Closed JoseSantosAMD closed 7 months ago
My test failed with:
$ omniperf profile -n mix_test -- ~/mixbench-hip-ro
___ _ __
/ _ \ _ __ ___ _ __ (_)_ __ ___ _ __ / _|
| | | | '_ ` _ \| '_ \| | '_ \ / _ \ '__| |_
| |_| | | | | | | | | | | |_) | __/ | | _|
\___/|_| |_| |_|_| |_|_| .__/ \___|_| |_|
|_|
Omniperf version: 2.0.0-Tech-Preview
Profiler choice: rocprofv1
Path: /home/colramos/GitHub/omniperf/workloads/mix_test/MI200
Target: MI200
Command: /home/colramos/mixbench-hip-ro
Kernel Selection: None
Dispatch Selection: None
IP Blocks: All
Traceback (most recent call last):
File "./src/omniperf", line 51, in <module>
main()
File "./src/omniperf", line 40, in main
omniperf.run_profiler()
File "/home/colramos/GitHub/omniperf/src/utils/utils.py", line 45, in wrap_function
result = function(*args, **kwargs)
File "/home/colramos/GitHub/omniperf/src/omniperf_base.py", line 228, in run_profiler
profiler.run_profiling(self.__version["ver"], config.prog)
File "/home/colramos/GitHub/omniperf/src/utils/utils.py", line 45, in wrap_function
result = function(*args, **kwargs)
File "/home/colramos/GitHub/omniperf/src/omniperf_profile/profiler_rocprof_v1.py", line 75, in run_profiling
super().run_profiling(version, prog)
File "/home/colramos/GitHub/omniperf/src/omniperf_profile/profiler_base.py", line 333, in run_profiling
if self.__args.kernel_verbose > 5:
AttributeError: 'Namespace' object has no attribute 'kernel_verbose'
This PR removes kernel-vebose from profile. It now only shows up in analyze and database mode. Instead of overwriting names in existing csvs, it reads in the csvs to a dataframe and renames kernel names in those dataframes. In database mode these updated dataframes are used instead of the csvs,. They are uploaded with pymongo to upload to mongodb
Closes #193
This PR removes kernel-vebose from profile. It now only shows up in analyze and database mode. Instead of overwriting names in existing csvs, it reads in the csvs to a dataframe and renames kernel names in those dataframes. In database mode these updated dataframes are used instead of the csvs,. They are uploaded with pymongo to upload to mongodb