As a 3rd party reviewing workloads in Grafana it would be nice to track/get better insights into how the app was invoked. App parms can make a huge difference in the profiling results.
For example, comparing babelstream data I see that a different rocm stack / CPU was present in each SUT, and some of the kernels are the same across the two data sets. I'd like to know babelstream parameters. What was on the command line when it was launched?
As a 3rd party reviewing workloads in Grafana it would be nice to track/get better insights into how the app was invoked. App parms can make a huge difference in the profiling results.
For example, comparing babelstream data I see that a different rocm stack / CPU was present in each SUT, and some of the kernels are the same across the two data sets. I'd like to know babelstream parameters. What was on the command line when it was launched?
cc: James Dezelle