search

clMathLibraries / clBLAS

a software library containing BLAS functions written in OpenCL
Apache License 2.0
843 stars 237 forks source link

clBLAS Performance on AMD APU #328

Open jiahaipeng opened 6 years ago

jiahaipeng commented 6 years ago

I want to know how the performance of CLBLAS on AMD embeded APU, such as RX-416GD.

I tested clBLAS on AMD RX-416GD, the performance of sgemm only up to 123GFlops. However, the peak performance of RX-416GD(GPU) is 480GFlops.

I want to know why the performance of sgemm is low or some mistakes occurs when I install clblas.

Thanks

tingxingdong commented 6 years ago

clBLAS is not tuned or optimized for embedded APU but for discrete GPU like R9 Fury Nano. I do not think any mistakes here.