Graphe des révisions

65 Révisions

Auteur SHA1 Message Date
Ben Richard 21b23a08d6 Work around crash when profiling multi-process/multi-GPU application (#376)
* Fix crash in multi-GPU scenario

Exclude -o option when invoking rocprof so that each rocprof process
writes to a different .csv file. Concatenate into a single .csv file
when finished.

Signed-off-by: benrichard-amd <ben.richard@amd.com>

* Only combine csv files when using rocprofv2

rocprofv1 does not have separate csv files

Signed-off-by: benrichard-amd <ben.richard@amd.com>

* Fix indices in combined CSV file

Use ignore_index flag to ensure there are no duplicate indices.

Signed-off-by: benrichard-amd <ben.richard@amd.com>

* Fix Dispatch_ID column and remove unnamed column

-Pandas was inserting an unnamed column (index column)
-Overwrite the Dispatch_ID column so that every row is unique, starting at 0
-Remove fixup_rocprofv2_dispatch_ids as no longer needed

Signed-off-by: benrichard-amd <ben.richard@amd.com>

* Fix code formatting

Signed-off-by: benrichard-amd <ben.richard@amd.com>

* Fix code formatting (for real this time)

Signed-off-by: benrichard-amd <ben.richard@amd.com>

---------

Signed-off-by: benrichard-amd <ben.richard@amd.com>

[ROCm/rocprofiler-compute commit: 69e5c32d52]
2024-08-09 09:46:42 -04:00
coleramos425 68387da116 Comply to formatting
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 7046ea15bd]
2024-06-03 13:47:45 -05:00
coleramos425 48c7833d07 Re-implementing HBM stack / XCD info for incoming product sku
Co-authored-by: Nicholas Curtis <nicholas.curtis@amd.com>
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 295b344646]
2024-06-03 13:47:45 -05:00
coleramos425 219914fa9e Format CMake and Python
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 1f370c9fe7]
2024-05-10 09:07:40 -06:00
coleramos425 e6b2b03f32 Add support for --quiet flag to roofline
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: dba868973b]
2024-05-10 09:07:40 -06:00
coleramos425 9a0d5f0663 Replace deprecated roofline warning with logging helper function
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 7d34e80567]
2024-04-25 18:43:20 +00:00
Karl W Schulz 2653dc850b additional mod needed to support roofline binaries potentially
executing from two different locations

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: b5011ff0ae]
2024-04-22 09:00:18 -05:00
Karl W Schulz ab7db47eb6 update logic to detect roofline binaries in two alternate paths
depending on whether user is running within local clone or from form
install.

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 093a4511ee]
2024-04-21 14:53:14 -05:00
Karl W Schulz c7d3c861be update logic to detect VERSION file to accommodate rocm packaging;
check two locations to cover case where user is running within local
git clone directly or alternatively, from package install.

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 3c562588ff]
2024-04-21 14:53:14 -05:00
coleramos425 55ebe7059c Move dispatch id patch to proper util func in utils. Enable in rocprofv2 post-processing
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: e0556f32ab]
2024-04-01 14:30:21 -05:00
Karl W Schulz 19a9e0431c apply formatter
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: f4070e3654]
2024-03-15 14:04:49 -05:00
colramos-amd fa0cc81720 Update roofline distro_map for Ubuntu22.04 (#319)
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 2dcad38028]
2024-03-15 12:47:30 -05:00
Karl W Schulz 8d6a8e3b06 update logging formatter handling for consistent behavior with error messsages
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 69af2ffc83]
2024-03-14 17:37:30 -05:00
colramos-amd ff7751724b Fix missing argv in console_warning()
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: e7cc326d5a]
2024-03-12 16:44:02 -05:00
Karl W Schulz 28f027c7bf move set_locale_encoding() to occur after logging is setup - allows
function to use console_error() directly (https://github.com/AMDResearch/omniperf/pull/308#discussion_r1519914083)

