Matthew Nicely comments

Results 113 comments of


                                            Matthew Nicely

[BUG] Illegal CUDA shared memory access in SM90 GEMM TMA Warpspecialized at ClusterBarrier::init

@kadeng did you resolve your issue?

[QST]Is this the complete set of valid parameters for performing fp16 matrix multiplication using tensor cores?

@zwshan There are no expectations that CUTLASS should match or exceed cuBLAS performance. The intent of CUTLASS is to provide developers with an _additional_ tool to cuBLAS to explore functionality...

[QST]Is this the complete set of valid parameters for performing fp16 matrix multiplication using tensor cores?

@zwshan has your issue been resolved?

[DOC] cute/02_layout_operations.md

Closing due to inactivity

[FEA] Can I use copy to store register value into shared memory?

Closing due to inactivitly

[QST] Any experience to help me optimize on a particular problem size(a small M)?

@HuaYZhao has your issue been resolved?

[QST] In markdown docs, the attached links used global path, it doesn't work well in all platforms

Closing due to inactivity

[BUG] Cutlass 3x gemms can't be compiled with clang

@ezhulenev can you provide the following versions? - Clang - CUDA - CUTLASS

[BUG] Cutlass 3x gemms can't be compiled with clang

Closing due to inactivity

[QST] how bank conflict in shared memory is fixed in depthwise conv

@yupatrick22 has your issue been resolved?