vrr/prima.cpp
2
0
Fork 0
mirror of https://github.com/Lizonghang/prima.cpp.git synced 2025-09-10 06:04:45 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

add f32, f16, q8, q4k speed test for cuda

Browse source
This commit is contained in:
Lizonghang 2024-11-10 23:41:13 +04:00
parent f4260bb346
commit 10f6f92c7e
3 changed files with 59 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

13
common/profiler.h
View file

@ -42,13 +42,16 @@ struct gpu_support {
struct gpu_props {
const char * name;
const char * description;
float memory_free; // in GB
float memory_total; // in GB
float metal_flops; // in GFLOPS
float cuda_flops; // in GFLOPS
float memory_free; // in GB
float memory_total; // in GB
float metal_flops; // in GFLOPS
float cuda_flops_f32; // in GFLOPS
float cuda_flops_f16; // in GFLOPS
float cuda_flops_q8; // in GFLOPS
float cuda_flops_q4k; // in GFLOPS
gpu_props()
: name(""), description(""), memory_free(0.0f), memory_total(0.0f), metal_flops(0.0f), cuda_flops(0.0f) {}
: name(""), description(""), memory_free(0.0f), memory_total(0.0f), metal_flops(0.0f), cuda_flops_f32(0.0f), cuda_flops_f16(0.0f), cuda_flops_q8(0.0f), cuda_flops_q4k(0.0f) {}
};
struct device_info {