Perfspect updates April, 07, 2023 (#25)

Author: ashrafMahgoub

README.md

@@ -1,14 +1,27 @@
 # PerfSpect · [![Build](https://github.com/intel/PerfSpect/actions/workflows/build.yml/badge.svg)](https://github.com/intel/PerfSpect/actions/workflows/build.yml)[![License](https://img.shields.io/badge/License-BSD--3-blue)](https://github.com/intel/PerfSpect/blob/master/LICENSE)
 
-[Quick Start](#quick-start-requires-perf-installed) | [Requirements](#requirements) | [Build from source](#build-from-source) | [Collection](#collection) | [Post-processing](#post-processing) | [Caveats](#caveats) | [How to contribute](#how-to-contribute)
+[Quick Start](#quick-start-requires-perf-installed) | [Requirements](#requirements) | [Build from source](#build-from-source) | [Caveats](#caveats) | [How to contribute](#how-to-contribute)
 
-PerfSpect is a system performance characterization tool based on linux perf targeting Intel microarchitectures.  
-The tool has two parts
+PerfSpect is a system performance characterization tool built on top of linux perf. It contains two parts
 
-1. perf collection to collect underlying PMU (Performance Monitoring Unit) counters
-2. post processing that generates csv output of performance metrics.
+perf-collect: Collects harware events
 
-### Quick start (requires perf installed)
+- Collection mode:
+  - `sudo ./perf-collect` _default system wide_
+  - `sudo ./perf-collect --socket`
+  - `sudo ./perf-collect --thread`
+  - `sudo ./perf-collect --pid <process-id>`
+  - `sudo ./perf-collect --cid <container-id1>;<container-id2>`
+- Duration:
+  - `sudo ./perf-collect` _default run until terminated_
+  - `sudo ./perf-collect --timeout 10` _run for 10 seconds_
+  - `sudo ./perf-collect --app "myapp.sh myparameter"` _runs for duration of another process_
+
+perf-postprocess: Calculates high level metrics from hardware events
+
+- `perf-postprocess -r results/perfstat.csv`
+
+## Quick start (requires perf installed)
 
 ```
 wget -qO- https://github.com/intel/PerfSpect/releases/latest/download/perfspect.tgz | tar xvz
@@ -17,7 +30,7 @@ sudo ./perf-collect --timeout 10
 sudo ./perf-postprocess -r results/perfstat.csv --html perfstat.html
 ```
 
-### Deploy in Kubernetes
+## Deploy in Kubernetes
 
 Modify the template [deamonset.yml](docs/daemonset.yml) to deploy in kubernetes
 
@@ -52,97 +65,13 @@ _Note: PerfSpect may work on other Linux distributions, but has not been thoroug
 
 ## Build from source
 
-Requires recent python
+Requires recent python. On successful build, binaries will be created in `dist` folder
 
 ```
 pip3 install -r requirements.txt
 make
 ```
 