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: e85037fdfb]
2024-03-11 14:19:01 -05:00
Karl W. Schulz 6b0a36cb27 Update src/utils/utils.py
Co-authored-by: Cole Ramos <colramos@amd.com>
Signed-off-by: Karl W. Schulz <koomie@users.noreply.github.com>

[ROCm/rocprofiler-compute commit: 5e21dc35fd]
2024-03-11 14:19:01 -05:00
Karl W. Schulz f287aa0fc8 Update src/utils/utils.py
Co-authored-by: Cole Ramos <colramos@amd.com>
Signed-off-by: Karl W. Schulz <koomie@users.noreply.github.com>

[ROCm/rocprofiler-compute commit: 8e45cc897f]
2024-03-11 14:19:01 -05:00
Karl W Schulz 9d7dde19b0 refactor setting of local encoding; remove from top-level omniperf
wrapper and push into base class via a companion utility function

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 27a57bdafb]
2024-03-11 14:19:01 -05:00
Karl W Schulz 2d84322ea1 update print_status() to use console logger; convert some info logs to debug logs
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 15c86ba424]
2024-03-11 14:19:01 -05:00
Karl W Schulz a9cda2ad33 add a exit keyword to console_error function; allows multiple error
lines prior to code exit

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 2b24375e8d]
2024-03-11 14:19:01 -05:00
Karl W Schulz 7756a5b00b updates for run_prof() routine:
* add a profileMode keyword option, used to indent profiling output

* update logic if forked process is not successful, output is
  displayed directly with INFO logging or lower; also dispaly output
  in ERROR mode if the process fails

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 377c642ac2]
2024-03-11 14:19:01 -05:00
Karl W Schulz ab3ce5ecb8 remove duplicate process output within run_prof()
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: dcdef37c34]
2024-03-11 14:19:01 -05:00
Karl W Schulz 89bc098c59 include ERROR delimiter directly for console_error() messsages
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 61267ee262]
2024-03-11 14:19:01 -05:00
Karl W Schulz a5e23e17fe add a profileMode option to capture_subprocess_output() - includes
indented output with profiler selection when enabled

Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: cc7a725ae8]
2024-03-11 14:19:01 -05:00
Karl W Schulz 587d55cd9d minor refactor of print_status()
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 6bb4210d94]
2024-03-11 14:19:01 -05:00
Karl W Schulz b948aa6ebf add indent support to console_log()
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 63129e4247]
2024-03-11 14:19:01 -05:00
colramos-amd 91434bcebf Fix dangling merge conflict
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 6c5a3623b6]
2024-03-11 14:19:01 -05:00
colramos-amd 98e5e16d0e Fix python formatting
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: f2aac37178]
2024-03-11 14:19:01 -05:00
colramos-amd cfdf288cba Dedicated logging function for each log level
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 233905071c]
2024-03-11 14:19:01 -05:00
Cole Ramos 4585750893 Force encoding before Omniperf execution (#299)
* Force encoding before Omniperf execution

Signed-off-by: colramos-amd <colramos@amd.com>

* Fix circular dependency in utils module

Signed-off-by: colramos-amd <colramos@amd.com>

---------

Signed-off-by: colramos-amd <colramos@amd.com>

[ROCm/rocprofiler-compute commit: f4511e7280]
2024-03-06 17:01:07 -06:00
colramos-amd ee3ba49168 Fix typo, patch issue in kernel filtering in standalone GUI, and error on version mismatch
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: ea9783f105]
2024-03-03 14:38:28 -06:00
coleramos425 9ebe1b804c Remove hardcoded gfx90a in gen_sysinfo util
Signed-off-by: coleramos425 <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 1c3e7598a6]
2024-03-01 15:50:42 -06:00
colramos-amd 9ff07b51ff Python formatting
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: d6f45411eb]
2024-03-01 15:50:42 -06:00
Nicholas Curtis dcb928b7c7 Re-implement machine-specs to contain documentation, units, etc. Print using tabulate
Signed-off-by: Nicholas Curtis <nicurtis@amd.com>


