct cpu_map *map,int idx,void *data) {struct perf_env *env = data;int core =-1 , cpu =perf_env__get_cpu (env, map, idx);if (cpu !=-1 ) { -int socket_id = env->cpu[cpu].socket_id; -/* -* Encode socket in upper 16 bits -* core_id is relative to socket, and +* Encode socket in upper 24 bits
please note we use upper 8 bits for socket number, the comments suggests it's 24 bits
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 997
Quoted 9 lines
tools/perf/builtin-stat.c +5
-2+* encode die id in upper 16 bits +* core_id is relative to socket and die, * we need a global id. So we combine -* socket + core id. +* socket + die id + core id */ - core = (socket_id <<16 ) | (env->cpu[cpu].core_id &0xffff ); +if (WARN_ONCE (env->cpu[cpu].socket_id >>8 , +"The socket_id number is too big. Please upgrade the perf tool.\n" ))
hum, how's perf tool upgrade going to help in here?
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 1007
Quoted 10 lines
tools/perf/builtin-stat.c +9
-0+return -1 ; + +if (WARN_ONCE (env->cpu[cpu].die_id >>8 , +"The die_id number is too big. Please upgrade the perf tool.\n" )) +return -1 ; + + core = (env->cpu[cpu].socket_id <<24 ) | + (env->cpu[cpu].die_id <<16 ) | + (env->cpu[cpu].core_id &0xffff ); }
other than comments above, the patchset looks good to me
thanks,
jirka
Re: [PATCH V2 3/5] perf stat: Support per-die aggregation
Quoted 10 lines · Jiri Olsa
On Thu, May 30, 2019 at 07:53:47AM -0700, [email protected] wrote:
SNIP
Quoted 6 lines
+
static intperf_env__get_core(struct cpu_map *map, int idx, void *data){
structperf_env*env = data;
int core = -1, cpu =perf_env__get_cpu(env, map, idx);
Quoted 10 lines
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 990
Quoted 7 lines
if (cpu != -1) {tools/perf/builtin-stat.c +1
-4-int socket_id = env->cpu[cpu].socket_id; -/* -* Encode socket in upper 16 bits -* core_id is relative to socket, and +* Encode socket in upper 24 bits please note we use upper 8 bits for socket number, the comments suggests it's 24 bits
I will fix it in v3.
Quoted 12 lines
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 997
Quoted 10 lines
tools/perf/builtin-stat.c +5
-2+* encode die id in upper 16 bits +* core_id is relative to socket and die, * we need a global id. So we combine -* socket + core id. +* socket + die id + core id */ - core = (socket_id <<16 ) | (env->cpu[cpu].core_id &0xffff ); +if (WARN_ONCE (env->cpu[cpu].socket_id >>8 , +"The socket_id number is too big. Please upgrade the perf tool.\n" ))hum, how's perf tool upgrade going to help in here?
I think there is nothing we can do for current encoding, if the socket number or die number is bigger than 8 bits. We have to design a new encoding method, which needs to update the perf tool. So I use a similar expression from process_cpu_topology().
if (do_core_id_test && nr != (u32)-1 && nr > (u32)cpu_nr) {
pr_debug("socket_id number is too big."
"You may need to upgrade the perf tool.\n");
goto free_cpu;
}Any suggestions for the warning message?
Quoted 14 lines
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 1007
Quoted 11 lines
tools/perf/builtin-stat.c +9
-0+return -1 ; + +if (WARN_ONCE (env->cpu[cpu].die_id >>8 , +"The die_id number is too big. Please upgrade the perf tool.\n" )) +return -1 ; + + core = (env->cpu[cpu].socket_id <<24 ) | + (env->cpu[cpu].die_id <<16 ) | + (env->cpu[cpu].core_id &0xffff ); }other than comments above, the patchset looks good to me
Thanks for the review. I will send out V3 after the comments are addressed.
Thanks,
Kan
Re: [PATCH V2 3/5] perf stat: Support per-die aggregation
Quoted 43 lines · Kan
Quoted 4 lines · Jiri Olsa
On Thu, May 30, 2019 at 07:53:47AM -0700, [email protected] wrote:
SNIP
Quoted 15 lines
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 990
Quoted 12 lines
tools/perf/builtin-stat.c +2
-4+static int perf_env__get_core (struct cpu_map *map,int idx,void *data) {struct perf_env *env = data;int core =-1 , cpu =perf_env__get_cpu (env, map, idx);if (cpu !=-1 ) { -int socket_id = env->cpu[cpu].socket_id; -/* -* Encode socket in upper 16 bits -* core_id is relative to socket, and +* Encode socket in upper 24 bits please note we use upper 8 bits for socket number, the comments suggests it's 24 bits
I will fix it in v3.