-On successful build, binaries will be created in `dist` folder
-
-## Collection:
-
-```
-(sudo) ./perf-collect (options)  -- Some options can be used only with root privileges
-
-usage: perf-collect [-h] [-t TIMEOUT | -a APP]
-                    [-p PID | -c CID | --thread | --socket] [-V] [-i INTERVAL]
-                    [-m MUXINTERVAL] [-o OUTCSV] [-v]
-
-optional arguments:
-  -h, --help            show this help message and exit
-  -t TIMEOUT, --timeout TIMEOUT
-                        perf event collection time
-  -a APP, --app APP     Application to run with perf-collect, perf collection
-                        ends after workload completion
-  -p PID, --pid PID     perf-collect on selected PID(s)
-  -c CID, --cid CID     perf-collect on selected container ids
-  --thread              Collect for thread metrics
-  --socket              Collect for socket metrics
-  -V, --version         display version info
-  -i INTERVAL, --interval INTERVAL
-                        interval in seconds for time series dump, default=1
-  -m MUXINTERVAL, --muxinterval MUXINTERVAL
-                        event mux interval in milli seconds, default=0 i.e.
-                        will use the system default
-  -o OUTCSV, --outcsv OUTCSV
-                        perf stat output in csv format,
-                        default=results/perfstat.csv
-  -v, --verbose         Display debugging information
-```
-
-### Examples
-
-1. sudo ./perf-collect (collect PMU counters using predefined architecture specific event file until collection is terminated)
-2. sudo ./perf-collect -a "myapp.sh myparameter" (collect perf for myapp.sh)
-3. sudo ./perf-collect --cid "one or more container IDs from docker or kubernetes seperated by semicolon"
-
-## Post-processing:
-
-```
-./perf-postprocess (options)
-
-usage: perf-postprocess [-h] [--version] [-m METRICFILE] [-o OUTFILE]
-                        [--persocket] [--percore] [-v] [--epoch] [-html HTML]
-                        [-r RAWFILE]
-
-perf-postprocess: perf post process
-
-optional arguments:
-  -h, --help            show this help message and exit
-  --version, -V         display version information
-  -m METRICFILE, --metricfile METRICFILE
-                        formula file, default metric file for the architecture
-  -o OUTFILE, --outfile OUTFILE
-                        perf stat outputs in csv format,
-                        default=results/metric_out.csv
-  --persocket           generate per socket metrics
-  --percore             generate per core metrics
-  -v, --verbose         include debugging information, keeps all intermediate
-                        csv files
-  --epoch               time series in epoch format, default is sample count
-  -html HTML, --html HTML
-                        Static HTML report
-
-required arguments:
-  -r RAWFILE, --rawfile RAWFILE
-                        Raw CSV output from perf-collect
-```
-
-### Examples
-
-./perf-postprocess -r results/perfstat.csv (post processes perfstat.csv and creates metric_out.csv, metric_out.average.csv, metric_out.raw.csv)
-
-./perf-postprocess -r results/perfstat.csv --html perfstat.html (creates a report for TMA analysis and system level metric charts.)
-
-### Notes
-
-1. metric_out.csv : Time series dump of the metrics. The metrics are defined in events/metric.json
-2. metric_out.averags.csv: Average of metrics over the collection period
-3. metric_out.raw.csv: csv file with raw events normalized per second
-4. Socket/core level metrics: Additonal csv files outputfile.socket.csv/outputfile.core.csv will be generated.
-
 ## Caveats
 
 1. The tool can collect only the counters supported by underlying linux perf version.

_version.txt

@@ -1 +1 @@
-1.2.5
+1.2.6

events/icx.txt

@@ -151,6 +151,13 @@ upi/event=0x2,umask=0xf,name='UNC_UPI_TxL_FLITS.ALL_DATA'/,
 upi/event=0x2,umask=0x97,name='UNC_UPI_TxL_FLITS.NON_DATA'/,
 upi/event=0x1,umask=0x0,name='UNC_UPI_CLOCKTICKS'/;
 
+cha/event=0x35,umask=0xc88ffe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_CRD_PREF'/,
+cha/event=0x35,umask=0xc80ffe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_CRD'/;
+
+cha/event=0x35,umask=0xc897fe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_PREF'/,
+cha/event=0x35,umask=0xc817fe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD'/,
+cha/event=0x35,umask=0xccd7fe01,name='UNC_CHA_TOR_INSERTS.IA_MISS_LLCPREFDATA'/;
+
 cha/event=0x35,umask=0xC816FE01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_LOCAL'/,
 cha/event=0x35,umask=0xC8177E01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_REMOTE'/,
 cha/event=0x35,umask=0xC896FE01,name='UNC_CHA_TOR_INSERTS.IA_MISS_DRD_PREF_LOCAL'/,

