-
Notifications
You must be signed in to change notification settings - Fork 758
Pull requests: pytorch/FBGEMM
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix HIP grid overflow in jagged_softmax_backward_cuda
cla signed
meta-exported
#5974
opened Jul 1, 2026 by
q10
Contributor
Loading…
Fix HIP grid overflow in jagged_softmax_forward_cuda
cla signed
meta-exported
#5973
opened Jul 1, 2026 by
q10
Contributor
Loading…
Apply stochastic rounding in the optimized HIP TBE backward kernel
cla signed
meta-exported
module: rocm
#5972
opened Jul 1, 2026 by
q10
Contributor
Loading…
Add two-level prefetch into the NBit autovec kernel
cla signed
meta-exported
#5971
opened Jul 1, 2026 by
ShuyangLiu
Loading…
Add two-level prefetch into the SVE kernels (#5970)
cla signed
meta-exported
#5970
opened Jul 1, 2026 by
ShuyangLiu
Loading…
Shared CPU two-level prefetch helpers + env knobs (#5969)
cla signed
meta-exported
#5969
opened Jun 30, 2026 by
ShuyangLiu
Loading…
Fix self-corrupting base conda in setup_miniconda (drop --update-deps) (#5967)
cla signed
meta-exported
#5967
opened Jun 30, 2026 by
gchalump
Contributor
Loading…
optimize sparse_permute_2d kernel (#2876)
cla signed
meta-exported
#5964
opened Jun 30, 2026 by
yaoyj11
Contributor
Loading…
Adaptive Regularization for sparse embeddings — TBE frontend dispatch + tests (#5962)
cla signed
meta-exported
#5962
opened Jun 29, 2026 by
hdmeta
Contributor
Loading…
Add UVM support for MTIA in TBE training ops and codegen
cla signed
meta-exported
#5955
opened Jun 25, 2026 by
crypt3lx2k
Loading…
Replace TBE type-dispatch macros with C++20 generic-lambda functions
cla signed
#5952
opened Jun 25, 2026 by
cyyever
Contributor
Loading…
Fix backward_dense health-check regression under the ci profile
cla signed
meta-exported
#5936
opened Jun 19, 2026 by
gchalump
Contributor
Loading…
Fix x86 CPU CI OOM by capping Inductor compile threads
cla signed
meta-exported
#5930
opened Jun 18, 2026 by
gchalump
Contributor
Loading…
Make enable_ssd_writeback outer check in FeatureEvict
cla signed
meta-exported
#5924
opened Jun 17, 2026 by
lizhe-ji
Loading…
Enable
use_subwarp_shuffle=True for CTA kernel on ROCm
ciflow/rocm
cla signed
meta-exported
module: rocm
#5917
opened Jun 17, 2026 by
spcyppt
Contributor
Loading…
Add RocksDB key-count and rows-written stats
cla signed
meta-exported
#5914
opened Jun 16, 2026 by
lizhe-ji
Loading…
Reserve temp_kv bucket space by miss count to avoid rehashing (#5912)
cla signed
meta-exported
#5912
opened Jun 16, 2026 by
meta-codesync
Bot
Loading…
Fix ROCm __syncthreads deadlock in compute_amax_and_quantize_kernel
ciflow/rocm
cla signed
meta-exported
module: rocm
#5894
opened Jun 12, 2026 by
q10
Contributor
Loading…
Make tbe.ssd.ssd_config canonical; ops_common is now a shim
cla signed
meta-exported
#5874
opened Jun 11, 2026 by
q10
Contributor
Loading…
dense_to_jagged_forward: realize total_L SymInt before empty
cla signed
meta-exported
#5873
opened Jun 11, 2026 by
haoyuz
Contributor
Loading…
Suppress type errors for Pyre upgrade
cla signed
meta-exported
#5872
opened Jun 10, 2026 by
spcyppt
Contributor
Loading…
[WIP] feat: Add autotune for jagged BMM Triton kernels
#5866
opened Jun 10, 2026 by
Rusty95
Loading…
Remove stale opcheck xsuccess entries (jagged/sparse/quantize)
cla signed
meta-exported
#5860
opened Jun 9, 2026 by
q10
Contributor
Loading…
Enable dirty-bit tracking through the sharded map
cla signed
meta-exported
#5852
opened Jun 9, 2026 by
lizhe-ji
Loading…
Add multithreading to table lookup (#5849)
cla signed
meta-exported
#5849
opened Jun 8, 2026 by
ShuyangLiu
Loading…
Previous Next
ProTip!
Adding no:label will show everything without a label.