Skip to contentSkip to conversation

perf vendor events intel: Update free running alderlake events

PATCH Applied Ian Rogers 2 participants 6 messages Apr 7, 2023 perf events

lore
Share & tools
lore

Ian Rogers

author reviewer
#0 lore raw

PATCH v2 1/5 perf vendor events intel: Update free running alderlake events

Fix the PMU name, event code and umask.

These updates were generated by:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
with this PR:
https://github.com/intel/perfmon/pull/66

---
 .../arch/x86/alderlake/uncore-memory.json        | 16 ++++++++++++----
 .../arch/x86/alderlaken/uncore-memory.json       | 16 ++++++++++++----
 2 files changed, 24 insertions(+), 8 deletions(-)
tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json +12 -4
index 2ccd9cf96957..163d7e7755c4 100644
@@ -1,29 +1,37 @@
 [
     {
         "BriefDescription": "Counts every 64B read  request entering the Memory Controller 0 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read entering the Memory Controller 1 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "ACT command for a read request sent to DRAM",
tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json +12 -4
index 2ccd9cf96957..163d7e7755c4 100644
@@ -1,29 +1,37 @@
 [
     {
         "BriefDescription": "Counts every 64B read  request entering the Memory Controller 0 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read entering the Memory Controller 1 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "ACT command for a read request sent to DRAM",
Signature · 1 line
2.40.0.577.gac1e443424-goog
author reviewer
#1 lore raw

PATCH v2 2/5 perf vendor events intel: Update free running icelakex events

Fix the PMU names, event code and umask. Remove UNC_IIO_BANDWIDTH_OUT events that aren't supported.

These updates were generated by:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
with this PR:
https://github.com/intel/perfmon/pull/66

---
 .../arch/x86/icelakex/uncore-memory.json      |  4 +-
 .../arch/x86/icelakex/uncore-other.json       | 84 ++++++-------------
 2 files changed, 30 insertions(+), 58 deletions(-)
tools/perf/pmu-events/arch/x86/icelakex/uncore-memory.json +3 -1
index 66bb4538c6f2..814d9599474d 100644
@@ -106,9 +106,11 @@
     },
     {
         "BriefDescription": "Free running counter that increments for the Memory Controller",
+        "EventCode": "0xff",
         "EventName": "UNC_M_CLOCKTICKS_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x10",
+        "Unit": "imc_free_running"
     },
     {
         "BriefDescription": "DRAM Precharge All Commands",
tools/perf/pmu-events/arch/x86/icelakex/uncore-other.json +27 -57

136 lines not on this page · raw message

Signature · 1 line
2.40.0.577.gac1e443424-goog
author reviewer
#2 lore raw

PATCH v2 3/5 perf vendor events intel: Correct knightslanding memory topic

In reply to Ian Rogers

Correct the memory topic of events for the imc related PMUs.

These updates were generated by:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
with this PR:
https://github.com/intel/perfmon/pull/66

---
 .../x86/knightslanding/uncore-memory.json     | 38 +++++++++++++++++++
 .../arch/x86/knightslanding/uncore-other.json | 36 ------------------
 2 files changed, 38 insertions(+), 36 deletions(-)
 create mode 100644 tools/perf/pmu-events/arch/x86/knightslanding/uncore-memory.json
tools/perf/pmu-events/arch/x86/knightslanding/uncore-memory.json new file +38 -0
new file mode 100644
index 000000000000..47da947b1a6e
@@ -0,0 +1,38 @@
+[
+    {
+        "BriefDescription": "CAS All",
+        "EventCode": "0x03",
+        "EventName": "UNC_M_CAS_COUNT.ALL",
+        "PerPkg": "1",
+        "UMask": "0x3",
+        "Unit": "iMC_DCLK"
+    },
+    {
+        "BriefDescription": "CAS Reads",
+        "EventCode": "0x03",
+        "EventName": "UNC_M_CAS_COUNT.RD",
+        "PerPkg": "1",
+        "UMask": "0x1",
+        "Unit": "iMC_DCLK"
+    },
+    {
+        "BriefDescription": "CAS Writes",
+        "EventCode": "0x03",
+        "EventName": "UNC_M_CAS_COUNT.WR",
+        "PerPkg": "1",
+        "UMask": "0x2",
+        "Unit": "iMC_DCLK"
+    },
+    {
+        "BriefDescription": "DCLK count",
+        "EventName": "UNC_M_D_CLOCKTICKS",
+        "PerPkg": "1",
+        "Unit": "iMC_DCLK"
+    },
+    {
+        "BriefDescription": "UCLK count",
+        "EventName": "UNC_M_U_CLOCKTICKS",
+        "PerPkg": "1",
+        "Unit": "iMC_UCLK"
+    }
+]
tools/perf/pmu-events/arch/x86/knightslanding/uncore-other.json +0 -36
index 491cb37ddab0..fc85e0c95318 100644
@@ -3621,41 +3621,5 @@
         "PerPkg": "1",
         "UMask": "0x4",
         "Unit": "M2PCIe"
-    },
-    {
-        "BriefDescription": "CAS All",
-        "EventCode": "0x03",
-        "EventName": "UNC_M_CAS_COUNT.ALL",
-        "PerPkg": "1",
-        "UMask": "0x3",
-        "Unit": "iMC_DCLK"
-    },
-    {
-        "BriefDescription": "CAS Reads",
-        "EventCode": "0x03",
-        "EventName": "UNC_M_CAS_COUNT.RD",
-        "PerPkg": "1",
-        "UMask": "0x1",
-        "Unit": "iMC_DCLK"
-    },
-    {
-        "BriefDescription": "CAS Writes",
-        "EventCode": "0x03",
-        "EventName": "UNC_M_CAS_COUNT.WR",
-        "PerPkg": "1",
-        "UMask": "0x2",
-        "Unit": "iMC_DCLK"
-    },
-    {
-        "BriefDescription": "DCLK count",
-        "EventName": "UNC_M_D_CLOCKTICKS",
-        "PerPkg": "1",
-        "Unit": "iMC_DCLK"
-    },
-    {
-        "BriefDescription": "UCLK count",
-        "EventName": "UNC_M_U_CLOCKTICKS",
-        "PerPkg": "1",
-        "Unit": "iMC_UCLK"
     }
 ]
Signature · 1 line
2.40.0.577.gac1e443424-goog
author reviewer
#3 lore raw