events/metric_bdx.json

@@ -173,159 +173,159 @@
 		"expression": "([UNC_C_TOR_INSERTS.OPCODE.0x1c8] + [UNC_C_TOR_INSERTS.OPCODE.0x180]) * 64 / 1000000"
 	},
 	{
-		"name": "metric_TMAM_Info_cycles_both_threads_active(%)",
+		"name": "metric_TMA_Info_cycles_both_threads_active(%)",
 		"expression": "100 * ( (1 - ([CPU_CLK_THREAD_UNHALTED.ONE_THREAD_ACTIVE] / ([CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY] / 2)) ) if [const_thread_count] > 1 else 0)"
 	},
 	{
-		"name": "metric_TMAM_Info_CoreIPC",
+		"name": "metric_TMA_Info_CoreIPC",
 		"expression": "[instructions] / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_Frontend_Bound(%)",
+		"name": "metric_TMA_Frontend_Bound(%)",
 		"expression": "100 * [IDQ_UOPS_NOT_DELIVERED.CORE] / (4 * ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))"
 	},
 	{
-		"name": "metric_TMAM_..Frontend_Latency(%)",
+		"name": "metric_TMA_..Frontend_Latency(%)",
 		"expression": "100 * [IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE] / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_....ICache_Misses(%)",
+		"name": "metric_TMA_....ICache_Misses(%)",
 		"expression": "100 * [ICACHE.IFDATA_STALL] / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....ITLB_Misses(%)",
+		"name": "metric_TMA_....ITLB_Misses(%)",
 		"expression": "100 * ((14 * [ITLB_MISSES.STLB_HIT]) + [ITLB_MISSES.WALK_DURATION_c1] + (7 * [ITLB_MISSES.WALK_COMPLETED] )) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....Branch_Resteers(%)",
+		"name": "metric_TMA_....Branch_Resteers(%)",
 		"expression": "100 * (([RS_EVENTS.EMPTY_CYCLES] - [ICACHE.IFDATA_STALL] - (14 * [ITLB_MISSES.STLB_HIT] + [ITLB_MISSES.WALK_DURATION_c1] + 7 * [ITLB_MISSES.WALK_COMPLETED])) / [RS_EVENTS.EMPTY_END]) * ([BR_MISP_RETIRED.ALL_BRANCHES] + [MACHINE_CLEARS.COUNT] + [BACLEARS.ANY]) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....DSB_Switches(%)",
+		"name": "metric_TMA_....DSB_Switches(%)",
 		"expression": "100 * 2 * [DSB2MITE_SWITCHES.PENALTY_CYCLES] / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....MS_Switches(%)",
+		"name": "metric_TMA_....MS_Switches(%)",
 		"expression": "100 * 2 * [IDQ.MS_SWITCHES] / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_..Frontend_Bandwidth(%)",
+		"name": "metric_TMA_..Frontend_Bandwidth(%)",
 		"expression": "100 * ([IDQ_UOPS_NOT_DELIVERED.CORE] - (4 * [IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE])) / (4 * [CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_Bad_Speculation(%)",
+		"name": "metric_TMA_Bad_Speculation(%)",
 		"expression": "100 * ([UOPS_ISSUED.ANY] - [UOPS_RETIRED.RETIRE_SLOTS] + ((4 * [INT_MISC.RECOVERY_CYCLES_ANY]) / [const_thread_count])) / (4 * ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])) "
 	},
 	{
-		"name": "metric_TMAM_..Branch_Mispredicts(%)",
+		"name": "metric_TMA_..Branch_Mispredicts(%)",
 		"expression": "([BR_MISP_RETIRED.ALL_BRANCHES] / ([BR_MISP_RETIRED.ALL_BRANCHES] + [MACHINE_CLEARS.COUNT])) * 100 * ([UOPS_ISSUED.ANY] - [UOPS_RETIRED.RETIRE_SLOTS] + (4 * [INT_MISC.RECOVERY_CYCLES_ANY] / [const_thread_count])) / (4 * [CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_..Machine_Clears(%)",
