url string | repository_url string | labels_url string | comments_url string | events_url string | html_url string | id int64 | node_id string | number int64 | title string | user dict | labels list | state string | locked bool | assignees list | milestone dict | comments int64 | created_at timestamp[ns, tz=UTC] | updated_at timestamp[ns, tz=UTC] | closed_at timestamp[ns, tz=UTC] | assignee dict | author_association string | type dict | active_lock_reason float64 | draft float64 | pull_request dict | body string | closed_by dict | reactions dict | timeline_url string | performed_via_github_app float64 | state_reason string | sub_issues_summary dict | issue_dependencies_summary dict | pinned_comment float64 | parent_issue_url string |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
https://api.github.com/repos/pytorch/pytorch/issues/177024 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177024/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177024/comments | https://api.github.com/repos/pytorch/pytorch/issues/177024/events | https://github.com/pytorch/pytorch/pull/177024 | 4,051,748,409 | PR_kwDOA-j9z87JYgFd | 177,024 | [DOCS] Fix typo "is wrong,please" -> "is wrong, please" | {
"avatar_url": "https://avatars.githubusercontent.com/u/253604130?v=4",
"events_url": "https://api.github.com/users/benediktjohannes/events{/privacy}",
"followers_url": "https://api.github.com/users/benediktjohannes/followers",
"following_url": "https://api.github.com/users/benediktjohannes/following{/other_us... | [
"open source",
"topic: not user facing"
] | open | false | [] | null | 1 | 2026-03-10T13:16:15 | 2026-03-10T13:16:21 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177024.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177024",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177024.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177024"
} | Fix typo "is wrong,please" -> "is wrong, please" | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177024/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177024/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177023 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177023/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177023/comments | https://api.github.com/repos/pytorch/pytorch/issues/177023/events | https://github.com/pytorch/pytorch/pull/177023 | 4,051,739,110 | PR_kwDOA-j9z87JYeKl | 177,023 | [ROCm] Directly access scalars if largeBar is enabled | {
"avatar_url": "https://avatars.githubusercontent.com/u/246502575?v=4",
"events_url": "https://api.github.com/users/pschlan-amd/events{/privacy}",
"followers_url": "https://api.github.com/users/pschlan-amd/followers",
"following_url": "https://api.github.com/users/pschlan-amd/following{/other_user}",
"gists_... | [
"module: rocm",
"open source",
"release notes: cuda"
] | open | false | [] | null | 2 | 2026-03-10T13:14:33 | 2026-03-10T13:16:22 | null | null | NONE | null | null | 1 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177023.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177023",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177023.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177023"
} | When using ROCm and the device has largeBar enabled, we can read scalars directly from GPU memory. That way, there's no need for a malloc and memcpy.
## Benchmark
Simple synthetic benchmark:
```python
import torch
import time
device = torch.device("cuda")
N = 1000000
tensor = torch.arange(N)
tensor = ten... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177023/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177023/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177022 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177022/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177022/comments | https://api.github.com/repos/pytorch/pytorch/issues/177022/events | https://github.com/pytorch/pytorch/issues/177022 | 4,051,663,881 | I_kwDOA-j9z87xf3wJ | 177,022 | DISABLED test_comprehensive_addmm_decomposed_cuda_float32 (__main__.TestInductorOpInfoCUDA) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: flaky-tests",
"skipped",
"oncall: pt2",
"module: inductor"
] | open | false | [] | null | 1 | 2026-03-10T13:01:19 | 2026-03-10T13:01:23 | null | null | NONE | null | null | null | null | Platforms: inductor
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_comprehensive_addmm_decomposed_cuda_float32&suite=TestInductorOpInfoCUDA&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66433358245).... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177022/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177022/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177020 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177020/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177020/comments | https://api.github.com/repos/pytorch/pytorch/issues/177020/events | https://github.com/pytorch/pytorch/issues/177020 | 4,051,315,481 | I_kwDOA-j9z87xeisZ | 177,020 | tensor cuda memory is not released when using nested_tensor_from_jagged | {
"avatar_url": "https://avatars.githubusercontent.com/u/22587711?v=4",
"events_url": "https://api.github.com/users/yiwangchunyu/events{/privacy}",
"followers_url": "https://api.github.com/users/yiwangchunyu/followers",
"following_url": "https://api.github.com/users/yiwangchunyu/following{/other_user}",
"gist... | [
"triage review",
"module: cuda",
"module: memory usage",
"module: nestedtensor",
"bot-triaged"
] | open | false | [] | null | 0 | 2026-03-10T12:01:19 | 2026-03-10T12:04:33 | null | null | NONE | null | null | null | null | ### 🐛 Describe the bug
I found it's interesting that the cuda memory can not be released by gc when I used `nested_tensor_from_jagged`.