[ROCm/rocprofiler-compute commit: 200961ffd9]
2024-03-01 15:50:42 -06:00
Nicholas Curtis d91521f2dc typo
Signed-off-by: Nicholas Curtis <nicurtis@amd.com>


[ROCm/rocprofiler-compute commit: 18dfc35c14]
2024-03-01 15:50:42 -06:00
colramos-amd c27927e4ad Hide L2Banks in sysinfo and update roofline module
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: a70dc4999d]
2024-03-01 15:50:42 -06:00
colramos-amd 2ecc5a4246 Renaming global spec vars. Replacing L2Banks with total_l2_chan (#251)
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 86b6c8e848]
2024-03-01 15:50:42 -06:00
colramos-amd 5352eafde4 Reorganize the specs module to reduce duplicate code
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 3b1b8d7b5b]
2024-03-01 15:50:42 -06:00
Karl W Schulz 6ac883e20e more code formatting updates
Signed-off-by: Karl W Schulz <karl.schulz@amd.com>


[ROCm/rocprofiler-compute commit: 76873dd0c4]
2024-02-22 16:02:25 -06:00
Nick Curtis fba2e174fb Fix clocks (#246)
Signed-off-by: Nicholas Curtis <nicurtis@amd.com>
Co-authored-by: Nicholas Curtis <nicurtis@amd.com>

[ROCm/rocprofiler-compute commit: 852cc13f2a]
2024-02-13 20:22:11 -06:00
Nick Curtis 0dcd6e35dd Fix hbmBW calculation, (#243)
* Fix hbmBW calculation, currently we're reading the wrong value of the mclk out of rocm-smi

Signed-off-by: Nicholas Curtis <nicurtis@amd.com>

* Fix for earlier smi

Signed-off-by: Nicholas Curtis <nicurtis@amd.com>

---------

Signed-off-by: Nicholas Curtis <nicurtis@amd.com>
Co-authored-by: Nicholas Curtis <nicurtis@amd.com>

[ROCm/rocprofiler-compute commit: c60f031558]
2024-02-13 20:19:56 -06:00
colramos-amd 6bedf1ebf6 Fix logic for roofline field in sysinfo
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 7bb211a465]
2024-02-12 17:07:55 -06:00
colramos-amd 8721490218 Revert sbios fix. Catch commas
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 33a2cd34ce]
2024-02-08 12:52:07 -06:00
colramos-amd bd22355b96 Add numPipes built-in-var for updated metrics
Co-authored-by: Nick Curtis <nicholas.curtis@amd.com>
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: d16f58b797]
2024-02-06 16:19:40 -06:00
colramos-amd 875ab7257e Fix reporting of --roof-only runs in sys_info
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 9b464503fe]
2024-02-04 12:44:47 -06:00
colramos-amd c0e2d759dd profile: fix parsing tcc per channel on mi300 (#234)
Co-authored-by: fei.zheng <fei.zheng@amd.com>
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: f44ce21252]
2024-01-31 09:34:50 -06:00
colramos-amd d8608a07a2 Enable database mode
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: e1867d34d3]
2024-01-30 10:43:30 -06:00
colramos-amd edccad450c Patch roofline debug logging
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: a1bfa83764]
2024-01-29 10:22:37 -06:00
colramos-amd 3f294b5dc0 Add missing input file to Mi300 flattening
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: 911943acee]
2024-01-26 10:32:13 -06:00
colramos-amd 995429869a Add missing rocm6 and ubuntu2204 elements to roofline binary
Signed-off-by: colramos-amd <colramos@amd.com>


[ROCm/rocprofiler-compute commit: be606e6ccb]
2024-01-25 12:55:32 -06:00