+		"name": "metric_TMA_..Machine_Clears(%)",
 		"expression": "([MACHINE_CLEARS.COUNT] / ([BR_MISP_RETIRED.ALL_BRANCHES] + [MACHINE_CLEARS.COUNT])) * 100 * ([UOPS_ISSUED.ANY] - [UOPS_RETIRED.RETIRE_SLOTS] + (4 * [INT_MISC.RECOVERY_CYCLES_ANY] / [const_thread_count])) / (4 * [CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_Backend_bound(%)",
+		"name": "metric_TMA_Backend_Bound(%)",
 		"expression": "100 - (100 * ([UOPS_ISSUED.ANY] - [UOPS_RETIRED.RETIRE_SLOTS] + 4 * ([INT_MISC.RECOVERY_CYCLES_ANY] / [const_thread_count]) + [IDQ_UOPS_NOT_DELIVERED.CORE] + [UOPS_RETIRED.RETIRE_SLOTS]) / (4 * [CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])) "
 	},
 	{
-		"name": "metric_TMAM_..Memory_Bound(%)",
+		"name": "metric_TMA_..Memory_Bound(%)",
 		"expression": "100 * (1 - (([UOPS_ISSUED.ANY] - [UOPS_RETIRED.RETIRE_SLOTS] + 4 * ([INT_MISC.RECOVERY_CYCLES_ANY] / [const_thread_count]) + [IDQ_UOPS_NOT_DELIVERED.CORE] + [UOPS_RETIRED.RETIRE_SLOTS]) / (4 * [CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))) * ([CYCLE_ACTIVITY.STALLS_MEM_ANY] + [RESOURCE_STALLS.SB]) / ([CYCLE_ACTIVITY.STALLS_TOTAL] + [UOPS_EXECUTED.CYCLES_GE_1_UOPS_EXEC] - ( [UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC] if ([instructions] / [cpu-cycles]) > 1.8 else  [UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC]) - ( [RS_EVENTS.EMPTY_CYCLES] if ([IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE] / [CPU_CLK_UNHALTED.THREAD_ANY]) > 0.1 else 0) + [RESOURCE_STALLS.SB])"
 	},
 	{
-		"name": "metric_TMAM_....L1_Bound(%)",
+		"name": "metric_TMA_....L1_Bound(%)",
 		"expression": "100 * ([CYCLE_ACTIVITY.STALLS_MEM_ANY] - [CYCLE_ACTIVITY.STALLS_L1D_MISS]) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_......DTLB_Load(%)",
+		"name": "metric_TMA_......DTLB_Load(%)",
 		"expression": "100 * ([DTLB_LOAD_MISSES.STLB_HIT] * 8 + [DTLB_LOAD_MISSES.WALK_DURATION_c1] + 7 * [DTLB_LOAD_MISSES.WALK_COMPLETED]) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_......Store_Fwd_Blk(%)",
+		"name": "metric_TMA_......Store_Fwd_Blk(%)",
 		"expression": "100 * (13 * [LD_BLOCKS.STORE_FORWARD]) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....L2_Bound(%)",
+		"name": "metric_TMA_....L2_Bound(%)",
 		"expression": "100 * ([CYCLE_ACTIVITY.STALLS_L1D_MISS] - [CYCLE_ACTIVITY.STALLS_L2_MISS]) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....L3_Bound(%)",
+		"name": "metric_TMA_....L3_Bound(%)",
 		"expression": "100 * [MEM_LOAD_UOPS_RETIRED.L3_HIT] / ([MEM_LOAD_UOPS_RETIRED.L3_HIT] + 7 * [MEM_LOAD_UOPS_RETIRED.L3_MISS]) * ([CYCLE_ACTIVITY.STALLS_L2_MISS] / [cpu-cycles])"
 	},
 	{
-		"name": "metric_TMAM_......L3_Latency(%)",
+		"name": "metric_TMA_......L3_Latency(%)",
 		"expression": "100 * 41 * [MEM_LOAD_UOPS_RETIRED.L3_HIT] * ( 1 + [MEM_LOAD_UOPS_RETIRED.HIT_LFB] / ( [MEM_LOAD_UOPS_RETIRED.L2_HIT] + [MEM_LOAD_UOPS_RETIRED.L3_HIT] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD] ) ) / [cpu-cycles] "
 	},
 	{
-		"name": "metric_TMAM_......Contested_Accesses(%)",
+		"name": "metric_TMA_......Contested_Accesses(%)",
