tools/perf/json: Add common metrics for s390
PATCH 1/6 tools/perf/json: Add common metrics for s390
Add 3 metrics for s390 machines:
- Cycles per instruction: Amount of CPU cycles used per instructions, named cpi.
- Problem state ratio: Ratio of instructions executed in problem state compared to total number of instructions, named prbstate.
- Level one instruction and data cache misses per 100 instructions, named l1mp.
For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
Outpuf after:
# ./perf stat -M cpi -- dd if=/dev/zero of=/dev/null bs=1M count=10K 10240+0 records in 10240+0 records out 10737418240 bytes (11 GB, 10 GiB) copied, 1.30151 s, 8.2 GB/s
Performance counter stats for 'dd if=/dev/zero of=/dev/null .....':
6,779,778,802 CPU_CYCLES # 1.96 cpi
3,461,975,090 INSTRUCTIONS
1.306873021 seconds time elapsed
0.001034000 seconds user
1.305677000 seconds sys
#
--- .../pmu-events/arch/s390/cf_z13/transaction.json | 15 +++++++++++++++ .../pmu-events/arch/s390/cf_z14/transaction.json | 15 +++++++++++++++ .../pmu-events/arch/s390/cf_z15/transaction.json | 15 +++++++++++++++ .../pmu-events/arch/s390/cf_z16/transaction.json | 15 +++++++++++++++ 4 files changed, 60 insertions(+)
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +15 -0
@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]
tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +15 -0
@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]
tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +15 -0
@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]
tools/perf/pmu-events/arch/s390/cf_z16/transaction.json +15 -0
@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]
Signature · 1 line
2.39.1
PATCH 2/6 tools/perf/json: Add cache metrics for s390 z16
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memory
For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
Outpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deleted
Performance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp
431,866 L1I_DIR_WRITES
2,395 IDCW_OFF_DRAWER_IV
0 ICW_OFF_DRAWER
0 IDCW_OFF_DRAWER_DRAWER_HIT
1,437 DCW_OFF_DRAWER
425,960,793 L1D_DIR_WRITES 12.165030699 seconds time elapsed
0.001037000 seconds user
12.162140000 seconds sys#
--- .../arch/s390/cf_z16/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)
tools/perf/pmu-events/arch/s390/cf_z16/transaction.json +25 -0
@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((DCW_REQ + DCW_REQ_IV + ICW_REQ + ICW_REQ_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((DCW_REQ_CHIP_HIT + DCW_ON_CHIP + DCW_ON_CHIP_IV + DCW_ON_CHIP_CHIP_HIT + ICW_REQ_CHIP_HIT + ICW_ON_CHIP + ICW_ON_CHIP_IV + ICW_ON_CHIP_CHIP_HIT) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((DCW_REQ_DRAWER_HIT + DCW_ON_CHIP_DRAWER_HIT + DCW_ON_MODULE + DCW_ON_DRAWER + IDCW_ON_MODULE_IV + IDCW_ON_MODULE_CHIP_HIT + IDCW_ON_MODULE_DRAWER_HIT + IDCW_ON_DRAWER_IV + IDCW_ON_DRAWER_CHIP_HIT + IDCW_ON_DRAWER_DRAWER_HIT + ICW_REQ_DRAWER_HIT + ICW_ON_CHIP_DRAWER_HIT + ICW_ON_MODULE + ICW_ON_DRAWER) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((DCW_OFF_DRAWER + IDCW_OFF_DRAWER_IV + IDCW_OFF_DRAWER_CHIP_HIT + IDCW_OFF_DRAWER_DRAWER_HIT + ICW_OFF_DRAWER) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((DCW_ON_CHIP_MEMORY + DCW_ON_MODULE_MEMORY + DCW_ON_DRAWER_MEMORY + DCW_OFF_DRAWER_MEMORY + ICW_ON_CHIP_MEMORY + ICW_ON_MODULE_MEMORY + ICW_ON_DRAWER_MEMORY + ICW_OFF_DRAWER_MEMORY) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ]
Signature · 1 line
2.39.1
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Quoted 30 lines · Thomas Richter
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys
Quoted 4 lines
#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>
Thanks,
Ian
Quoted 42 lines
--- .../arch/s390/cf_z16/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)tools/perf/pmu-events/arch/s390/cf_z16/transaction.json +25
-1@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((DCW_REQ + DCW_REQ_IV + ICW_REQ + ICW_REQ_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((DCW_REQ_CHIP_HIT + DCW_ON_CHIP + DCW_ON_CHIP_IV + DCW_ON_CHIP_CHIP_HIT + ICW_REQ_CHIP_HIT + ICW_ON_CHIP + ICW_ON_CHIP_IV + ICW_ON_CHIP_CHIP_HIT) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((DCW_REQ_DRAWER_HIT + DCW_ON_CHIP_DRAWER_HIT + DCW_ON_MODULE + DCW_ON_DRAWER + IDCW_ON_MODULE_IV + IDCW_ON_MODULE_CHIP_HIT + IDCW_ON_MODULE_DRAWER_HIT + IDCW_ON_DRAWER_IV + IDCW_ON_DRAWER_CHIP_HIT + IDCW_ON_DRAWER_DRAWER_HIT + ICW_REQ_DRAWER_HIT + ICW_ON_CHIP_DRAWER_HIT + ICW_ON_MODULE + ICW_ON_DRAWER) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((DCW_OFF_DRAWER + IDCW_OFF_DRAWER_IV + IDCW_OFF_DRAWER_CHIP_HIT + IDCW_OFF_DRAWER_DRAWER_HIT + ICW_OFF_DRAWER) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((DCW_ON_CHIP_MEMORY + DCW_ON_MODULE_MEMORY + DCW_ON_DRAWER_MEMORY + DCW_OFF_DRAWER_MEMORY + ICW_ON_CHIP_MEMORY + ICW_ON_MODULE_MEMORY + ICW_ON_DRAWER_MEMORY + ICW_OFF_DRAWER_MEMORY) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ] --2.39.1
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 31 lines · Thomas Richter
Quoted 30 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys
Quoted 6 lines
Quoted 4 lines
#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>Acked-by Ian Rogers <[email protected]>
Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Quoted 37 lines · Arnaldo Carvalho de Melo
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 36 lines · Thomas Richter
Quoted 34 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>
Quoted 9 lines
Quoted 1 line
Acked-by Ian Rogers <[email protected]>Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
I would really prefer the current implementation without using "ScaleUnit": "100%" The reason is that these formulars are given to me from the s390 Performance team. They want to use the exact same formulars on all platforms running on s390 which includes z/OS and z/VM. This way they are sure to get the same numbers.
