Skip to content

Pull requests: pytorch/FBGEMM

New pull request New
547 Open 4,233 Closed
547 Open 4,233 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Buid time optimize (part2) cla signed fb-exported meta-exported
#5000 opened Oct 13, 2025 by gchalump Loading…
2
Add Paged Attention to FMHA Cutlass Blackwell Forward kernel for fixed length cla signed fb-exported meta-exported
#4999 opened Oct 13, 2025 by sarithad-meta Loading…
2
Add several features. cla signed
#4998 opened Oct 13, 2025 by jiayus-nvidia Loading…
3
add monitroing metrics for dram cache perf -- metadata read & write cla signed fb-exported meta-exported
#4996 opened Oct 10, 2025 by kathyxuyy Loading…
1
FP8 Convolution Kernel cla signed fb-exported meta-exported
#4994 opened Oct 10, 2025 by jwfromm Loading…
2
support loading libraries when installing multiple FBGEMM targets cla signed fb-exported meta-exported
#4993 opened Oct 10, 2025 by q10 Loading…
3
Try adding device sync before attention. (#2008) cla signed fb-exported meta-exported
#4992 opened Oct 9, 2025 by jwfromm Loading…
2
Adding python api to support sync trigger evict cla signed fb-exported meta-exported
#4984 opened Oct 7, 2025 by EddyLXJ Loading…
2
[fbgemm_gpu] Fix CMake deps cla signed
#4978 opened Oct 6, 2025 by q10 Loading…
2
: Add double type to be supported by permute_1D_sparse_data cla signed fb-exported meta-exported
#4968 opened Oct 2, 2025 by Shuchangd Loading…
2
Build time optimiaztion cla signed fb-exported meta-exported
#4954 opened Sep 30, 2025 by gchalump Loading…
2
Switch back to NVIDIA/cutlass, and upgrade to v4.2.1 cla signed
#4949 opened Sep 30, 2025 by jasl Loading…
10
Gate invalid triton autotune configs in AOTInductor for GFX95+ cla signed fb-exported meta-exported
#4940 opened Sep 26, 2025 by JChunX Loading…
2
Back out "Update to use Python 3.9 syntax" cla signed fb-exported meta-exported
#4928 opened Sep 24, 2025 by q10 Loading…
3
forward performance tuning for MI350 cla signed module: rocm
#4925 opened Sep 24, 2025 by liligwu Loading…
5
more hipify v2 fixes (#4854) cla signed fb-exported meta-exported module: rocm
#4921 opened Sep 23, 2025 by q10 Loading…
13
Support bf16 in blackwell cutlass decode attention kernel cla signed fb-exported meta-exported
#4916 opened Sep 23, 2025 by Aya-ZIbra Loading…
3
Resolve wgrad grouped gemm relocation issue in fbgemm cla signed fb-exported meta-exported
#4915 opened Sep 23, 2025 by jiawenliu64 Loading…
2
- Clean torch.check cla signed fb-exported meta-exported
#4871 opened Sep 12, 2025 by flaviotruzzi Loading…
3
dequantize_fp8_cache_kernel: Move D=128 device-side-assertion check to host cla signed fb-exported meta-exported
#4869 opened Sep 12, 2025 by ColinPeppler Loading…
4
symmetric quantization to FBGEMM prefill token-wise FP8 (fixed) cla signed fb-exported meta-exported
#4868 opened Sep 12, 2025 by ColinPeppler Loading…
3
- Reland D75563906 ci-no-td cla signed fb-exported meta-exported
#4865 opened Sep 11, 2025 by flaviotruzzi Loading…
4
Migrate GenAI quantize kernels to FBGEMM_LAUNCH_KERNEL, pt 4 cla signed fb-exported
#4863 opened Sep 11, 2025 by q10 Loading…
2
Add cutlass decode kernel to TritonBench cla signed fb-exported meta-exported
#4853 opened Sep 10, 2025 by Aya-ZIbra Loading…
5
Bump ASMJIT to 1.18 cla signed
#4850 opened Sep 10, 2025 by cyyever • Draft
5
ProTip! What’s not been updated in a month: updated:<2025-09-13.