forked from NVIDIA/cutlass
-
Notifications
You must be signed in to change notification settings - Fork 70
Pull requests: intel/sycl-tla
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Add split reduction kernel for Flash Attention decoding
#671
opened Dec 18, 2025 by
wuxun-zhang
Loading…
4 of 7 tasks
fix output shape for FMHA forward kernel
#670
opened Dec 18, 2025 by
wuxun-zhang
Loading…
1 of 3 tasks
Example changes for streamk and mixed dtype for new atom API
#665
opened Dec 17, 2025 by
rajprince-intel
Loading…
1 of 7 tasks
Support for cached KV and paged KV in the New Flash Attention Kernel
#661
opened Dec 8, 2025 by
ClarkChin08
Loading…
MXFP4/MXFP8/int4 weights support in CuTe interface MoE GEMM example
#640
opened Nov 21, 2025 by
sanchitintel
•
Draft
Added limitation check in gemm_with_epilogue_softmax example test
#636
opened Nov 19, 2025 by
kausikmaiti
•
Draft
[Experiment] Evaluate perf impact of striped vs. blocked SLM read/write 1D copy atoms
#631
opened Nov 15, 2025 by
sanchitintel
•
Draft
1 task done
Previous Next
ProTip!
Adding no:label will show everything without a label.