Hope this background info helps.
Thanks a lot.
Signature · 5 lines
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany -- Vorsitzender des Aufsichtsrats: Gregor Pillen Geschäftsführung: David Faller Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Quoted 49 lines · Thomas Richter
Quoted 37 lines · Arnaldo Carvalho de Melo
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 36 lines · Thomas Richter
Quoted 34 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>Quoted 9 lines
Quoted 1 line
Acked-by Ian Rogers <[email protected]>Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
Quoted 6 lines
I would really prefer the current implementation without using "ScaleUnit": "100%" The reason is that these formulars are given to me from the s390 Performance team. They want to use the exact same formulars on all platforms running on s390 which includes z/OS and z/VM. This way they are sure to get the same numbers.
Hope this background info helps.
For the series:
Using ScaleUnit won't change the result. A ScaleUnit of "100%" means scale the result up by multiplying by 100 and then apply the % after the value. Another nit is having metrics that place their units in the name, like _percent, is usually a sign the name can be better. Perhaps we can follow up with some clean up.
Thanks,
Ian
Quoted 8 lines
Thanks a lot.
--
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany --
Vorsitzender des Aufsichtsrats: Gregor Pillen
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Em Tue, Mar 14, 2023 at 09:34:46AM -0700, Ian Rogers escreveu:
Quoted 55 lines · Thomas Richter
Quoted 54 lines · Arnaldo Carvalho de Melo
Quoted 37 lines
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 36 lines · Thomas Richter
Quoted 34 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>Quoted 9 lines
Quoted 1 line
Acked-by Ian Rogers <[email protected]>Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
I would really prefer the current implementation without using "ScaleUnit": "100%" The reason is that these formulars are given to me from the s390 Performance team. They want to use the exact same formulars on all platforms running on s390 which includes z/OS and z/VM. This way they are sure to get the same numbers.
Quoted 4 lines
Quoted 1 line
Hope this background info helps.
For the series:
Acked-by Ian Rogers <[email protected]>
Thanks, applied.
- Arnaldo
Quoted 17 lines
Using ScaleUnit won't change the result. A ScaleUnit of "100%" means scale the result up by multiplying by 100 and then apply the % after the value. Another nit is having metrics that place their units in the name, like
_percent, is usually a sign the name can be better. Perhaps we can follow up with some clean up.Thanks,
IanQuoted 8 lines
Thanks a lot.
--
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany --
Vorsitzender des Aufsichtsrats: Gregor Pillen
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
Signature · 1 line
- Arnaldo
Arnaldo Carvalho de Melo
maintainerRe: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Em Tue, Mar 14, 2023 at 06:36:24PM -0300, Arnaldo Carvalho de Melo escreveu:
Quoted 56 lines
Em Tue, Mar 14, 2023 at 09:34:46AM -0700, Ian Rogers escreveu:
Quoted 55 lines · Thomas Richter
Quoted 54 lines · Arnaldo Carvalho de Melo
Quoted 37 lines
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 36 lines · Thomas Richter
Quoted 34 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>Quoted 9 lines
Quoted 1 line
Acked-by Ian Rogers <[email protected]>Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
I would really prefer the current implementation without using "ScaleUnit": "100%" The reason is that these formulars are given to me from the s390 Performance team. They want to use the exact same formulars on all platforms running on s390 which includes z/OS and z/VM. This way they are sure to get the same numbers.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
Hope this background info helps.
For the series:
Acked-by Ian Rogers <[email protected]>Thanks, applied.
- Arnaldo
While trying to cross build to s390 on:
ubuntu:18.04
using python3
CC /tmp/build/perf/tests/parse-events.o
Exception processing pmu-events/arch/s390/cf_z16/extended.json Traceback (most recent call last):
File "pmu-events/jevents.py", line 997, in <module>
main()
File "pmu-events/jevents.py", line 979, in main
ftw(arch_path, [], preprocess_one_file) File "pmu-events/jevents.py", line 935, in ftw
ftw(item.path, parents + [item.name], action)
File "pmu-events/jevents.py", line 933, in ftw
action(parents, item)
File "pmu-events/jevents.py", line 514, in preprocess_one_file
for event in read_json_events(item.path, topic):
File "pmu-events/jevents.py", line 388, in read_json_events
events = json.load(open(path), object_hook=JsonEvent)
File "/usr/lib/python3.6/json/__init__.py", line 296, in load
return loads(fp.read(),
File "/usr/lib/python3.6/encodings/ascii.py", line 26, in decode
return codecs.ascii_decode(input, self.errors)[0]
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 4271: ordinal not in range(128)
Quoted 22 lines
Quoted 18 lines
Using ScaleUnit won't change the result. A ScaleUnit of "100%" means scale the result up by multiplying by 100 and then apply the % after the value. Another nit is having metrics that place their units in the name, like
_percent, is usually a sign the name can be better. Perhaps we can follow up with some clean up.Thanks,
IanQuoted 8 lines
Thanks a lot.
--
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany --
Vorsitzender des Aufsichtsrats: Gregor Pillen
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294Signature · 1 line
- Arnaldo
Signature · 1 line
- Arnaldo
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Quoted 80 lines · Arnaldo Carvalho de Melo
Em Tue, Mar 14, 2023 at 06:36:24PM -0300, Arnaldo Carvalho de Melo escreveu:
Quoted 56 lines
Em Tue, Mar 14, 2023 at 09:34:46AM -0700, Ian Rogers escreveu:
Quoted 55 lines · Thomas Richter
Quoted 54 lines · Arnaldo Carvalho de Melo
Quoted 37 lines
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 36 lines · Thomas Richter
Quoted 34 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>Quoted 9 lines
Quoted 1 line
Acked-by Ian Rogers <[email protected]>Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
I would really prefer the current implementation without using "ScaleUnit": "100%" The reason is that these formulars are given to me from the s390 Performance team. They want to use the exact same formulars on all platforms running on s390 which includes z/OS and z/VM. This way they are sure to get the same numbers.
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
Hope this background info helps.
For the series:
Acked-by Ian Rogers <[email protected]>Thanks, applied.
