-
-
-
https://github.com/DF4FM/CUDA-Agent/tree/main/results 有人fork的早，可以继续吃瓜
-
https://newsletter.semianalysis.com/p/inferencex-v2-nvidia-blackwell-vs -
https://github.com/patrick-toulme/justabyte/blob/main/cutile_blackwell_post/run_moe_dump_all.py -
https://github.com/facebookexperimental/triton/pull/866 Add distilled PTX ISA 9.1 knowledge base for Claude -
https://github.com/YJMSTR/cutile-flash-linear-attention -
https://github.com/triton-lang/triton/tree/main/python/triton/tools/triton_to_gluon_translater -
RewriteTensorPointerPass -
# NOTES(Chenggang): current `shape/strides` are `int64_t`, while `offsets/block_shape` are `int32_t` shape = _convert_to_ir_values(builder, shape)