```python
import torch
import gc
def run():
a_torch = torch.empty([1152, 32, 32], dtype=torch.float32, device="cuda")
torch.nested.nested_tensor_from_jagged(
... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177020/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177020/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177019 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177019/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177019/comments | https://api.github.com/repos/pytorch/pytorch/issues/177019/events | https://github.com/pytorch/pytorch/pull/177019 | 4,051,238,425 | PR_kwDOA-j9z87JW83P | 177,019 | [Inductor] Support reference-type OpaqueObject with CUDA graphs | {
"avatar_url": "https://avatars.githubusercontent.com/u/19503980?v=4",
"events_url": "https://api.github.com/users/kshitij12345/events{/privacy}",
"followers_url": "https://api.github.com/users/kshitij12345/followers",
"following_url": "https://api.github.com/users/kshitij12345/following{/other_user}",
"gist... | [
"open source",
"module: inductor",
"ciflow/inductor",
"ciflow/torchtitan"
] | open | false | [] | null | 3 | 2026-03-10T11:47:08 | 2026-03-10T13:11:30 | null | null | COLLABORATOR | null | null | 1 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177019.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177019",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177019.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177019"
} | Related: https://github.com/pytorch/pytorch/issues/176615
Reference-type opaque objects passed as custom_op args appear as FakeScriptObject in example_inputs at compile time, which failed the cudagraph eligibility check and silently disabled CUDA graphs.
Add FakeScriptObject to the eligibility check, relax the re... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177019/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177019/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177017 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177017/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177017/comments | https://api.github.com/repos/pytorch/pytorch/issues/177017/events | https://github.com/pytorch/pytorch/issues/177017 | 4,050,881,800 | I_kwDOA-j9z87xc40I | 177,017 | Incompatibility between FSDP and nn.utils.parametrize | {
"avatar_url": "https://avatars.githubusercontent.com/u/6229650?v=4",
"events_url": "https://api.github.com/users/BenjaminBossan/events{/privacy}",
"followers_url": "https://api.github.com/users/BenjaminBossan/followers",
"following_url": "https://api.github.com/users/BenjaminBossan/following{/other_user}",
... | [
"oncall: distributed",
"bot-triaged"
] | open | false | [] | null | 0 | 2026-03-10T10:36:46 | 2026-03-10T10:38:14 | null | null | NONE | null | null | null | null | ### 🐛 Describe the bug
I have a use case that requires using `nn.utils.parametrize` dynamically during the `forward` call. When combining this with FSDP, it leads to an error where a `parametrize` check fails because the parameter is present on the module when it shouldn't be, triggered here:
https://github.com/pyto... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177017/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177017/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177016 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177016/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177016/comments | https://api.github.com/repos/pytorch/pytorch/issues/177016/events | https://github.com/pytorch/pytorch/issues/177016 | 4,050,881,708 | I_kwDOA-j9z87xc4ys | 177,016 | [CUDA] nn.LSTM numerical inconsistency: Sample Independence Violation (NaN in batch vs valid individually) at float32 edge values | {
"avatar_url": "https://avatars.githubusercontent.com/u/181111767?v=4",
"events_url": "https://api.github.com/users/zifan6699/events{/privacy}",
"followers_url": "https://api.github.com/users/zifan6699/followers",
"following_url": "https://api.github.com/users/zifan6699/following{/other_user}",
"gists_url": ... | [
"needs reproduction",
"bot-triaged"
] | open | false | [] | null | 2 | 2026-03-10T10:36:44 | 2026-03-10T12:04:48 | null | null | NONE | null | null | null | null | ### 🐛 Describe the bug
I am reporting a severe discrepancy between the documented behavior of "Batched computations" and the actual results observed in nn.LSTM on CUDA.