- Arnaldo
While trying to cross build to s390 on:
ubuntu:18.04
using python3
CC
/tmp/build/perf/tests/parse-events.o
Exception processingpmu-events/arch/s390/cf_z16/extended.jsonTraceback (most recent call last):
File "pmu-events/jevents.py", line 997, in <module>main() File "pmu-events/jevents.py", line 979, in main ftw(arch_path, [], preprocess_one_file)
Quoted 15 lines
File "
pmu-events/jevents.py", line 935, in ftw
ftw(item.path, parents + [item.name], action)
File "pmu-events/jevents.py", line 933, in ftw
action(parents, item)
File "pmu-events/jevents.py", line 514, inpreprocess_one_file
for event inread_json_events(item.path, topic):
File "pmu-events/jevents.py", line 388, inread_json_events
events =json.load(open(path), object_hook=JsonEvent)
File "/usr/lib/python3.6/json/__init__.py", line 296, in load
return loads(fp.read(),
File "/usr/lib/python3.6/encodings/ascii.py", line 26, in decode
returncodecs.ascii_decode(input, self.errors)[0]
UnicodeDecodeError: 'ascii' codec can't decode byte0xe2in position 4271: ordinal not in range(128)
Hmmm, this is very strange. After reading this mail I installed Ubuntu 18.04 on my s390 system. The build works fine, no errors at all.
# pmu-events/jevents.py s390 all pmu-events/arch pmu-events/pmu-events.c # ll pmu-events/pmu-events.c -rw-r--r-- 1 root root 317284 Mar 23 10:46 pmu-events/pmu-events.c #
The file has the correct contents and the build works fine too.
# make .... Auto-detecting system features: ... dwarf: [ on ] ... dwarf_getlocations: [ on ] ... glibc: [ on ] ... libbfd: [ on ] ... libbfd-buildid: [ on ] ... libcap: [ OFF ] ... libelf: [ on ] ... libnuma: [ on ] ... numa_num_possible_cpus: [ on ] ... libperl: [ on ] ... libpython: [ on ] ... libcrypto: [ on ] ... libunwind: [ OFF ] ... libdw-dwarf-unwind: [ on ] ... zlib: [ on ] ... lzma: [ on ] ... get_cpuid: [ OFF ] ... bpf: [ on ] ... libaio: [ on ] ... libzstd: [ OFF ]
INSTALL libsubcmd_headers
INSTALL libsymbol_headers
INSTALL libperf_headers
INSTALL libapi_headers
INSTALL libbpf_headers
CC pmu-events/pmu-events.o
LD pmu-events/pmu-events-in.o
LINK perf
# ./perf list | grep -A 20 basic:
basic:
CPU_CYCLES
[Cycle Count. Unit: cpum_cf]
INSTRUCTIONS
[Instruction Count. Unit: cpum_cf]
L1D_DIR_WRITES
[Level-1 D-Cache Directory Write Count. Unit: cpum_cf]
L1D_PENALTY_CYCLES
[Level-1 D-Cache Penalty Cycle Count. Unit: cpum_cf]
L1I_DIR_WRITES
[Level-1 I-Cache Directory Write Count. Unit: cpum_cf]
L1I_PENALTY_CYCLES
[Level-1 I-Cache Penalty Cycle Count. Unit: cpum_cf]
PROBLEM_STATE_CPU_CYCLES
[Problem-State Cycle Count. Unit: cpum_cf]
PROBLEM_STATE_INSTRUCTIONS
[Problem-State Instruction Count. Unit: cpum_cf]So everythings works as usual.
Signature · 5 lines
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany -- Vorsitzender des Aufsichtsrats: Gregor Pillen Geschäftsführung: David Faller Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Quoted 35 lines · Thomas Richter
Quoted 14 lines · Arnaldo Carvalho de Melo
While trying to cross build to s390 on:
ubuntu:18.04
using python3
CC
/tmp/build/perf/tests/parse-events.o
Exception processingpmu-events/arch/s390/cf_z16/extended.jsonTraceback (most recent call last):
File "pmu-events/jevents.py", line 997, in <module>main() File "pmu-events/jevents.py", line 979, in main ftw(arch_path, [], preprocess_one_file)Quoted 15 lines
File "
pmu-events/jevents.py", line 935, in ftw
ftw(item.path, parents + [item.name], action)
File "pmu-events/jevents.py", line 933, in ftw
action(parents, item)
File "pmu-events/jevents.py", line 514, inpreprocess_one_file
for event inread_json_events(item.path, topic):
File "pmu-events/jevents.py", line 388, inread_json_events
events =json.load(open(path), object_hook=JsonEvent)
File "/usr/lib/python3.6/json/__init__.py", line 296, in load
return loads(fp.read(),
File "/usr/lib/python3.6/encodings/ascii.py", line 26, in decode
returncodecs.ascii_decode(input, self.errors)[0]
UnicodeDecodeError: 'ascii' codec can't decode byte0xe2in position 4271: ordinal not in range(128)Hmmm, this is very strange. After reading this mail I installed Ubuntu 18.04 on my s390 system. The build works fine, no errors at all.
Quoted 7 lines
# pmu-events/jevents.py s390 all pmu-events/arch pmu-events/pmu-events.c # ll pmu-events/pmu-events.c -rw-r--r-- 1 root root 317284 Mar 23 10:46 pmu-events/pmu-events.c #The file has the correct contents and the build works fine too.
# make
The file contains UTF-8 characters, which were already present before your patch. Guess you need to provide an addon patch which converts to plain ASCII.
Arnaldo Carvalho de Melo
maintainerRe: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
Em Thu, Mar 23, 2023 at 11:06:15AM +0100, Heiko Carstens escreveu:
Quoted 41 lines · Thomas Richter
Quoted 40 lines · Arnaldo Carvalho de Melo
Quoted 14 lines
While trying to cross build to s390 on:
ubuntu:18.04
using python3
CC
/tmp/build/perf/tests/parse-events.o
Exception processingpmu-events/arch/s390/cf_z16/extended.jsonTraceback (most recent call last):
File "pmu-events/jevents.py", line 997, in <module>main() File "pmu-events/jevents.py", line 979, in main ftw(arch_path, [], preprocess_one_file)Quoted 15 lines
File "
pmu-events/jevents.py", line 935, in ftw
ftw(item.path, parents + [item.name], action)
File "pmu-events/jevents.py", line 933, in ftw
action(parents, item)
File "pmu-events/jevents.py", line 514, inpreprocess_one_file
for event inread_json_events(item.path, topic):
File "pmu-events/jevents.py", line 388, inread_json_events
events =json.load(open(path), object_hook=JsonEvent)
File "/usr/lib/python3.6/json/__init__.py", line 296, in load
return loads(fp.read(),
File "/usr/lib/python3.6/encodings/ascii.py", line 26, in decode
returncodecs.ascii_decode(input, self.errors)[0]
UnicodeDecodeError: 'ascii' codec can't decode byte0xe2in position 4271: ordinal not in range(128)Hmmm, this is very strange. After reading this mail I installed Ubuntu 18.04 on my s390 system. The build works fine, no errors at all.