Quoted 12 lines
tools/perf/builtin-stat.c static int perf_env__get_socket(struct cpu_map *map, int idx, void *data) · line 997
Quoted 10 lines
tools/perf/builtin-stat.c +5
-2+* encode die id in upper 16 bits +* core_id is relative to socket and die, * we need a global id. So we combine -* socket + core id. +* socket + die id + core id */ - core = (socket_id <<16 ) | (env->cpu[cpu].core_id &0xffff ); +if (WARN_ONCE (env->cpu[cpu].socket_id >>8 , +"The socket_id number is too big. Please upgrade the perf tool.\n" ))hum, how's perf tool upgrade going to help in here?
I think there is nothing we can do for current encoding, if the socket number or die number is bigger than 8 bits. We have to design a new encoding method, which needs to update the perf tool.
So I use a similar expression fromprocess_cpu_topology().
Quoted 7 lines
if (do_core_id_test && nr != (u32)-1 && nr > (u32)cpu_nr) { pr_debug("socket_id number is too big." "You may need to upgrade the perf tool.\n"); goto free_cpu; }Any suggestions for the warning message?
not really :-\ perhaps the original is more accurate and does not "claim" the upgrade will help.. it's just a nit.. feel free to keep it
thanks,
jirka
PATCH v2 4/5 perf header: Rename "sibling cores" to "sibling sockets"
In reply to [email protected]From: Kan Liang <[email protected]>
The "sibling cores" actually shows the sibling CPUs of a socket. The name "sibling cores" is very misleading.
Rename "sibling cores" to "sibling sockets"
---
No changes since V1
tools/perf/Documentation/perf.data-file-format.txt | 2 +- tools/perf/util/header.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
tools/perf/Documentation/perf.data-file-format.txt +1 -1
@@ -168,7 +168,7 @@ struct { }; Example: - sibling cores : 0-8 + sibling sockets : 0-8 sibling dies : 0-3 sibling dies : 4-7 sibling threads : 0-1
tools/perf/util/header.c +1 -1
@@ -1460,7 +1460,7 @@ static void print_cpu_topology(struct feat_fd *ff, FILE *fp) str = ph->env.sibling_cores;for (i =0 ; i < nr; i++) { -fprintf (fp,"# sibling cores : %s\n" , str); +fprintf (fp,"# sibling sockets : %s\n" , str); str +=strlen (str) +1 ; }
Signature · 1 line
2.7.4
PATCH v2 5/5 perf tools: Apply new CPU topology sysfs attributes
In reply to [email protected]From: Kan Liang <[email protected]>
The existing "thread_siblings" and "thread_siblings_list" attribute will be deprecated.
Use the new CPU topology sysfs attributes, "core_cpus" and "core_cpus_list", which are synonymous with the deprecated attributes.
Check the new name first. If not available, use the deprecated name to be compatible with old kernel.
---
New for V2
tools/perf/util/cputopo.c | 8 +++++++- tools/perf/util/smt.c | 8 ++++++-- 2 files changed, 13 insertions(+), 3 deletions(-)
tools/perf/util/cputopo.c +7 -1
@@ -15,6 +15,8 @@ "%s/devices/system/cpu/cpu%d/topology/die_cpus_list"#define THRD_SIB_FMT \ "%s/devices/system/cpu/cpu%d/topology/thread_siblings_list" +#define THRD_SIB_FMT_NEW \ + "%s/devices/system/cpu/cpu%d/topology/core_cpus_list"#define NODE_ONLINE_FMT \ "%s/devices/system/node/online"#define NODE_MEMINFO_FMT \
@@ -91,8 +93,12 @@ static int build_cpu_topology(struct cpu_topology *tp, int cpu) ret =0 ; try_threads: -scnprintf (filename,MAXPATHLEN ,THRD_SIB_FMT , +scnprintf (filename,MAXPATHLEN ,THRD_SIB_FMT_NEW ,sysfs__mountpoint (), cpu); +if (access (filename,F_OK ) ==-1 ) { +scnprintf (filename,MAXPATHLEN ,THRD_SIB_FMT , +sysfs__mountpoint (), cpu); + } fp =fopen (filename,"r" );if (!fp) goto done;
tools/perf/util/smt.c +6 -2
@@ -23,8 +23,12 @@ int smt_on(void)char fn[256 ];snprintf (fn,sizeof fn, -"devices/system/cpu/cpu%d/topology/thread_siblings" , - cpu); +"devices/system/cpu/cpu%d/topology/core_cpus" , cpu); +if (access (fn,F_OK ) ==-1 ) { +snprintf (fn,sizeof fn, +"devices/system/cpu/cpu%d/topology/thread_siblings" , + cpu); + }if (sysfs__read_str (fn, &str, &strlen) <0 )continue ;/* Entry is hex, but does not have 0x, so need custom parser */
Signature · 1 line
2.7.4