Add roofline PDF output to general profiling runs (#774)

Change when Roofline PDFs are generated- during general profiling and --roof-only profiling (skip only when --no-roof option is present)

---------

Signed-off-by: Carrie Fallows <Carrie.Fallows@amd.com>
This commit is contained in:
cfallows-amd
2025-06-25 01:19:28 -04:00
committato da GitHub
parent 181e4f44c7
commit 630bc149ff
3 ha cambiato i file con 6 aggiunte e 3 eliminazioni
+1
Vedi File
@@ -77,6 +77,7 @@ Full documentation for ROCm Compute Profiler is available at [https://rocm.docs.
* VGPR Writes
* Total FLOPs (consider fp6 and fp4 ops)
* Update Dash to >=3.0.0 (for web UI)
* Change when Roofline PDFs are generated- during general profiling and --roof-only profiling (skip only when --no-roof option is present)
### Resolved issues
+3 -1
Vedi File
@@ -196,7 +196,9 @@ an Instinct MI210 vs an Instinct MI250.
Additionally, you will notice a few extra files. An SoC parameters file,
``sysinfo.csv``, is created to reflect the target device settings. All
profiling output is stored in ``log.txt``. Roofline-specific benchmark
results are stored in ``roofline.csv``.
results are stored in ``roofline.csv`` and roofline plots are outputted into PDFs as
``empirRoof_gpu-0_[datatype1]_..._[datatypeN].pdf`` where datatypes requested through
--roofline-data-type option are listed in the file name.
.. code-block:: shell-session
+2 -2
Vedi File
@@ -89,7 +89,7 @@ class Roofline:
# Set roofline run parameters from args
if hasattr(self.__args, "path") and not run_parameters:
self.__run_parameters["workload_dir"] = self.__args.path
if hasattr(self.__args, "roof_only") and self.__args.roof_only:
if hasattr(self.__args, "no_roof") and self.__args.no_roof == False:
self.__run_parameters["is_standalone"] = True
if hasattr(self.__args, "kernel_names") and self.__args.kernel_names:
self.__run_parameters["include_kernel_names"] = True
@@ -104,7 +104,7 @@ class Roofline:
if self.__run_parameters["include_kernel_names"] and (
not self.__run_parameters["is_standalone"]
):
console_error("--roof-only is required for --kernel-names")
console_error("--kernel-names cannot be used with --no-roof option")
def roof_setup(self):
# Setup the workload directory for roofline profiling.