# pmu-events/jevents.py s390 all pmu-events/arch pmu-events/pmu-events.c # ll pmu-events/pmu-events.c -rw-r--r-- 1 root root 317284 Mar 23 10:46 pmu-events/pmu-events.c #
Quoted 6 lines
Quoted 2 lines
The file has the correct contents and the build works fine too.
# makeThe file contains UTF-8 characters, which were already present before your patch. Guess you need to provide an addon patch which converts to plain ASCII.
Yeah, and in this s390 perf test build container in the past I didn't have the needed python3-dev package needed to build jevents.py, so it was being disabled and the problem was left unnoticed.
Now that it is a opt-out feature, I installed python3-dev, jevents.py got built and then this failure surfaced.
- Arnaldo
Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16
In reply to Ian RogersQuoted 60 lines · Ian Rogers
Quoted 49 lines · Thomas Richter
Quoted 37 lines · Arnaldo Carvalho de Melo
Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
Quoted 36 lines · Thomas Richter
Quoted 34 lines
Add metrics for s390 z16
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K .... dd output deletedPerformance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
0 IDCW_OFF_DRAWER_CHIP_HIT # 0.00 l4rp 431,866 L1I_DIR_WRITES 2,395 IDCW_OFF_DRAWER_IV 0 ICW_OFF_DRAWER 0 IDCW_OFF_DRAWER_DRAWER_HIT 1,437 DCW_OFF_DRAWER 425,960,793 L1D_DIR_WRITES12.165030699 seconds time elapsed 0.001037000 seconds user 12.162140000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>Quoted 9 lines
Quoted 1 line
Acked-by Ian Rogers <[email protected]>Thanks, applied the first two patches, please address the review suggestions for patches 3-6 and resubmit only those.
The patches will be in the public perf-tools-next branch later today.
- Arnaldo
Quoted 6 lines
I would really prefer the current implementation without using "ScaleUnit": "100%" The reason is that these formulars are given to me from the s390 Performance team. They want to use the exact same formulars on all platforms running on s390 which includes z/OS and z/VM. This way they are sure to get the same numbers.
Hope this background info helps.
For the series:
Acked-by Ian Rogers <[email protected]>
Quoted 8 lines
Using ScaleUnit won't change the result. A ScaleUnit of "100%" means scale the result up by multiplying by 100 and then apply the % after the value. Another nit is having metrics that place their units in the name, like
_percent, is usually a sign the name can be better. Perhaps we can follow up with some clean up.Thanks,
Ian
Thanks Ian,
I put the ScaleUnit thing on my todo list and will provide a clean up...
Signature · 5 lines
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany -- Vorsitzender des Aufsichtsrats: Gregor Pillen Geschäftsführung: David Faller Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294
PATCH 3/6 tools/perf/json: Add cache metrics for s390 z15
In reply to Thomas RichterAdd metrics for s390 z15
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memory
For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
Outpuf after:
# ./perf stat -M l4rp -- find / .... find output deleted
Performance counter stats for 'find /':
5 L1I_OFFDRAWER_L4_SOURCED_WRITES # 0.01 l4rp
187 L1D_OFFDRAWER_L4_SOURCED_WRITES
0 L1I_OFFDRAWER_L3_SOURCED_WRITES
231,333,165 L1I_DIR_WRITES
3,303 L1D_OFFDRAWER_L3_SOURCED_WRITES
47,461 L1D_OFFDRAWER_L3_SOURCED_WRITES_IV
0 L1I_OFFDRAWER_L3_SOURCED_WRITES_IV
126,706,244 L1D_DIR_WRITES27.870355461 seconds time elapsed
0.521562000 seconds user
12.494503000 seconds sys
#
--- .../arch/s390/cf_z15/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)
tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +25 -0
@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((L1D_L2D_SOURCED_WRITES + L1I_L2I_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((L1D_ONCHIP_L3_SOURCED_WRITES + L1D_ONCHIP_L3_SOURCED_WRITES_IV + L1I_ONCHIP_L3_SOURCED_WRITES + L1I_ONCHIP_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((L1D_ONCLUSTER_L3_SOURCED_WRITES + L1D_ONCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONDRAWER_L4_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES_IV + L1I_ONDRAWER_L4_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONCHIP_L3_SOURCED_WRITES_RO + L1I_OFFCLUSTER_L3_SOURCED_WRITES + L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((L1D_OFFDRAWER_L3_SOURCED_WRITES + L1D_OFFDRAWER_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_L4_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_L4_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ]
Signature · 1 line
2.39.1
Ian Rogers
reviewerRe: [PATCH 3/6] tools/perf/json: Add cache metrics for s390 z15
Quoted 51 lines · Thomas Richter
Add metrics for s390 z15
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- find / .... find output deletedPerformance counter stats for 'find /':
5 L1I_OFFDRAWER_L4_SOURCED_WRITES # 0.01 l4rp 187 L1D_OFFDRAWER_L4_SOURCED_WRITES 0 L1I_OFFDRAWER_L3_SOURCED_WRITES 231,333,165 L1I_DIR_WRITES 3,303 L1D_OFFDRAWER_L3_SOURCED_WRITES 47,461 L1D_OFFDRAWER_L3_SOURCED_WRITES_IV 0 L1I_OFFDRAWER_L3_SOURCED_WRITES_IV 126,706,244 L1D_DIR_WRITES27.870355461 seconds time elapsed 0.521562000 seconds user 12.494503000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>--- .../arch/s390/cf_z15/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +5
-0@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((L1D_L2D_SOURCED_WRITES + L1I_L2I_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100"
tools/perf/pmu-events/arch/s390/cf_z15/transaction.json line 35
Quoted 10 lines
tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +10
-0+ }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((L1D_ONCHIP_L3_SOURCED_WRITES + L1D_ONCHIP_L3_SOURCED_WRITES_IV + L1I_ONCHIP_L3_SOURCED_WRITES + L1I_ONCHIP_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((L1D_ONCLUSTER_L3_SOURCED_WRITES + L1D_ONCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONDRAWER_L4_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES_IV + L1I_ONDRAWER_L4_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONCHIP_L3_SOURCED_WRITES_RO + L1I_OFFCLUSTER_L3_SOURCED_WRITES + L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100"
It is more typical for percentages to change the ScaleUnit to "100%" and not to do the "* 100". Otherwise these look good.