PATCH v2 4/5 perf vendor events intel: Update free running snowridgex events

In reply to Ian Rogers

Fix the PMU names, event code and umask. Remove UNC_IIO_BANDWIDTH_OUT events that aren't supported.

These updates were generated by:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
with this PR:
https://github.com/intel/perfmon/pull/66

---
 .../arch/x86/snowridgex/uncore-memory.json    |  4 +-
 .../arch/x86/snowridgex/uncore-other.json     | 92 ++++++-------------
 2 files changed, 30 insertions(+), 66 deletions(-)
tools/perf/pmu-events/arch/x86/snowridgex/uncore-memory.json +3 -1
index 7dc0910694ed..530e9b71b92a 100644
@@ -127,10 +127,12 @@
     },
     {
         "BriefDescription": "Free running counter that increments for the Memory Controller",
+        "EventCode": "0xff",
         "EventName": "UNC_M_CLOCKTICKS_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "UNC_M_CLOCKTICKS_FREERUN",
-        "Unit": "iMC"
+        "UMask": "0x10",
+        "Unit": "imc_free_running"
     },
     {
         "BriefDescription": "DRAM Precharge All Commands",
tools/perf/pmu-events/arch/x86/snowridgex/uncore-other.json +27 -65

152 lines not on this page · raw message

Signature · 1 line
2.40.0.577.gac1e443424-goog
author reviewer
#4 lore raw

PATCH v2 5/5 perf vendor events intel: Update free running tigerlake events

In reply to Ian Rogers

Fix the topic, PMU name, event code and umask.

These updates were generated by:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
with this PR:
https://github.com/intel/perfmon/pull/66

