Skip to content

Conversation

@dvalinrh
Copy link
Contributor

@dvalinrh dvalinrh commented Jan 12, 2026

Description

Fixes the wrapper code to save the benchmarks results in the pcp logs.

Before/After Comparison

Before
No output for results in pcp
o.w.iteration o.w.running o.w.numthreads o.w.runtime o.w.throughput o.w.latency

15:41:12 N/A N/A N/A N/A N/A N/A
15:41:13 0.000 0.000 0.000 NaN NaN NaN
15:41:14 0.000 0.000 0.000 NaN NaN NaN

After
o.w.iteration o.w.SUMM_ME o.w.CPU_avx512 o.w.SUMM_CPU o.w.ME_THREADED o.w.ME_LATENCY o.w.ME_LARGE o.w.ME_WRITE o.w.ME_READ_L o.w.ME_READ_S o.w.ME_ALLOC_S o.w.m_CPU_enc_ECDSA o.w.m_CPU_enc_AES o.w.m_CPU_enc_SHA o.w.CPU_avx o.w.CPU_fma o.w.CPU_sse o.w.CPU_mm o.w.CPU_MATRIX_MULT_SSE o.w.CPU_PHYSICS o.w.CPU_SINGLETHREAD o.w.CPU_COMPRESSION o.w.CPU_ENCRYPTION o.w.CPU_SORTING o.w.CPU_PRIME o.w.CPU_FLOATINGPOINT_MATH o.w.CPU_INTEGER_MATH

16:33:26 1.000 2119.817 4294.222 3949.275 24418.757 52.891 14487.500 9952.606 11545.292 22341.960 1360.445 453445788.206 3937983257.671 510818534.269 2577.061 3652.387 1235.218 344.081 2939.722 478.939 1803.238 49590.234 1558.383 5974.223 23.804 6057.154 11261.641

Clerical Stuff

This closes #43
to close the issue out automatically.

Mention the JIRA ticket of the issue, this helps keep
everything together so we can find this PR easily.

Relates to JIRA: RPOPC-763

Test information

Command executed
#!/bin/bash
/home/ec2-user/workloads/passmark-wrapper/passmark/passmark_run --run_user ec2-user --home_parent /home --iterations 1 --tuned_setting tuned_none_sys_file_ --host_config "m5.xlarge" --sysname "m5.xlarge" --sys_type aws --no-overrides --use_pcp

===============================
csv file

CPU_INTEGER_MATH:11261.641333333331
CPU_FLOATINGPOINT_MATH:6057.1541365866988
CPU_PRIME:23.803852076272896
CPU_SORTING:5974.2228394045142
CPU_ENCRYPTION:1558.3825366168708
CPU_COMPRESSION:49590.234479685467
CPU_SINGLETHREAD:1803.23847150815
CPU_PHYSICS:478.93937053621914
CPU_MATRIX_MULT_SSE:2939.7218804837939
CPU_mm:344.08071706262359
CPU_sse:1235.2182204515746
CPU_fma:3652.3866208054092
CPU_avx:2577.0606889375385
CPU_avx512:4294.2219917406546
m_CPU_enc_SHA:510818534.26934922
m_CPU_enc_AES:3937983257.6712565
m_CPU_enc_ECDSA:453445788.20610982
ME_ALLOC_S:1360.4445269428206
ME_READ_S:22341.960156249999
ME_READ_L:11545.2919921875
ME_WRITE:9952.6064453125
ME_LARGE:14487.5
ME_LATENCY:52.891450104725969
ME_THREADED:24418.75732421875
SUMM_CPU:3949.2746791240029
SUMM_ME:2119.8167112484348

================================================
Partial pcp output pmrep -p -a

     o.w.iteration  o.w.SUMM_ME  o.w.CPU_avx512  o.w.SUMM_CPU  o.w.ME_THREADED  o.w.ME_LATENCY  o.w.ME_LARGE  o.w.ME_WRITE  o.w.ME_READ_L  o.w.ME_READ_S  o.w.ME_ALLOC_S  o.w.m_CPU_enc_ECDSA  o.w.m_CPU_enc_AES  o.w.m_CPU_enc_SHA  o.w.CPU_avx  o.w.CPU_fma  o.w.CPU_sse  o.w.CPU_mm  o.w.CPU_MATRIX_MULT_SSE  o.w.CPU_PHYSICS  o.w.CPU_SINGLETHREAD  o.w.CPU_COMPRESSION  o.w.CPU_ENCRYPTION  o.w.CPU_SORTING  o.w.CPU_PRIME  o.w.CPU_FLOATINGPOINT_MATH  o.w.CPU_INTEGER_MATH

16:33:26 1.000 2119.817 4294.222 3949.275 24418.757 52.891 14487.500 9952.606 11545.292 22341.960 1360.445 453445788.206 3937983257.671 510818534.269 2577.061 3652.387 1235.218 344.081 2939.722 478.939 1803.238 49590.234 1558.383 5974.223 23.804

========================================
Screen output from the test.

passmark_out.txt

@dvalinrh dvalinrh requested a review from kdvalin January 12, 2026 17:05
@github-actions
Copy link

This relates to RPOPC-763

@dvalinrh dvalinrh changed the title Fix it so the pcp metrics are saved properly to the pcp log files. Add the open metrics for passmark results. Jan 13, 2026
@dvalinrh dvalinrh requested a review from malucius-rh January 13, 2026 11:27
Copy link
Member

@kdvalin kdvalin left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As a quick fix, this LGTM.

We may want to move the function to test_tools-wrapper so we can reduce duplicated code for resetting the metrics.

Comment on lines +7 to +32
CPU_INTEGER_MATH NaN
CPU_FLOATINGPOINT_MATH NaN
CPU_PRIME NaN
CPU_SORTING NaN
CPU_ENCRYPTION NaN
CPU_COMPRESSION NaN
CPU_SINGLETHREAD NaN
CPU_PHYSICS NaN
CPU_MATRIX_MULT_SSE NaN
CPU_mm NaN
CPU_sse NaN
CPU_fma NaN
CPU_avx NaN
CPU_avx512 NaN
m_CPU_enc_SHA NaN
m_CPU_enc_AES NaN
m_CPU_enc_ECDSA NaN
ME_ALLOC_S NaN
ME_READ_S NaN
ME_READ_L NaN
ME_WRITE NaN
ME_LARGE NaN
ME_LATENCY NaN
ME_THREADED NaN
SUMM_CPU NaN
SUMM_ME NaN
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We might want to consider how best to separate the test results from each benchmark.. Maybe openmetrics.workload.<test name>.<metric name> in future?

This is NOT needed to merge this PR

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Or testname_metricname, since dots are special for OM.

@malucius-rh
Copy link

As a quick fix, this LGTM.

We may want to move the function to test_tools-wrapper so we can reduce duplicated code for resetting the metrics.

There's a reset_pcp_om function available in pcp_commands which can reset the openmetrics values

Copy link

@malucius-rh malucius-rh left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@dvalinrh dvalinrh merged commit 9ddb586 into main Jan 14, 2026
5 checks passed
@dvalinrh dvalinrh deleted the fix_pcp branch January 14, 2026 13:54
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add test results metrics to pcp logs.

4 participants