Thanks,
Ian
Quoted 15 lines
tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +10
-1+ }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((L1D_OFFDRAWER_L3_SOURCED_WRITES + L1D_OFFDRAWER_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_L4_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_L4_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ] --2.39.1
PATCH 4/6 tools/perf/json: Add cache metrics for s390 z14
In reply to Thomas RichterAdd metrics for s390 z14
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memory
For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
Outpuf after:
# ./perf stat -M l4rp -- find / .... find output deleted
Performance counter stats for 'find /':
0 L1I_OFFDRAWER_L4_SOURCED_WRITES # 0.01 l4rp
84 L1D_OFFDRAWER_L4_SOURCED_WRITES
0 L1I_OFFDRAWER_L3_SOURCED_WRITES
71,535,353 L1I_DIR_WRITES
219 L1D_OFFDRAWER_L3_SOURCED_WRITES
16,436 L1D_OFFDRAWER_L3_SOURCED_WRITES_IV
0 L1I_OFFDRAWER_L3_SOURCED_WRITES_IV
46,343,940 L1D_DIR_WRITES10.530805537 seconds time elapsed
0.774396000 seconds user
1.602714000 seconds sys
#
--- .../arch/s390/cf_z14/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)
tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +25 -0
@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((L1D_L2D_SOURCED_WRITES + L1I_L2I_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((L1D_ONCHIP_L3_SOURCED_WRITES + L1D_ONCHIP_L3_SOURCED_WRITES_IV + L1I_ONCHIP_L3_SOURCED_WRITES + L1I_ONCHIP_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((L1D_ONCLUSTER_L3_SOURCED_WRITES + L1D_ONCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONDRAWER_L4_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES_IV + L1I_ONDRAWER_L4_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONCHIP_L3_SOURCED_WRITES_RO + L1I_OFFCLUSTER_L3_SOURCED_WRITES + L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((L1D_OFFDRAWER_L3_SOURCED_WRITES + L1D_OFFDRAWER_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_L4_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_L4_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ]
Signature · 1 line
2.39.1
Ian Rogers
reviewerRe: [PATCH 4/6] tools/perf/json: Add cache metrics for s390 z14
Quoted 44 lines · Thomas Richter
Add metrics for s390 z14
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M l4rp -- find / .... find output deletedPerformance counter stats for 'find /':
0 L1I_OFFDRAWER_L4_SOURCED_WRITES # 0.01 l4rp 84 L1D_OFFDRAWER_L4_SOURCED_WRITES 0 L1I_OFFDRAWER_L3_SOURCED_WRITES 71,535,353 L1I_DIR_WRITES 219 L1D_OFFDRAWER_L3_SOURCED_WRITES 16,436 L1D_OFFDRAWER_L3_SOURCED_WRITES_IV 0 L1I_OFFDRAWER_L3_SOURCED_WRITES_IV 46,343,940 L1D_DIR_WRITES10.530805537 seconds time elapsed
0.774396000 seconds user 1.602714000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>--- .../arch/s390/cf_z14/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +0
-0
tools/perf/pmu-events/arch/s390/cf_z14/transaction.json line 25
Quoted 8 lines
tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +5
-0@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((L1D_L2D_SOURCED_WRITES + L1I_L2I_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100"
Some comment as patch #3 wrt ScaleUnit of "100%" rather than "* 100". I can see from the metric above this way is being consistent.
Thanks,
Ian
Quoted 25 lines
tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +20
-1+ }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((L1D_ONCHIP_L3_SOURCED_WRITES + L1D_ONCHIP_L3_SOURCED_WRITES_IV + L1I_ONCHIP_L3_SOURCED_WRITES + L1I_ONCHIP_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((L1D_ONCLUSTER_L3_SOURCED_WRITES + L1D_ONCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONDRAWER_L4_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES + L1I_ONCLUSTER_L3_SOURCED_WRITES_IV + L1I_ONDRAWER_L4_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES + L1D_OFFCLUSTER_L3_SOURCED_WRITES_IV + L1D_ONCHIP_L3_SOURCED_WRITES_RO + L1I_OFFCLUSTER_L3_SOURCED_WRITES + L1I_OFFCLUSTER_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((L1D_OFFDRAWER_L3_SOURCED_WRITES + L1D_OFFDRAWER_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_L4_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES + L1I_OFFDRAWER_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_L4_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ] --2.39.1
PATCH 5/6 tools/perf/json: Add cache metrics for s390 z13
In reply to Thomas RichterAdd metrics for s390 z13
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memory
For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
Output after:
# ./perf stat -M l4rp -- find / ...find output deleted
Performance counter stats for 'find /':
2 L1I_OFFDRAWER_SCOL_L4_SOURCED_WRITES # 0.02 l4rp
252 L1D_ONDRAWER_L4_SOURCED_WRITES
3,465 L1D_ONDRAWER_L3_SOURCED_WRITES_IV
80 L1D_OFFDRAWER_SCOL_L4_SOURCED_WRITES
761 L1D_ONDRAWER_L3_SOURCED_WRITES
0 L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES
131,817,067 L1I_DIR_WRITES
1 L1I_OFFDRAWER_FCOL_L4_SOURCED_WRITES
447 L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES
22 L1D_OFFDRAWER_FCOL_L4_SOURCED_WRITES
7 L1I_ONDRAWER_L4_SOURCED_WRITES
0 L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES
1,071 L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES
3 L1I_ONDRAWER_L3_SOURCED_WRITES
13,352 L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV
15,252 L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV
0 L1I_ONDRAWER_L3_SOURCED_WRITES_IV
0 L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV
57,431,083 L1D_DIR_WRITES
0 L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV 15.386502874 seconds time elapsed
0.647348000 seconds user
3.537041000 seconds sys#
--- .../arch/s390/cf_z13/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +25 -0