According to PyTorch documentation, results from batched computations might be "slightly different" due to low-level optimizations. However, my findi... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177016/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177016/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177015 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177015/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177015/comments | https://api.github.com/repos/pytorch/pytorch/issues/177015/events | https://github.com/pytorch/pytorch/pull/177015 | 4,050,454,635 | PR_kwDOA-j9z87JUhTw | 177,015 | [FSDP2] Use separate NCCL communicator for reduce-scatter to enable AG/RS overlap | {
"avatar_url": "https://avatars.githubusercontent.com/u/134637289?v=4",
"events_url": "https://api.github.com/users/weifengpy/events{/privacy}",
"followers_url": "https://api.github.com/users/weifengpy/followers",
"following_url": "https://api.github.com/users/weifengpy/following{/other_user}",
"gists_url": ... | [
"release notes: distributed (fsdp)",
"ciflow/inductor",
"ciflow/torchtitan"
] | open | false | [] | null | 1 | 2026-03-10T09:14:51 | 2026-03-10T11:52:39 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177015.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177015",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177015.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177015"
} | Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom):
* __->__ #177015
@ngimel pointed that the same NCCL communicator serialize all-gather and reduce-scatter. Using different cuda streams are not enough. We must also use different PGs
This is confirmed with unit test and profiler traces be... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177015/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177015/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177014 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177014/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177014/comments | https://api.github.com/repos/pytorch/pytorch/issues/177014/events | https://github.com/pytorch/pytorch/issues/177014 | 4,050,413,193 | I_kwDOA-j9z87xbGaJ | 177,014 | `cudaErrorInvalidDeviceFunction` error due to use of `__CUDA_ARCH__` in host code | {
"avatar_url": "https://avatars.githubusercontent.com/u/309017?v=4",
"events_url": "https://api.github.com/users/Flamefire/events{/privacy}",
"followers_url": "https://api.github.com/users/Flamefire/followers",
"following_url": "https://api.github.com/users/Flamefire/following{/other_user}",
"gists_url": "ht... | [
"oncall: quantization",
"bot-triaged"
] | open | false | [] | null | 0 | 2026-03-10T09:06:23 | 2026-03-10T09:08:04 | null | null | COLLABORATOR | null | null | null | null | ### 🐛 Describe the bug
Calling e.g. `torch._convert_weight_to_int4pack` fails with `cudaErrorInvalidDeviceFunction` (only visible in subsequent checks like trying to print a CUDA tensor)
Reason is a potentially unsupported use of `__CUDA_ARCH__`
In this case the relevant code is https://github.com/pytorch/pytorch/b... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177014/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177014/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177013 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177013/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177013/comments | https://api.github.com/repos/pytorch/pytorch/issues/177013/events | https://github.com/pytorch/pytorch/issues/177013 | 4,050,400,954 | I_kwDOA-j9z87xbDa6 | 177,013 | [ONNX Export] TransformerEncoderLayer export fails when MHA fastpath is enabled | {
"avatar_url": "https://avatars.githubusercontent.com/u/18723104?v=4",
"events_url": "https://api.github.com/users/latstars/events{/privacy}",
"followers_url": "https://api.github.com/users/latstars/followers",
"following_url": "https://api.github.com/users/latstars/following{/other_user}",
"gists_url": "htt... | [
"module: onnx",
"triaged",
"enhancement",
"has workaround",
"bot-triaged"
] | open | false | [] | null | 0 | 2026-03-10T09:03:56 | 2026-03-10T09:08:20 | null | null | NONE | null | null | null | null | ### 🐛 Describe the bug
## Problem
PyTorch's TransformerEncoderLayer fails ONNX export.
The following script will produce errors.
