Skip to content

Pull requests: ROCm/FBGEMM

New pull request New
47 Open 104 Closed
47 Open 104 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

Fix output_type=BF16 test_backward_adagrad_large_dims test
#151 opened Mar 25, 2026 by avbokovoy Loading…
@avbokovoy
Fix test_cache_int32_overflow test failure on ROCm
#149 opened Mar 25, 2026 by avbokovoy Loading…
@avbokovoy
Make warp segment threshold consitent with host function
#148 opened Mar 24, 2026 by avbokovoy Loading…
@avbokovoy
Fix bug in momentum type declaration in HIP TBE kernel
#147 opened Mar 19, 2026 by aryaman-gupta Loading…
Generalizing the HIP-optimized warp_per_row kernel
#146 opened Mar 19, 2026 by aryaman-gupta • Draft
Fix UVM FP16 performance regression on MI300
#145 opened Mar 10, 2026 by avbokovoy Loading…
@avbokovoy
1
Implement pre-sorting, caching and contigous warp processing in group_index_select
#144 opened Mar 3, 2026 by avbokovoy Loading…
@avbokovoy
24
Split Kernel Optimization for group_index_select_or_add_2d_kernel enhancement New feature or request
#142 opened Feb 2, 2026 by aryaman-gupta Loading…
@aryaman-gupta
20
Implement cached member_id upper bound search enhancement New feature or request
#141 opened Feb 2, 2026 by avbokovoy Loading…
@avbokovoy
3
sync pytorch/FBGEMM to rocm/FBGEMM
#140 opened Jan 18, 2026 by keerthana-bidar Loading…
1 task
Implement asynchronous LDS loads for MI350 enhancement New feature or request
#138 opened Dec 19, 2025 by avbokovoy Loading…
@avbokovoy
7
Optimizations for index_select_scalar_cumsum_kernel
#137 opened Dec 16, 2025 by amd-wsung102 Loading…
1 task
@amd-wsung102
21
Optimize group_index_select_or_add_2d_kernel by adding a separate codepath for small embedding dimensions
#135 opened Dec 16, 2025 by aryaman-gupta Loading…
@liligwu
Fixes bug in one specialized HIP instantiation of the warp-per-row kernel
#134 opened Dec 5, 2025 by aryaman-gupta Loading…
1
group_index_select_or_add_2d_kernel optimization
#131 opened Nov 11, 2025 by shbiswas834 Loading…
Yandai/temp opt codegen
#129 opened Oct 28, 2025 by yadaish Loading…
1 task
Replace Vec4TAcc load with llvm_amdgcn_raw_buffer_load_fp32
#126 opened Sep 25, 2025 by huizzhan • Draft
1 task
fwd optimizations
#125 opened Sep 23, 2025 by shbiswas834 Loading…
Remove fwd and warmup from benchmark profiling
#124 opened Sep 17, 2025 by huizzhan • Draft
1 task
added malloc pitch on merged pool embedding
#123 opened Sep 11, 2025 by kudomcho Loading…
added optimized merged pool embedding script
#122 opened Sep 10, 2025 by kudomcho Loading…
apply unroll and prefetch optimization
#119 opened Aug 27, 2025 by zhiding512 Loading…
apply Vec4T on vbe forward
#118 opened Aug 27, 2025 by JaxChen29 Loading…
1 task
tuned grid size by reducing num_warps_per_threadblock to 4
#117 opened Aug 26, 2025 by kudomcho Loading…
1 task
apply Vec4T on vbe forward
#115 opened Aug 21, 2025 by JaxChen29 Loading…
1 task
ProTip! Filter pull requests by the default branch with base:main.