AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Upgrade ROCm-SMI to AMD SMI (#86)

Browse Source 
* Integrating amd-smi into rocprofiler-systems due to rocm-smi deprecation.
* No functionality changes to users other than naming conventions.
* New tracks available in perfetto- gpu busy percentage metrics now splits gfx busy into separate gfx, umc, and mm engine measurements.

---------

Signed-off-by: Carrie Fallows <Carrie.Fallows@amd.com>
Co-authored-by: David Galiffi <David.Galiffi@amd.com>
This commit is contained in:
cfallows-amd
2025-01-30 21:32:07 -05:00
committed by GitHub
parent af3a8f695f
commit 0c32dfd6bc
32 changed files with 533 additions and 363 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/conceptual/rocprof-sys-feature-set.rst
+1 -1
View File
@@ -52,7 +52,7 @@ GPU metrics
* HIP kernel tracing
* HSA API tracing
* HSA operation tracing
* System-level sampling (via rocm-smi)
* System-level sampling (via amd-smi)
* Memory usage
* Power usage
docs/data/rocprof-sys-perfetto.png
BIN
View File
Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 313 KiB

After

Width:  |  Height:  |  Size: 433 KiB

docs/how-to/configuring-runtime-options.rst
+7 -7
View File
@@ -229,7 +229,7 @@ Generating a default configuration file
ROCPROFSYS_USE_SAMPLING = false
ROCPROFSYS_USE_PROCESS_SAMPLING = true
ROCPROFSYS_USE_ROCM = true
ROCPROFSYS_USE_ROCM_SMI = true
ROCPROFSYS_USE_AMD_SMI = true
ROCPROFSYS_USE_KOKKOSP = false
ROCPROFSYS_USE_CODE_COVERAGE = false
ROCPROFSYS_USE_PID = true
@@ -384,7 +384,7 @@ Viewing the setting descriptions
| ROCPROFSYS_USE_OMPT | Enable support for OpenMP-Tools |
| ROCPROFSYS_TRACE | Enable perfetto backend |
| ROCPROFSYS_USE_PID | Enable tagging filenames with proces... |
| ROCPROFSYS_USE_ROCM_SMI | Enable sampling GPU power, temp, uti... |
| ROCPROFSYS_USE_AMD_SMI | Enable sampling GPU power, temp, uti... |
| ROCPROFSYS_USE_ROCM | Enable ROCM tracing |
| ROCPROFSYS_USE_SAMPLING | Enable statistical sampling of call-... |
| ROCPROFSYS_USE_PROCESS_SAMPLING | Enable a background thread which sam... |
@@ -461,11 +461,11 @@ Viewing components
| sampling_wall_clock | Wall-clock timing. Derived from statistic... |
| sampling_cpu_clock | CPU-clock timing. Derived from statistica... |
| sampling_percent | Fraction of wall-clock time spent in func... |
| sampling_gpu_power | GPU Power Usage via ROCm-SMI. Derived fro... |
| sampling_gpu_temp | GPU Temperature via ROCm-SMI. Derived fro... |
| sampling_gpu_busy | GPU Utilization (% busy) via ROCm-SMI. De... |
| sampling_vcn_busy | GPU VCN Utilization (% activity) via ROCm... |
| sampling_gpu_memory_usage | GPU Memory Usage via ROCm-SMI. Derived fr... |
| sampling_gpu_power | GPU Power Usage via AMD-SMI. Derived from... |
| sampling_gpu_temp | GPU Temperature via AMD-SMI. Derived from... |
| sampling_gpu_busy | GPU Utilization (% busy) via AMD-SMI. Der... |
| sampling_vcn_busy | GPU VCN Utilization (% activity) via AMD-... |
| sampling_gpu_memory_usage | GPU Memory Usage via AMD-SMI. Derived fro... |
|-----------------------------------|----------------------------------------------|
Viewing hardware counters
docs/how-to/instrumenting-rewriting-binary-application.rst
+1 -1
View File
@@ -246,7 +246,7 @@ view the help menu.
libprofiler.so
libpthread.so.0
libresolv.so.2
librocm_smi64.so
libamd_smi.so
librocmtools.so
librocprofiler64.so
libroctracer64.so
docs/how-to/sampling-call-stack.rst
+5 -5
View File
@@ -213,9 +213,9 @@ View the help menu of ``rocprof-sys-sample`` with the ``-h`` / ``--help`` option
[BACKEND OPTIONS] These options control region information captured w/o sampling or instrumentation
-I, --include [ all | kokkosp | mpip | mutex-locks | ompt | rcclp | rocm-smi | rocprofiler | roctracer | roctx | rw-locks | spin-locks ]
-I, --include [ all | kokkosp | mpip | mutex-locks | ompt | rcclp | amd-smi | rocprofiler | roctracer | roctx | rw-locks | spin-locks ]
Include data from these backends (count: unlimited)
-E, --exclude [ all | kokkosp | mpip | mutex-locks | ompt | rcclp | rocm-smi | rocprofiler | roctracer | roctx | rw-locks | spin-locks ]
-E, --exclude [ all | kokkosp | mpip | mutex-locks | ompt | rcclp | amd-smi | rocprofiler | roctracer | roctx | rw-locks | spin-locks ]
Exclude data from these backends (count: unlimited)
[HARDWARE COUNTER OPTIONS] See also: rocprof-sys-avail -H
@@ -293,7 +293,7 @@ The following snippets show how ``rocprof-sys-sample`` runs with various environ
ROCPROFSYS_TRACE=true
ROCPROFSYS_USE_PROCESS_SAMPLING=true
ROCPROFSYS_USE_RCCLP=true
ROCPROFSYS_USE_ROCM_SMI=true
ROCPROFSYS_USE_AMD_SMI=true
ROCPROFSYS_USE_ROCM=true
ROCPROFSYS_USE_SAMPLING=true
ROCPROFSYS_PROFILE=true
@@ -323,7 +323,7 @@ The following snippets show how ``rocprof-sys-sample`` runs with various environ
ROCPROFSYS_TRACE=true
ROCPROFSYS_USE_PROCESS_SAMPLING=true
ROCPROFSYS_USE_RCCLP=false
ROCPROFSYS_USE_ROCM_SMI=false
ROCPROFSYS_USE_AMD_SMI=false
ROCPROFSYS_USE_ROCM=false
ROCPROFSYS_USE_SAMPLING=true
ROCPROFSYS_PROFILE=true
@@ -354,7 +354,7 @@ Here is the full output from the previous
ROCPROFSYS_USE_OMPT=false
ROCPROFSYS_USE_PROCESS_SAMPLING=true
ROCPROFSYS_USE_RCCLP=false
ROCPROFSYS_USE_ROCM_SMI=false
ROCPROFSYS_USE_AMD_SMI=false
ROCPROFSYS_USE_ROCM=false
ROCPROFSYS_USE_SAMPLING=true
[rocprof-sys][dl][1785877] rocprofsys_main