Search results

Why does single-stage runs faster than multi-stage for triton_fused_softmax.py?

Status: Open.

#386 In xlite-dev/LeetCUDA;

· yurilxc opened on Sep 8, 2025

Shape mismatch for tl.load when using small num columns in triton_fused_softmax.py

Status: Open.

#385 In xlite-dev/LeetCUDA;

· yurilxc opened on Sep 8, 2025

hgemm-tn correctness

Status: Open.

#384 In xlite-dev/LeetCUDA;

· Adoni opened on Sep 5, 2025

elementwise 算子问题

Status: Open.

#381 In xlite-dev/LeetCUDA;

· Crucifixion-Fxl opened on Aug 27, 2025

sigmoid.cu文件修改Ss Ks之后，kernel输出与pytorch输出不能对齐

Status: Open.

#379 In xlite-dev/LeetCUDA;

· GeneralJing opened on Aug 18, 2025

Typo at kernels/nvidia-nsight/README.md

Status: Open.

#377 In xlite-dev/LeetCUDA;

· wili-65535 opened on Aug 17, 2025

可以贡献SwiGLU激活函数算子吗？

Status: Open.

#372 In xlite-dev/LeetCUDA;

· napleon-liu opened on Aug 8, 2025

flash_attn_mma_split_q.cu 存在对于K的gmem->smem重复拷贝的问题

stale

Status: Open.

#364 In xlite-dev/LeetCUDA;

· zhanggj821 opened on Jul 15, 2025