Closed crtrott closed 5 years ago
We have failing tests with Cuda through Clang4: the single vector spmv kernel fails as well as the spmv_mv for LayoutLeft. The sparse_spgemm tests seems to hang.
We don't care about clang4 for CUDA anymore. Minimum CUDA clang build is with 6.
We have failing tests with Cuda through Clang4: the single vector spmv kernel fails as well as the spmv_mv for LayoutLeft. The sparse_spgemm tests seems to hang.