Skip to content

Pull requests: ROCm/pytorch

New pull request New
79 Open 2,637 Closed
79 Open 2,637 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

HipDNN batchnorm integration
#3010 opened Feb 27, 2026 by dnikolaev-amd Loading…
10
[Release/2.11] Do not compile nccl_device.h for ROCm
#3009 opened Feb 27, 2026 by ethanwee1 • Draft
1
Support for AMD GPU address sanitizer.
#3007 opened Feb 26, 2026 by carlobertolli Loading…
1
[NO CP][release/2.8][ROCm][inductor] Inductor heuristic upstream backports
#3006 opened Feb 26, 2026 by naromero77amd Loading…
4
Atomics on 16 bits: prevent reading 4 bytes for 2-byte locations.
#3005 opened Feb 26, 2026 by carlobertolli • Draft
1 task
1
[rocm7.2_internal_testing] update related_commit
#3003 opened Feb 26, 2026 by amd-sriram Loading…
@amd-sriram
1
[release/2.9] Enable expandable segments (#173330)
#3002 opened Feb 25, 2026 by jeffdaily Loading…
1
[release/2.10] Enable expandable segments (#173330)
#3001 opened Feb 25, 2026 by jeffdaily Loading…
1
[AUTOGENERATED] release/2.11_IFU_20260224
#3000 opened Feb 24, 2026 by pragupta Loading…
3
[release/2.9] Fix deprecated import and skip test_ring_flex_attention for Navi3X and Navi4X
#2994 opened Feb 19, 2026 by mstankov-amd Loading…
2
[AUTOGENERATED] develop_IFU_20260218
#2989 opened Feb 18, 2026 by pragupta Loading…
3
[Release/2.7] add torch.version.rocm, distinct from torch.version.hip
#2985 opened Feb 17, 2026 by amd-sriram • Draft
@amd-sriram
1
[Release/2.8] add torch.version.rocm, distinct from torch.version.hip (#168…
#2984 opened Feb 17, 2026 by amd-sriram • Draft
@amd-sriram
[release/2.9] add torch.version.rocm, distinct from torch.version.hip
#2983 opened Feb 17, 2026 by amd-sriram • Draft
@amd-sriram
1
[release/2.10] add mkl
#2979 opened Feb 12, 2026 by ethanwee1 Loading…
3
[release/2.9] add mkl
#2978 opened Feb 12, 2026 by ethanwee1 Loading…
3
[release/2.8] add mkl
#2977 opened Feb 12, 2026 by ethanwee1 Loading…
3
[release/2.7] add mkl
#2976 opened Feb 12, 2026 by ethanwee1 Loading…
2
[rocm7.2_internal_testing] Enable variable-length attention unit tests (#170969) and Bump AOTriton to 0.11.2b (#174105)
#2964 opened Feb 5, 2026 by xinyazhang Loading…
1
Fix test_call_count_tunableop to extract kernel names from profiler
#2954 opened Feb 3, 2026 by umarkovi-amd Loading…
1
[release/2.10] Reverting prefered hipBLASLt backend for RDNA 3 and RDNA 3.5 (Strix Point/Halo)
#2944 opened Jan 26, 2026 by slojosic-amd Loading…
1
[release/2.10] Add torch.backends.cuda.math_sdp.fp32_precision
#2941 opened Jan 26, 2026 by anatoliylitv Loading…
2
limit the total number of threads for triu_tril_kernel on ROCm
#2940 opened Jan 24, 2026 by dnikolaev-amd • Draft
2
Add missing kpack triton compile options
#2935 opened Jan 23, 2026 by leonling-ll • Draft
@leonling-ll
3
[release/2.8] Revert to prev sccache by ROCm
#2928 opened Jan 22, 2026 by jithunnair-amd Loading…
1
ProTip! Add no:assignee to see everything that’s not assigned.