```python
import torch
from torch.nn import TransformerEncoderLayer
class Model(torch.nn.Module):
def __init__(self):
super().__init__()
self.encoder_layer = Transformer... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177013/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177013/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177012 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177012/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177012/comments | https://api.github.com/repos/pytorch/pytorch/issues/177012/events | https://github.com/pytorch/pytorch/pull/177012 | 4,050,221,788 | PR_kwDOA-j9z87JTwAe | 177,012 | Accelerate SDPA on Arm CPUs: Update OpenBLAS to v0.3.32 | {
"avatar_url": "https://avatars.githubusercontent.com/u/115173828?v=4",
"events_url": "https://api.github.com/users/fadara01/events{/privacy}",
"followers_url": "https://api.github.com/users/fadara01/followers",
"following_url": "https://api.github.com/users/fadara01/following{/other_user}",
"gists_url": "ht... | [
"open source",
"topic: not user facing"
] | open | false | [] | null | 2 | 2026-03-10T08:26:39 | 2026-03-10T11:23:09 | null | null | COLLABORATOR | null | null | 1 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177012.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177012",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177012.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177012"
} | Stack from [ghstack](https://github.com/ezyang/ghstack/tree/0.12.0) (oldest at bottom):
* __->__ #177012
* #177009
* #176881
OpenBLAS v0.3.31 adds support for BGEMM on SVE128, SVE256 machines
and general optimizations for SBGEMM/BGEMM: https://github.com/OpenMathLib/OpenBLAS/pull/5419, https://github.com/OpenMat... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177012/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177012/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177011 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177011/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177011/comments | https://api.github.com/repos/pytorch/pytorch/issues/177011/events | https://github.com/pytorch/pytorch/pull/177011 | 4,050,186,107 | PR_kwDOA-j9z87JTofT | 177,011 | [FSDP] Use separate NCCL communicator for reduce-scatter to enable AG/RS overlap | {
"avatar_url": "https://avatars.githubusercontent.com/u/134637289?v=4",
"events_url": "https://api.github.com/users/weifengpy/events{/privacy}",
"followers_url": "https://api.github.com/users/weifengpy/followers",
"following_url": "https://api.github.com/users/weifengpy/following{/other_user}",
"gists_url": ... | [
"release notes: distributed (fsdp)",
"ciflow/inductor",
"ciflow/torchtitan"
] | open | false | [] | null | 1 | 2026-03-10T08:19:10 | 2026-03-10T11:01:28 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177011.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177011",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177011.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177011"
} | Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom):
* __->__ #177011
Give all-gather and reduce-scatter different ProcessGroups so they go through
separate NCCL communicators and can truly overlap in the backward pass. Previously
both used the same shard_process_group, which serialized them on ... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177011/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177011/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177010 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177010/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177010/comments | https://api.github.com/repos/pytorch/pytorch/issues/177010/events | https://github.com/pytorch/pytorch/issues/177010 | 4,050,146,768 | I_kwDOA-j9z87xaFXQ | 177,010 | [RFC] Introduce the unified RNG-related API to torch.accelerator | {
"avatar_url": "https://avatars.githubusercontent.com/u/106960996?v=4",
"events_url": "https://api.github.com/users/guangyey/events{/privacy}",
"followers_url": "https://api.github.com/users/guangyey/followers",
"following_url": "https://api.github.com/users/guangyey/following{/other_user}",
"gists_url": "ht... | [
"feature",
"triaged",
"module: random",
"module: accelerator",
"bot-triaged"
] | open | false | [] | null | 0 | 2026-03-10T08:11:43 | 2026-03-10T08:18:35 | null | null | COLLABORATOR | null | null | null | null | # Motivation
This RFC proposes adding unified RNG-related APIs to `torch.accelerator`, including:
- `torch.accelerator.initial_seed`
- `torch.accelerator.seed`
- `torch.accelerator.seed_all`
- `torch.accelerator.manual_seed`
- `torch.accelerator.manual_seed_all`
- `torch.accelerator.set_rng_state`
- `torch.accelerator.... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177010/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177010/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177009 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177009/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177009/comments | https://api.github.com/repos/pytorch/pytorch/issues/177009/events | https://github.com/pytorch/pytorch/pull/177009 | 4,050,030,031 | PR_kwDOA-j9z87JTHw4 | 177,009 | Accelerate SDPA on Arm CPUs: Unroll exp_sum and max_mul kernels | {
"avatar_url": "https://avatars.githubusercontent.com/u/115173828?v=4",
"events_url": "https://api.github.com/users/fadara01/events{/privacy}",
"followers_url": "https://api.github.com/users/fadara01/followers",
"following_url": "https://api.github.com/users/fadara01/following{/other_user}",
"gists_url": "ht... | [
"module: cpu",
"open source",
"topic: not user facing",
"ciflow/linux-aarch64"
] | open | false | [] | null | 3 | 2026-03-10T07:44:17 | 2026-03-10T09:27:55 | null | null | COLLABORATOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/177009.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/177009",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/177009.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/177009"
} | Stack from [ghstack](https://github.com/ezyang/ghstack/tree/0.12.0) (oldest at bottom):
* #177012
* __->__ #177009
* #176881
We noticed that (e.g. in whisper with: B=8, num_heads=20, seqlen=1500, head_dim=64), 40% of our time in scaled-dot-production-attention is spent in the `_exp_reduce_sum_fusion_kernel`.
So... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177009/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177009/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177007 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177007/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177007/comments | https://api.github.com/repos/pytorch/pytorch/issues/177007/events | https://github.com/pytorch/pytorch/issues/177007 | 4,049,847,115 | I_kwDOA-j9z87xY8NL | 177,007 | DISABLED test_nan_assert_float8_e4m3fn (__main__.ProcessGroupNCCLGroupTest) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"oncall: distributed",
"module: flaky-tests",
"skipped"
] | open | false | [] | null | 1 | 2026-03-10T07:01:41 | 2026-03-10T07:01:43 | null | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_nan_assert_float8_e4m3fn&suite=ProcessGroupNCCLGroupTest&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66367223823).
Over the past 6... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177007/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177007/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177006 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177006/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177006/comments | https://api.github.com/repos/pytorch/pytorch/issues/177006/events | https://github.com/pytorch/pytorch/issues/177006 | 4,049,846,973 | I_kwDOA-j9z87xY8K9 | 177,006 | DISABLED test_nan_assert_float32 (__main__.ProcessGroupNCCLGroupTest) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"oncall: distributed",
"module: flaky-tests",
"skipped"
] | open | false | [] | null | 1 | 2026-03-10T07:01:37 | 2026-03-10T07:01:40 | null | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_nan_assert_float32&suite=ProcessGroupNCCLGroupTest&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66391096529).
Over the past 6 hours... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177006/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177006/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177005 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177005/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177005/comments | https://api.github.com/repos/pytorch/pytorch/issues/177005/events | https://github.com/pytorch/pytorch/issues/177005 | 4,049,846,664 | I_kwDOA-j9z87xY8GI | 177,005 | DISABLED test_rms_norm_bwd_bfloat16_split_reductions_False_shape0_max_autotune_False_initial_xblock_1_add_1dim_True (__main__.NoMixOrderReductionTest) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"triaged",
"module: flaky-tests",
"skipped",
"oncall: pt2",
"module: inductor"
] | open | false | [] | null | 2 | 2026-03-10T07:01:35 | 2026-03-10T13:00:49 | null | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_rms_norm_bwd_bfloat16_split_reductions_False_shape0_max_autotune_False_initial_xblock_1_add_1dim_True&suite=NoMixOrderReductionTest&limit=100) and the most recent trunk [workflow log... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177005/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177005/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177004 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177004/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177004/comments | https://api.github.com/repos/pytorch/pytorch/issues/177004/events | https://github.com/pytorch/pytorch/issues/177004 | 4,049,846,354 | I_kwDOA-j9z87xY8BS | 177,004 | DISABLED test_rms_norm_bwd_bfloat16_split_reductions_False_shape0_max_autotune_False_initial_xblock_1_add_1dim_True (__main__.MixOrderReductionTest) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"triaged",
"module: flaky-tests",
"skipped",
"oncall: pt2",
"module: inductor"
] | open | false | [] | null | 2 | 2026-03-10T07:01:32 | 2026-03-10T13:00:46 | null | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_rms_norm_bwd_bfloat16_split_reductions_False_shape0_max_autotune_False_initial_xblock_1_add_1dim_True&suite=MixOrderReductionTest&limit=100) and the most recent trunk [workflow logs]... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177004/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177004/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177003 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177003/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177003/comments | https://api.github.com/repos/pytorch/pytorch/issues/177003/events | https://github.com/pytorch/pytorch/issues/177003 | 4,049,846,032 | I_kwDOA-j9z87xY78Q | 177,003 | DISABLED test_unbacked_symbol_closure (__main__.HigherOrderOpTestsWithCompiledAutograd) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"triaged",
"module: flaky-tests",
"skipped",
"oncall: pt2",
"module: inductor"
] | open | false | [] | null | 2 | 2026-03-10T07:01:29 | 2026-03-10T13:00:43 | null | null | NONE | null | null | null | null | Platforms: linux, slow
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_unbacked_symbol_closure&suite=HigherOrderOpTestsWithCompiledAutograd&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66390147776).
... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177003/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177003/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177002 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177002/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177002/comments | https://api.github.com/repos/pytorch/pytorch/issues/177002/events | https://github.com/pytorch/pytorch/issues/177002 | 4,049,845,875 | I_kwDOA-j9z87xY75z | 177,002 | DISABLED test_dynamic_lstm_cpp_serdes (__main__.CppSerdesTestExport) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: flaky-tests",
"skipped",
"oncall: pt2",
"oncall: export"
] | open | false | [] | null | 1 | 2026-03-10T07:01:26 | 2026-03-10T07:01:32 | null | null | NONE | null | null | null | null | Platforms: asan, linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_dynamic_lstm_cpp_serdes&suite=CppSerdesTestExport&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66366502665).
