Skip to content

Pull requests: pytorch/FBGEMM

New pull request New
528 Open 4,075 Closed
528 Open 4,075 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

report L per table cla signed fb-exported
#4819 opened Sep 4, 2025 by YanXiong-Meta Loading…
3
add the wasted tflops script (#1843) cla signed fb-exported
#4818 opened Sep 4, 2025 by sijiac Loading…
4
Migrate TBE UVM cache kernels to FBGEMM_LAUNCH_KERNEL cla signed fb-exported
#4817 opened Sep 4, 2025 by q10 Loading…
2
Add 2d-2d support to MXFP8 Grouped GEMM cla signed fb-exported
#4816 opened Sep 4, 2025 by danielvegamyhre Loading…
12
gather_scatter_bench : argparse -> click args cla signed fb-exported
#4815 opened Sep 4, 2025 by ionuthristodorescu Loading…
4
add loggings for num of rows cla signed fb-exported
#4814 opened Sep 3, 2025 by kathyxuyy Loading…
2
SWA attention backward optimization 1 cla signed fb-exported
#4808 opened Sep 3, 2025 by henrylhtsang Loading…
4
split dequant functions from kv_cache.cu cla signed fb-exported
#4787 opened Aug 28, 2025 by bottler Loading…
8
[fbgemm_gpu] CUDA 13 support cla signed
#4783 opened Aug 27, 2025 by q10 Loading…
1
remove cpu check and hardcoded row alignment to 8 cla signed fb-exported
#4781 opened Aug 27, 2025 by chenyuzhcy Loading…
2
remove FBGEMM_USE_GPU macro cla signed fb-exported
#4779 opened Aug 27, 2025 by emlin Loading…
9
Add export trace to UVM and cache benchmark (#1800) cla signed fb-exported
#4777 opened Aug 27, 2025 by spcyppt Loading…
2
[fbgemm_gpu] [WIP] Upgrade python versions cla signed
#4771 opened Aug 25, 2025 by q10 Loading…
1
Migrate linearize cache indices kernels to FBGEMM_LAUNCH_KERNEL cla signed fb-exported
#4768 opened Aug 25, 2025 by q10 Loading…
2
Enable tf32 precision acceleration for h100 cla signed fb-exported
#4767 opened Aug 24, 2025 by RandySheriff Loading…
6
training QPS improvement cla signed fb-exported
#4763 opened Aug 22, 2025 by kathyxuyy Loading…
2
apply unroll in backward module: rocm
#4747 opened Aug 20, 2025 by zhiding512 Loading…
4
[fbgemm_gpu] [TEST] ROCm build changes ciflow/rocm cla signed module: rocm
#4742 opened Aug 20, 2025 by q10 Loading…
2
reuse checkpont handle cla signed fb-exported
#4736 opened Aug 19, 2025 by jiayulu Loading…
3
bug fix when sent and recv sizes are both 0 cla signed fb-exported
#4730 opened Aug 19, 2025 by sunfish2010 Loading…
5
Backward optimizations + rocm and gfx950 build support cla signed module: rocm
#4729 opened Aug 19, 2025 by shbiswas834 • Draft
4
Make fp8 ck gemm support both OCP and FNUZ FP8 cla signed fb-exported
#4728 opened Aug 19, 2025 by Alkaid-Benetnash Loading…
5
Migrate sparse ops kernels to FBGEMM_LAUNCH_KERNEL, pt 5B cla signed fb-exported
#4726 opened Aug 19, 2025 by q10 Loading…
2
Enable fp32 tests for Windows cla signed
#4722 opened Aug 18, 2025 by cyyever • Draft
4
fix typo: quantize_packed_fp8_symmetric -> dequantize_packed_fp8_symmetric (#1740) cla signed fb-exported
#4717 opened Aug 15, 2025 by ColinPeppler Loading…
3
ProTip! Exclude everything labeled bug with -label:bug.