---
 .../arch/x86/tigerlake/uncore-memory.json     | 50 +++++++++++++++++++
 .../arch/x86/tigerlake/uncore-other.json      | 36 -------------
 2 files changed, 50 insertions(+), 36 deletions(-)
 create mode 100644 tools/perf/pmu-events/arch/x86/tigerlake/uncore-memory.json
tools/perf/pmu-events/arch/x86/tigerlake/uncore-memory.json new file +50 -0
new file mode 100644
index 000000000000..99fb5259fd25
@@ -0,0 +1,50 @@
+[
+    {
+        "BriefDescription": "Counts every read (RdCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
+        "EventCode": "0xff",
+        "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
+        "PerPkg": "1",
+        "UMask": "0x20",
+        "Unit": "imc_free_running_0"
+    },
+    {
+        "BriefDescription": "Counts every 64B read and write request entering the Memory Controller to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
+        "EventName": "UNC_MC0_TOTAL_REQCOUNT_FREERUN",
+        "PerPkg": "1",
+        "UMask": "0x10",
+        "Unit": "imc_free_running_0"
+    },
+    {
+        "BriefDescription": "Counts every write (WrCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
+        "EventCode": "0xff",
+        "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
+        "PerPkg": "1",
+        "UMask": "0x30",
+        "Unit": "imc_free_running_0"
+    },
+    {
+        "BriefDescription": "Counts every read (RdCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
+        "EventCode": "0xff",
+        "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
+        "PerPkg": "1",
+        "UMask": "0x20",
+        "Unit": "imc_free_running_1"
+    },
+    {
+        "BriefDescription": "Counts every 64B read and write request entering the Memory Controller to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
+        "EventName": "UNC_MC1_TOTAL_REQCOUNT_FREERUN",
+        "PerPkg": "1",
+        "UMask": "0x10",
+        "Unit": "imc_free_running_1"
+    },
+    {
+        "BriefDescription": "Counts every write (WrCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
+        "EventCode": "0xff",
+        "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
+        "PerPkg": "1",
+        "UMask": "0x30",
+        "Unit": "imc_free_running_1"
+    }
+]
tools/perf/pmu-events/arch/x86/tigerlake/uncore-other.json +0 -36
index a5a254327ae9..6e43aaf64e28 100644
@@ -93,41 +93,5 @@
         "EventName": "UNC_CLOCK.SOCKET",
         "PerPkg": "1",
         "Unit": "CLOCK"
-    },
-    {
-        "BriefDescription": "Counts every read (RdCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
-        "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
-        "PerPkg": "1",
-        "Unit": "imc"
-    },
-    {
-        "BriefDescription": "Counts every 64B read and write request entering the Memory Controller to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
-        "EventName": "UNC_MC0_TOTAL_REQCOUNT_FREERUN",
-        "PerPkg": "1",
-        "Unit": "imc"
-    },
-    {
-        "BriefDescription": "Counts every write (WrCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
-        "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
-        "PerPkg": "1",
-        "Unit": "imc"
-    },
-    {
-        "BriefDescription": "Counts every read (RdCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
-        "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
-        "PerPkg": "1",
-        "Unit": "imc"
-    },
-    {
-        "BriefDescription": "Counts every 64B read and write request entering the Memory Controller to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
-        "EventName": "UNC_MC1_TOTAL_REQCOUNT_FREERUN",
-        "PerPkg": "1",
-        "Unit": "imc"
-    },
-    {
-        "BriefDescription": "Counts every write (WrCAS) issued by the Memory Controller to DRAM (sum of all channels). All requests result in 64 byte data transfers from DRAM.",
-        "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
-        "PerPkg": "1",
-        "Unit": "imc"
     }
 ]
Signature · 1 line
2.40.0.577.gac1e443424-goog
maintainer Applied
#5 lore raw
In reply to Ian Rogers

Em Thu, Apr 06, 2023 at 05:13:18PM -0700, Ian Rogers escreveu:

Quoted 6 lines

Fix the PMU name, event code and umask.

These updates were generated by:
https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.py
with this PR:
https://github.com/intel/perfmon/pull/66

Thanks, applied.

- Arnaldo

Quoted 101 lines
---
 .../arch/x86/alderlake/uncore-memory.json        | 16 ++++++++++++----
 .../arch/x86/alderlaken/uncore-memory.json       | 16 ++++++++++++----
 2 files changed, 24 insertions(+), 8 deletions(-)
tools/perf/pmu-events/arch/x86/alderlake/uncore-memory.json +12 -4
index 2ccd9cf96957..163d7e7755c4 100644
@@ -1,29 +1,37 @@
 [
     {
         "BriefDescription": "Counts every 64B read  request entering the Memory Controller 0 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read entering the Memory Controller 1 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "ACT command for a read request sent to DRAM",
tools/perf/pmu-events/arch/x86/alderlaken/uncore-memory.json +12 -4
index 2ccd9cf96957..163d7e7755c4 100644
@@ -1,29 +1,37 @@
 [
     {
         "BriefDescription": "Counts every 64B read  request entering the Memory Controller 0 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read request entering the Memory Controller 0 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 0 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC0_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_0"
     },
     {
         "BriefDescription": "Counts every 64B read request entering the Memory Controller 1 to DRAM (sum of all channels).",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_RDCAS_COUNT_FREERUN",
         "PerPkg": "1",
         "PublicDescription": "Counts every 64B read entering the Memory Controller 1 to DRAM (sum of all channels).",
-        "Unit": "iMC"
+        "UMask": "0x20",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "Counts every 64B write request entering the Memory Controller 1 to DRAM (sum of all channels). Each write request counts as a new request incrementing this counter. However, same cache line write requests (both full and partial) are combined to a single 64 byte data transfer to DRAM.",
+        "EventCode": "0xff",
         "EventName": "UNC_MC1_WRCAS_COUNT_FREERUN",
         "PerPkg": "1",
-        "Unit": "iMC"
+        "UMask": "0x30",
+        "Unit": "imc_free_running_1"
     },
     {
         "BriefDescription": "ACT command for a read request sent to DRAM",
Signature · 1 line
2.40.0.577.gac1e443424-goog
Signature · 1 line
- Arnaldo