ROCm / rocFFT

Next generation FFT implementation for ROCm
https://rocm.docs.amd.com/projects/rocFFT/en/latest/
Other
175 stars 84 forks source link

Compiler does not gracefully handle multi-core builds #463

Open alerque opened 8 months ago

alerque commented 8 months ago

Re:

https://github.com/ROCm/rocFFT/blob/1680fbde578bbc65d1abfbaadf53e49dd5e0494c/library/src/rtc_compile.cpp

This is what CPU utilization looks like:

20240226_11h22m05s_grim

Somehow the usage is artificially capped at 50/70% of each core rather than just getting the job done.