@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((L1D_L2D_SOURCED_WRITES + L1I_L2I_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((L1D_ONCHIP_L3_SOURCED_WRITES + L1D_ONCHIP_L3_SOURCED_WRITES_IV + L1I_ONCHIP_L3_SOURCED_WRITES + L1I_ONCHIP_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((L1D_ONNODE_L4_SOURCED_WRITES + L1D_ONNODE_L3_SOURCED_WRITES_IV + L1D_ONNODE_L3_SOURCED_WRITES + L1I_ONNODE_L4_SOURCED_WRITES + L1I_ONNODE_L3_SOURCED_WRITES_IV + L1I_ONNODE_L3_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((L1D_ONDRAWER_L4_SOURCED_WRITES + L1D_ONDRAWER_L3_SOURCED_WRITES_IV + L1D_ONDRAWER_L3_SOURCED_WRITES + L1D_OFFDRAWER_SCOL_L4_SOURCED_WRITES + L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES + L1D_OFFDRAWER_FCOL_L4_SOURCED_WRITES + L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES + L1I_ONDRAWER_L4_SOURCED_WRITES + L1I_ONDRAWER_L3_SOURCED_WRITES_IV + L1I_ONDRAWER_L3_SOURCED_WRITES + L1I_OFFDRAWER_SCOL_L4_SOURCED_WRITES + L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES + L1I_OFFDRAWER_FCOL_L4_SOURCED_WRITES + L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((L1D_ONNODE_MEM_SOURCED_WRITES + L1D_ONDRAWER_MEM_SOURCED_WRITES + L1D_OFFDRAWER_MEM_SOURCED_WRITES + L1D_ONCHIP_MEM_SOURCED_WRITES + L1I_ONNODE_MEM_SOURCED_WRITES + L1I_ONDRAWER_MEM_SOURCED_WRITES + L1I_OFFDRAWER_MEM_SOURCED_WRITES + L1I_ONCHIP_MEM_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ]
Signature · 1 line
2.39.1
Ian Rogers
reviewerRe: [PATCH 5/6] tools/perf/json: Add cache metrics for s390 z13
Quoted 57 lines · Thomas Richter
Add metrics for s390 z13
- Percentage sourced from Level 2 cache
- Percentage sourced from Level 3 on same chip cache
- Percentage sourced from Level 4 Local cache on same book
- Percentage sourced from Level 4 Remote cache on different book
- Percentage sourced from memoryFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutput after:
# ./perf stat -M l4rp -- find / ...find output deletedPerformance counter stats for 'find /':
2 L1I_OFFDRAWER_SCOL_L4_SOURCED_WRITES # 0.02 l4rp 252 L1D_ONDRAWER_L4_SOURCED_WRITES 3,465 L1D_ONDRAWER_L3_SOURCED_WRITES_IV 80 L1D_OFFDRAWER_SCOL_L4_SOURCED_WRITES 761 L1D_ONDRAWER_L3_SOURCED_WRITES 0 L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES 131,817,067 L1I_DIR_WRITES 1 L1I_OFFDRAWER_FCOL_L4_SOURCED_WRITES 447 L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES 22 L1D_OFFDRAWER_FCOL_L4_SOURCED_WRITES 7 L1I_ONDRAWER_L4_SOURCED_WRITES 0 L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES 1,071 L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES 3 L1I_ONDRAWER_L3_SOURCED_WRITES 13,352 L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV 15,252 L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV 0 L1I_ONDRAWER_L3_SOURCED_WRITES_IV 0 L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV 57,431,083 L1D_DIR_WRITES 0 L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV15.386502874 seconds time elapsed 0.647348000 seconds user 3.537041000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>--- .../arch/s390/cf_z13/transaction.json | 25 +++++++++++++++++++ 1 file changed, 25 insertions(+)tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +0
-0
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json line 25
Quoted 8 lines
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +5
-0@@ -18,5 +18,30 @@ "BriefDescription": "Level One Miss per 100 Instructions", "MetricName": "l1mp", "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 2 cache", + "MetricName": "l2p", + "MetricExpr": "((L1D_L2D_SOURCED_WRITES + L1I_L2I_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100"
Looks good but the same comment about using ScaleUnit.
Thanks,
Ian
Quoted 25 lines
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +20
-1+ }, + { + "BriefDescription": "Percentage sourced from Level 3 on same chip cache", + "MetricName": "l3p", + "MetricExpr": "((L1D_ONCHIP_L3_SOURCED_WRITES + L1D_ONCHIP_L3_SOURCED_WRITES_IV + L1I_ONCHIP_L3_SOURCED_WRITES + L1I_ONCHIP_L3_SOURCED_WRITES_IV) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Local cache on same book", + "MetricName": "l4lp", + "MetricExpr": "((L1D_ONNODE_L4_SOURCED_WRITES + L1D_ONNODE_L3_SOURCED_WRITES_IV + L1D_ONNODE_L3_SOURCED_WRITES + L1I_ONNODE_L4_SOURCED_WRITES + L1I_ONNODE_L3_SOURCED_WRITES_IV + L1I_ONNODE_L3_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from Level 4 Remote cache on different book", + "MetricName": "l4rp", + "MetricExpr": "((L1D_ONDRAWER_L4_SOURCED_WRITES + L1D_ONDRAWER_L3_SOURCED_WRITES_IV + L1D_ONDRAWER_L3_SOURCED_WRITES + L1D_OFFDRAWER_SCOL_L4_SOURCED_WRITES + L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_SCOL_L3_SOURCED_WRITES + L1D_OFFDRAWER_FCOL_L4_SOURCED_WRITES + L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV + L1D_OFFDRAWER_FCOL_L3_SOURCED_WRITES + L1I_ONDRAWER_L4_SOURCED_WRITES + L1I_ONDRAWER_L3_SOURCED_WRITES_IV + L1I_ONDRAWER_L3_SOURCED_WRITES + L1I_OFFDRAWER_SCOL_L4_SOURCED_WRITES + L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_SCOL_L3_SOURCED_WRITES + L1I_OFFDRAWER_FCOL_L4_SOURCED_WRITES + L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES_IV + L1I_OFFDRAWER_FCOL_L3_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Percentage sourced from memory", + "MetricName": "memp", + "MetricExpr": "((L1D_ONNODE_MEM_SOURCED_WRITES + L1D_ONDRAWER_MEM_SOURCED_WRITES + L1D_OFFDRAWER_MEM_SOURCED_WRITES + L1D_ONCHIP_MEM_SOURCED_WRITES + L1I_ONNODE_MEM_SOURCED_WRITES + L1I_ONDRAWER_MEM_SOURCED_WRITES + L1I_OFFDRAWER_MEM_SOURCED_WRITES + L1I_ONCHIP_MEM_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" } ] --2.39.1
PATCH 6/6 tools/perf/json: Add metric for tlb and cache s390
In reply to Thomas RichterAdd metrics for tlb and cache statistics:
- finite_cpi: Cycles per Instructions from Finite cache/memory
- est_cpi: Estimated Instruction Complexity CPI infinite Level 1
