Franscobec/gpufetch
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[v0.30] Add uarch detection to AMD GPUs

Browse Source 
Similarly to NVIDIA and Intel GPUs, we now detect microarchitecture,
also with manufacturing process and specific chip name. We infer all
of this from the gfx name (in the code we use the term llvm_target),
altough it's not clear yet that this method is completely reliable (see
comments for more details). In the future we might want to replace that
with a better way. Once we have the gfx name, we *should* be able to
infer the specific chip, and from the chip we can easily infer the
microarchitecture.

This commit also includes some refactorings and code improvements on
the HSA backend.
This commit is contained in:
Dr-Noob
2025-10-15 08:23:28 +02:00
parent b29b17d14f
commit 5df85aea2c
7 changed files with 407 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
src/common/printer.cpp
View File

@@ -11,6 +11,7 @@
#include "../intel/uarch.hpp"
#include "../intel/intel.hpp"
#include "../hsa/hsa.hpp"
#include "../hsa/uarch.hpp"
#include "../cuda/cuda.hpp"
#include "../cuda/uarch.hpp"
@@ -490,10 +491,18 @@ bool print_gpufetch_amd(struct gpu_info* gpu, STYLE s, struct color** cs, struct
return false;
char* gpu_name = get_str_gpu_name(gpu);
char* gpu_chip = get_str_chip(gpu->arch);
char* uarch = get_str_uarch_hsa(gpu->arch);
char* manufacturing_process = get_str_process(gpu->arch);
char* sms = get_str_cu(gpu);
char* max_frequency = get_str_freq(gpu);
setAttribute(art, ATTRIBUTE_NAME, gpu_name);
if (gpu_chip != NULL) {
setAttribute(art, ATTRIBUTE_CHIP, gpu_chip);
}
setAttribute(art, ATTRIBUTE_UARCH, uarch);
setAttribute(art, ATTRIBUTE_TECHNOLOGY, manufacturing_process);
setAttribute(art, ATTRIBUTE_FREQUENCY, max_frequency);
setAttribute(art, ATTRIBUTE_STREAMINGMP, sms);