Closed matrix97317 closed 7 months ago
I have solved the above problem. Main Method: Select cuSparseLt Version. But I find a new problem for cuSparseLt 0.4.0
Hi @matrix97317, we need more context to understand the problem. First, my suggestion is to use the latest version of the library. Then, can you please provide the GPU model that you are using? are you using the example without modifications?
GPU: gtx3080ti GPU Driver: 520.61.05 CUDA: 11.8 CusparseLt: 0.4.0 Code: https://github.com/NVIDIA/CUDALibrarySamples/tree/master/cuSPARSELt/matmul_advanced
I don't change any example code.
Please try with the latest version 0.5.2.
gtx3080ti
is out of scope for our testing. Our targets are datacenter GPUs, not desktop ones. Anyway, we will run a quick check on this GPU model.
Thanks for your reply. For the latest version 0.5.2, it's correct.
Code: https://github.com/NVIDIA/CUDALibrarySamples/tree/master/cuSPARSELt/matmul_advanced