- scpl1m: Estimated Sourcing Cycles per Level 1 Miss
- tlb_percent: Estimated TLB CPU percentage of Total CPU
- tlb_miss: Estimated Cycles per TLB Miss
For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
Output after:
# ./perf stat -M tlb_miss -- dd if=/dev/zero of=/dev/null bs=1M count=10K ... dd output removed
Performance counter stats for
'dd if=/dev/zero of=/dev/null bs=1M count=10K':
667,726 DTLB2_MISSES # 440.96 tlb_miss
198 ITLB2_WRITES
795,170,260 L1C_TLB2_MISSES
9,478 ITLB2_MISSES
820 DTLB2_WRITES
1,197,126,869 L1D_PENALTY_CYCLES
2,457,447 L1I_PENALTY_CYCLES 1.249342187 seconds time elapsed
0.001030000 seconds user
1.248105000 seconds sys#
--- .../arch/s390/cf_z13/transaction.json | 30 +++++++++++++++++++ .../arch/s390/cf_z14/transaction.json | 25 ++++++++++++++++ .../arch/s390/cf_z15/transaction.json | 25 ++++++++++++++++ .../arch/s390/cf_z16/transaction.json | 25 ++++++++++++++++ 4 files changed, 105 insertions(+)
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +30 -0
@@ -43,5 +43,35 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((L1D_ONNODE_MEM_SOURCED_WRITES + L1D_ONDRAWER_MEM_SOURCED_WRITES + L1D_OFFDRAWER_MEM_SOURCED_WRITES + L1D_ONCHIP_MEM_SOURCED_WRITES + L1I_ONNODE_MEM_SOURCED_WRITES + L1I_ONDRAWER_MEM_SOURCED_WRITES + L1I_OFFDRAWER_MEM_SOURCED_WRITES + L1I_ONCHIP_MEM_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB1_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB1_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB1_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB1_MISSES + ITLB1_MISSES) / CPU_CYCLES) * (L1C_TLB1_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB1_MISSES + ITLB1_MISSES) / (DTLB1_WRITES + ITLB1_WRITES)) * (L1C_TLB1_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" + }, + { + "BriefDescription": "Page Table Entry misses", + "MetricName": "pte_miss", + "MetricExpr": "(TLB2_PTE_WRITES / (DTLB1_WRITES + ITLB1_WRITES)) * 100" } ]
tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +25 -0
@@ -43,5 +43,30 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB2_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB2_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB2_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / CPU_CYCLES) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / (DTLB2_WRITES + ITLB2_WRITES)) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" } ]
tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +25 -0
@@ -43,5 +43,30 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB2_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB2_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB2_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / CPU_CYCLES) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / (DTLB2_WRITES + ITLB2_WRITES)) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" } ]
tools/perf/pmu-events/arch/s390/cf_z16/transaction.json +25 -0
@@ -43,5 +43,30 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((DCW_ON_CHIP_MEMORY + DCW_ON_MODULE_MEMORY + DCW_ON_DRAWER_MEMORY + DCW_OFF_DRAWER_MEMORY + ICW_ON_CHIP_MEMORY + ICW_ON_MODULE_MEMORY + ICW_ON_DRAWER_MEMORY + ICW_OFF_DRAWER_MEMORY) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB2_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB2_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB2_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / CPU_CYCLES) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / (DTLB2_WRITES + ITLB2_WRITES)) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" } ]
Signature · 1 line
2.39.1
Ian Rogers
reviewerRe: [PATCH 6/6] tools/perf/json: Add metric for tlb and cache s390
Quoted 60 lines · Thomas Richter
Add metrics for tlb and cache statistics:
-finite_cpi: Cycles per Instructions from Finite cache/memory
-est_cpi: Estimated Instruction Complexity CPI infinite Level 1
- scpl1m: Estimated Sourcing Cycles per Level 1 Miss
-tlb_percent: Estimated TLB CPU percentage of Total CPU
-tlb_miss: Estimated Cycles per TLB MissFor details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutput after:
# ./perf stat -M tlb_miss -- dd if=/dev/zero of=/dev/null bs=1M count=10K ... dd output removedPerformance counter stats for
'dd if=/dev/zero of=/dev/null bs=1M count=10K': 667,726 DTLB2_MISSES # 440.96 tlb_miss 198 ITLB2_WRITES 795,170,260 L1C_TLB2_MISSES 9,478 ITLB2_MISSES 820 DTLB2_WRITES 1,197,126,869 L1D_PENALTY_CYCLES 2,457,447 L1I_PENALTY_CYCLES 1.249342187 seconds time elapsed 0.001030000 seconds user 1.248105000 seconds sys#
Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>--- .../arch/s390/cf_z13/transaction.json | 30 +++++++++++++++++++ .../arch/s390/cf_z14/transaction.json | 25 ++++++++++++++++ .../arch/s390/cf_z15/transaction.json | 25 ++++++++++++++++ .../arch/s390/cf_z16/transaction.json | 25 ++++++++++++++++ 4 files changed, 105 insertions(+)tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +10
-0@@ -43,5 +43,35 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((L1D_ONNODE_MEM_SOURCED_WRITES + L1D_ONDRAWER_MEM_SOURCED_WRITES + L1D_OFFDRAWER_MEM_SOURCED_WRITES + L1D_ONCHIP_MEM_SOURCED_WRITES + L1I_ONNODE_MEM_SOURCED_WRITES + L1I_ONDRAWER_MEM_SOURCED_WRITES + L1I_OFFDRAWER_MEM_SOURCED_WRITES + L1I_ONCHIP_MEM_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB1_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB1_MISSES / INSTRUCTIONS)"
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json line 65
Quoted 10 lines
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +10
-0+ }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB1_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB1_MISSES + ITLB1_MISSES) / CPU_CYCLES) * (L1C_TLB1_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100"
Looks good again but perhaps the ScaleUnit change. If you'd prefer to keep as-is for consistency I'm happy to add my Acked-by.