 		"expression": "100 * 60 * ([MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS]) * ( 1 + [MEM_LOAD_UOPS_RETIRED.HIT_LFB] / ( [MEM_LOAD_UOPS_RETIRED.L2_HIT] + [MEM_LOAD_UOPS_RETIRED.L3_HIT] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD] ) ) / [cpu-cycles] "
 	},
 	{
-		"name": "metric_TMAM_......Data_Sharing(%)",
+		"name": "metric_TMA_......Data_Sharing(%)",
 		"expression": "100 * 43 * [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT] * ( 1 + [MEM_LOAD_UOPS_RETIRED.HIT_LFB] / ( [MEM_LOAD_UOPS_RETIRED.L2_HIT] + [MEM_LOAD_UOPS_RETIRED.L3_HIT] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM] + [MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM] + [MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD] ) ) / [cpu-cycles] "
 	},
 	{
-		"name": "metric_TMAM_......SQ_Full(%)",
+		"name": "metric_TMA_......SQ_Full(%)",
 		"expression": "100 * ([OFFCORE_REQUESTS_BUFFER.SQ_FULL] / [const_thread_count]) / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_....MEM_Bound(%)",
+		"name": "metric_TMA_....MEM_Bound(%)",
 		"expression": "100 * (1 - ( [MEM_LOAD_UOPS_RETIRED.L3_HIT] / ([MEM_LOAD_UOPS_RETIRED.L3_HIT] + 7 * [MEM_LOAD_UOPS_RETIRED.L3_MISS])) ) * ([CYCLE_ACTIVITY.STALLS_L2_MISS] / [cpu-cycles])"
 	},
 	{
-		"name": "metric_TMAM_......MEM_Bandwidth(%)",
+		"name": "metric_TMA_......MEM_Bandwidth(%)",
 		"expression": "100 * (min([OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD_c4], [cpu-cycles])) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_......MEM_Latency(%)",
+		"name": "metric_TMA_......MEM_Latency(%)",
 		"expression": "100 * (min([OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DATA_RD], [cpu-cycles]) - min([OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD_c4], [cpu-cycles])) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_....Stores_Bound(%)",
+		"name": "metric_TMA_....Store_Bound(%)",
 		"expression": "100 * [RESOURCE_STALLS.SB] / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_......DTLB_Store(%)",
+		"name": "metric_TMA_......DTLB_Store(%)",
 		"expression": "100 * (7 * [DTLB_STORE_MISSES.STLB_HIT] + [DTLB_STORE_MISSES.WALK_DURATION_c1]) / [cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_..Core_Bound(%)",
+		"name": "metric_TMA_..Core_Bound(%)",
 		"expression": "100 * ( 1 - (( [UOPS_ISSUED.ANY] - [UOPS_RETIRED.RETIRE_SLOTS] + 4 * ([INT_MISC.RECOVERY_CYCLES_ANY] / [const_thread_count]) + [IDQ_UOPS_NOT_DELIVERED.CORE] + [UOPS_RETIRED.RETIRE_SLOTS] ) / ( 4 * [CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))) * (1 - (([CYCLE_ACTIVITY.STALLS_MEM_ANY] + [RESOURCE_STALLS.SB]) / ([CYCLE_ACTIVITY.STALLS_TOTAL] + [UOPS_EXECUTED.CYCLES_GE_1_UOPS_EXEC] - ( [UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC] if ([instructions] / [cpu-cycles]) > 1.8 else [UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC]) - ([RS_EVENTS.EMPTY_CYCLES] if ([IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE] / [CPU_CLK_UNHALTED.THREAD_ANY]) > 0.1 else 0) + [RESOURCE_STALLS.SB])))"
 	},
 	{
-		"name": "metric_TMAM_....Divider(%)",
+		"name": "metric_TMA_....Divider(%)",