Over the past 6 ... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177002/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177002/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177001 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177001/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177001/comments | https://api.github.com/repos/pytorch/pytorch/issues/177001/events | https://github.com/pytorch/pytorch/issues/177001 | 4,049,845,695 | I_kwDOA-j9z87xY72_ | 177,001 | DISABLED test_memory_stats_of_multiple_generators_and_graphs (__main__.TestCuda) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: cuda",
"triaged",
"module: flaky-tests",
"skipped"
] | open | false | [] | null | 1 | 2026-03-10T07:01:23 | 2026-03-10T07:01:25 | null | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_memory_stats_of_multiple_generators_and_graphs&suite=TestCuda&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66391096518).
Over the p... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177001/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177001/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/177000 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/177000/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/177000/comments | https://api.github.com/repos/pytorch/pytorch/issues/177000/events | https://github.com/pytorch/pytorch/issues/177000 | 4,049,845,522 | I_kwDOA-j9z87xY70S | 177,000 | DISABLED test_mempool_expandable (__main__.TestMemPool) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: cuda",
"triaged",
"module: flaky-tests",
"skipped"
] | open | false | [] | null | 2 | 2026-03-10T07:01:19 | 2026-03-10T13:00:40 | null | null | NONE | null | null | null | null | Platforms: linux, slow
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_mempool_expandable&suite=TestMemPool&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66390147776).
Over the past 6 hours, it has... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/177000/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/177000/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176999 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176999/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176999/comments | https://api.github.com/repos/pytorch/pytorch/issues/176999/events | https://github.com/pytorch/pytorch/issues/176999 | 4,049,845,354 | I_kwDOA-j9z87xY7xq | 176,999 | DISABLED test_memory_stats_of_multiple_generators_and_graphs (__main__.TestCuda) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: cuda",
"triaged",
"module: flaky-tests",
"skipped"
] | closed | false | [] | null | 1 | 2026-03-10T07:01:17 | 2026-03-10T13:00:15 | 2026-03-10T13:00:15 | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_memory_stats_of_multiple_generators_and_graphs&suite=TestCuda&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66391096518).
Over the p... | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176999/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176999/timeline | null | completed | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176998 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176998/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176998/comments | https://api.github.com/repos/pytorch/pytorch/issues/176998/events | https://github.com/pytorch/pytorch/issues/176998 | 4,049,844,898 | I_kwDOA-j9z87xY7qi | 176,998 | DISABLED test_dynamic_lstm_retraceability_strict (__main__.RetraceExportTestExport) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: flaky-tests",
"skipped",
"oncall: pt2",
"oncall: export"
] | open | false | [] | null | 1 | 2026-03-10T07:01:14 | 2026-03-10T07:01:19 | null | null | NONE | null | null | null | null | Platforms: asan, linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_dynamic_lstm_retraceability_strict&suite=RetraceExportTestExport&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66365591712).
O... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176998/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176998/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176997 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176997/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176997/comments | https://api.github.com/repos/pytorch/pytorch/issues/176997/events | https://github.com/pytorch/pytorch/issues/176997 | 4,049,844,746 | I_kwDOA-j9z87xY7oK | 176,997 | DISABLED test_comprehensive_linalg_cond_cuda_float32 (__main__.TestInductorOpInfoCUDA) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: flaky-tests",
"skipped",
"oncall: pt2",
"module: inductor"
] | open | false | [] | null | 1 | 2026-03-10T07:01:11 | 2026-03-10T07:01:14 | null | null | NONE | null | null | null | null | Platforms: inductor
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_comprehensive_linalg_cond_cuda_float32&suite=TestInductorOpInfoCUDA&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66397747238).
O... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176997/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176997/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176996 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176996/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176996/comments | https://api.github.com/repos/pytorch/pytorch/issues/176996/events | https://github.com/pytorch/pytorch/issues/176996 | 4,049,844,409 | I_kwDOA-j9z87xY7i5 | 176,996 | DISABLED test_unexpected_c_return_events (__main__.TestPythonTracer) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: flaky-tests",
"skipped",
"oncall: profiler"
] | open | false | [] | null | 1 | 2026-03-10T07:01:08 | 2026-03-10T07:01:11 | null | null | NONE | null | null | null | null | Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_unexpected_c_return_events&suite=TestPythonTracer&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66366502781).
