Xiao Fu（付晓）

icon company Meituan & NUDT icon location Master degree in National University of Defence Technology (NUDT) & DL Software Engineer at Meituan Inc. Specialize in HPC and LLM inference.

Results 4 comments of Xiao Fu（付晓）

Very poor performance of cblas_dgemm_batch with BLIS

> The second operation, generically, is something like: `C_i = \sum_j A_ij B_ij`. You might want to think of this either a) as a sequence of `axpy`s, or b) as...

Puzzles about using FLASH_LU_piv

Thanks, I'll try it. What can I do to utilize the function 'FLASH_LU_piv' correctly?

Puzzles about using FLASH_LU_piv

Thanks, It is my fault for misunderstanding the function 'FLASH_LU_piv' usage. Not only the matrix 'A_FLASH' should be a hierarchical object, but also the matrix 'IPIV'.

Are task dependencies (depend clauses) in bolt working?

> This testing does not cover all compilers and architectures. If you failed this `depend` test with a newer GCC+BOLT, we would appreciate any information (e.g., GCC/Clang version, CPU, etc)...