 		"expression": "100 * [ARITH.FPU_DIV_ACTIVE] / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_....Ports_Utilization(%)",
+		"name": "metric_TMA_....Ports_Utilization(%)",
 		"expression": "100 * (( [CYCLE_ACTIVITY.STALLS_TOTAL] + [UOPS_EXECUTED.CYCLES_GE_1_UOPS_EXEC] - ([UOPS_EXECUTED.CYCLES_GE_3_UOPS_EXEC] if ([instructions] / [cpu-cycles]) > 1.8 else [UOPS_EXECUTED.CYCLES_GE_2_UOPS_EXEC]) - ([RS_EVENTS.EMPTY_CYCLES] if ([IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE] / [CPU_CLK_UNHALTED.THREAD_ANY]) > 0.1 else 0) + [RESOURCE_STALLS.SB]) - [RESOURCE_STALLS.SB] - [CYCLE_ACTIVITY.STALLS_MEM_ANY] ) /[cpu-cycles]"
 	},
 	{
-		"name": "metric_TMAM_......0_Port_Utilized(%)",
+		"name": "metric_TMA_......0_Port_Utilized(%)",
 		"expression": "100 * (([UOPS_EXECUTED.CORE_i1_c1] / [const_thread_count]) if ([const_thread_count] > 1) else ([RS_EVENTS.EMPTY_CYCLES] if ([CYCLE_ACTIVITY.STALLS_TOTAL] - ([IDQ_UOPS_NOT_DELIVERED.CYCLES_0_UOPS_DELIV.CORE] / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])) ) > 0.1 else 0)) / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]) "
 	},
 	{
-		"name": "metric_TMAM_......1_Port_Utilized(%)",
+		"name": "metric_TMA_......1_Port_Utilized(%)",
 		"expression": "100 * (([UOPS_EXECUTED.CORE_c1] - [UOPS_EXECUTED.CORE_c2]) / [const_thread_count]) / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_......2_Port_Utilized(%)",
+		"name": "metric_TMA_......2_Port_Utilized(%)",
 		"expression": "100 * (([UOPS_EXECUTED.CORE_c2] - [UOPS_EXECUTED.CORE_c3]) / [const_thread_count]) / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_......3m_Ports_Utilized(%)",
+		"name": "metric_TMA_......3m_Ports_Utilized(%)",
 		"expression": "100 * ([UOPS_EXECUTED.CORE_c3] / [const_thread_count]) / ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count])"
 	},
 	{
-		"name": "metric_TMAM_Retiring(%)",
+		"name": "metric_TMA_Retiring(%)",
 		"expression": "100 * [UOPS_RETIRED.RETIRE_SLOTS] / (4 * ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))"
 	},
 	{
-		"name": "metric_TMAM_..Base(%)",
+		"name": "metric_TMA_..Base(%)",
 		"expression": "100 *(([UOPS_RETIRED.RETIRE_SLOTS] / (4 * ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))) - (([UOPS_RETIRED.RETIRE_SLOTS] / [UOPS_ISSUED.ANY]) * [IDQ.MS_UOPS] / (4 * ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))))"
 	},
 	{
-		"name": "metric_TMAM_..Microcode_Sequencer(%)",
+		"name": "metric_TMA_..Microcode_Sequencer(%)",
 		"expression": "100 * (([UOPS_RETIRED.RETIRE_SLOTS] / [UOPS_ISSUED.ANY]) * [IDQ.MS_UOPS] )/ (4 * ([CPU_CLK_UNHALTED.THREAD_ANY] / [const_thread_count]))"
 	}
 ]

events/metric_icx.json

@@ -130,7 +130,7 @@
     },
     {
         "name": "metric_LLC code read MPI (demand+prefetch)",
-        "expression": "([UNC_CHA_TOR_INSERTS.IA_MISS_LLCPREFCODE] + [UNC_CHA_TOR_INSERTS.IA_MISS_CRD] + [UNC_CHA_TOR_INSERTS.IA_MISS_CRD_PREF]) / [instructions]"
+        "expression": "([UNC_CHA_TOR_INSERTS.IA_MISS_CRD] + [UNC_CHA_TOR_INSERTS.IA_MISS_CRD_PREF]) / [instructions]"
     },
     {
         "name": "metric_LLC data read MPI (demand+prefetch)",