Thanks,
Ian
Quoted 120 lines
tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +10
-0+ }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB1_MISSES + ITLB1_MISSES) / (DTLB1_WRITES + ITLB1_WRITES)) * (L1C_TLB1_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" + }, + { + "BriefDescription": "Page Table Entry misses", + "MetricName": "pte_miss", + "MetricExpr": "(TLB2_PTE_WRITES / (DTLB1_WRITES + ITLB1_WRITES)) * 100" } ]tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +25
-0@@ -43,5 +43,30 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB2_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB2_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB2_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / CPU_CYCLES) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / (DTLB2_WRITES + ITLB2_WRITES)) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" } ]tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +25
-0@@ -43,5 +43,30 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((L1D_ONCHIP_MEMORY_SOURCED_WRITES + L1D_ONCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1D_OFFDRAWER_MEMORY_SOURCED_WRITES + L1I_ONCHIP_MEMORY_SOURCED_WRITES + L1I_ONCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFCLUSTER_MEMORY_SOURCED_WRITES + L1I_OFFDRAWER_MEMORY_SOURCED_WRITES) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB2_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB2_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB2_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / CPU_CYCLES) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / (DTLB2_WRITES + ITLB2_WRITES)) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" } ]tools/perf/pmu-events/arch/s390/cf_z16/transaction.json +25
-1@@ -43,5 +43,30 @@ "BriefDescription": "Percentage sourced from memory", "MetricName": "memp", "MetricExpr": "((DCW_ON_CHIP_MEMORY + DCW_ON_MODULE_MEMORY + DCW_ON_DRAWER_MEMORY + DCW_OFF_DRAWER_MEMORY + ICW_ON_CHIP_MEMORY + ICW_ON_MODULE_MEMORY + ICW_ON_DRAWER_MEMORY + ICW_OFF_DRAWER_MEMORY) / (L1I_DIR_WRITES + L1D_DIR_WRITES)) * 100" + }, + { + "BriefDescription": "Cycles per Instructions from Finite cache/memory", + "MetricName": "finite_cpi", + "MetricExpr": "L1C_TLB2_MISSES / INSTRUCTIONS" + }, + { + "BriefDescription": "Estimated Instruction Complexity CPI infinite Level 1", + "MetricName": "est_cpi", + "MetricExpr": "(CPU_CYCLES / INSTRUCTIONS) - (L1C_TLB2_MISSES / INSTRUCTIONS)" + }, + { + "BriefDescription": "Estimated Sourcing Cycles per Level 1 Miss", + "MetricName": "scpl1m", + "MetricExpr": "L1C_TLB2_MISSES / (L1I_DIR_WRITES + L1D_DIR_WRITES)" + }, + { + "BriefDescription": "Estimated TLB CPU percentage of Total CPU", + "MetricName": "tlb_percent", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / CPU_CYCLES) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES)) * 100" + }, + { + "BriefDescription": "Estimated Cycles per TLB Miss", + "MetricName": "tlb_miss", + "MetricExpr": "((DTLB2_MISSES + ITLB2_MISSES) / (DTLB2_WRITES + ITLB2_WRITES)) * (L1C_TLB2_MISSES / (L1I_PENALTY_CYCLES + L1D_PENALTY_CYCLES))" } ] --2.39.1
Quoted 23 lines · Thomas Richter
Add 3 metrics for s390 machines:
- Cycles per instruction: Amount of CPU cycles used per instructions, named cpi.
- Problem state ratio: Ratio of instructions executed in problem state compared to total number of instructions, named prbstate.
- Level one instruction and data cache misses per 100 instructions, named l1mp.For details about the formulas see this documentation:
https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdfOutpuf after:
# ./perf stat -M cpi -- dd if=/dev/zero of=/dev/null bs=1M count=10K 10240+0 records in 10240+0 records out 10737418240 bytes (11 GB, 10 GiB) copied, 1.30151 s, 8.2 GB/sPerformance counter stats for 'dd if=/dev/zero of=/dev/null .....':
6,779,778,802
CPU_CYCLES# 1.96 cpi
3,461,975,090 INSTRUCTIONS
Quoted 8 lines
1.306873021 seconds time elapsed
0.001034000 seconds user
1.305677000 seconds sys
#Signed-off-by Thomas Richter <[email protected]>Acked-by Sumanth Korikkar <[email protected]>
Thanks,
Ian
Quoted 110 lines
--- .../pmu-events/arch/s390/cf_z13/transaction.json | 15 +++++++++++++++ .../pmu-events/arch/s390/cf_z14/transaction.json | 15 +++++++++++++++ .../pmu-events/arch/s390/cf_z15/transaction.json | 15 +++++++++++++++ .../pmu-events/arch/s390/cf_z16/transaction.json | 15 +++++++++++++++ 4 files changed, 60 insertions(+)tools/perf/pmu-events/arch/s390/cf_z13/transaction.json +15
-0@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]tools/perf/pmu-events/arch/s390/cf_z14/transaction.json +15
-0@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]tools/perf/pmu-events/arch/s390/cf_z15/transaction.json +15
-0@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ]tools/perf/pmu-events/arch/s390/cf_z16/transaction.json +15
-1@@ -3,5 +3,20 @@ "BriefDescription": "Transaction count", "MetricName": "transaction", "MetricExpr": "TX_C_TEND + TX_NC_TEND + TX_NC_TABORT + TX_C_TABORT_SPECIAL + TX_C_TABORT_NO_SPECIAL" + }, + { + "BriefDescription": "Cycles per Instruction", + "MetricName": "cpi", + "MetricExpr": "CPU_CYCLES / INSTRUCTIONS" + }, + { + "BriefDescription": "Problem State Instruction Ratio", + "MetricName": "prbstate", + "MetricExpr": "(PROBLEM_STATE_INSTRUCTIONS / INSTRUCTIONS) * 100" + }, + { + "BriefDescription": "Level One Miss per 100 Instructions", + "MetricName": "l1mp", + "MetricExpr": "((L1I_DIR_WRITES + L1D_DIR_WRITES) / INSTRUCTIONS) * 100" } ] --2.39.1