Over the past 6 hours,... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176996/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176996/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176995 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176995/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176995/comments | https://api.github.com/repos/pytorch/pytorch/issues/176995/events | https://github.com/pytorch/pytorch/issues/176995 | 4,049,844,254 | I_kwDOA-j9z87xY7ge | 176,995 | DISABLED test_dynamic_lstm (__main__.TestExport) | {
"avatar_url": "https://avatars.githubusercontent.com/in/40112?v=4",
"events_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/events{/privacy}",
"followers_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/followers",
"following_url": "https://api.github.com/users/pytorch-bot%5Bbot%5D/following{... | [
"module: flaky-tests",
"skipped",
"oncall: pt2",
"oncall: export"
] | open | false | [] | null | 1 | 2026-03-10T07:01:05 | 2026-03-10T07:01:10 | null | null | NONE | null | null | null | null | Platforms: asan, linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_dynamic_lstm&suite=TestExport&limit=100) and the most recent trunk [workflow logs](https://github.com/pytorch/pytorch/runs/66396005481).
Over the past 6 hours, it has been d... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176995/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176995/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176994 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176994/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176994/comments | https://api.github.com/repos/pytorch/pytorch/issues/176994/events | https://github.com/pytorch/pytorch/pull/176994 | 4,049,669,933 | PR_kwDOA-j9z87JR8GF | 176,994 | [Inductor] Improve materialization heuristic for a chain of computations | {
"avatar_url": "https://avatars.githubusercontent.com/u/23176869?v=4",
"events_url": "https://api.github.com/users/BoyuanFeng/events{/privacy}",
"followers_url": "https://api.github.com/users/BoyuanFeng/followers",
"following_url": "https://api.github.com/users/BoyuanFeng/following{/other_user}",
"gists_url"... | [
"ciflow/trunk",
"topic: not user facing",
"module: inductor",
"ciflow/inductor",
"ciflow/torchtitan"
] | open | false | [] | null | 1 | 2026-03-10T06:15:36 | 2026-03-10T11:52:43 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176994.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176994",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176994.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176994"
} | When inductor compiles a model without graph breaks, multi-user pointwise buffers (e.g. residual adds in transformers) may not be materialized because should_realize_on_reuse only checked num_reads > 4. For a residual add with 2 reads and 3 users, this never triggered — causing each subsequent reduction kernel to inlin... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176994/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176994/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176993 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176993/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176993/comments | https://api.github.com/repos/pytorch/pytorch/issues/176993/events | https://github.com/pytorch/pytorch/issues/176993 | 4,049,657,008 | I_kwDOA-j9z87xYNyw | 176,993 | Raspberry Pi 4 Illegal instruction | {
"avatar_url": "https://avatars.githubusercontent.com/u/32973228?v=4",
"events_url": "https://api.github.com/users/Warfields/events{/privacy}",
"followers_url": "https://api.github.com/users/Warfields/followers",
"following_url": "https://api.github.com/users/Warfields/following{/other_user}",
"gists_url": "... | [
"triage review",
"module: binaries",
"module: crash",
"module: arm",
"bot-triaged"
] | open | false | [] | null | 2 | 2026-03-10T06:11:53 | 2026-03-10T13:15:32 | null | null | NONE | null | null | null | null | ### 🐛 Describe the bug
Importing torch results in an illegal instruction being executed on a Raspberry Pi 4
```
swarfield@pi-server:~/Code/mumble-sip $ uvx --verbose pocket-tts
DEBUG uv 0.10.9
DEBUG Using request connect timeout of 10s and read timeout of 30s
DEBUG Searching for default Python interpreter in managed ... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176993/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176993/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176992 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176992/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176992/comments | https://api.github.com/repos/pytorch/pytorch/issues/176992/events | https://github.com/pytorch/pytorch/pull/176992 | 4,049,474,937 | PR_kwDOA-j9z87JRTgi | 176,992 | [shard prop] norm ops | {
"avatar_url": "https://avatars.githubusercontent.com/u/32071673?v=4",
"events_url": "https://api.github.com/users/pianpwk/events{/privacy}",
"followers_url": "https://api.github.com/users/pianpwk/followers",
"following_url": "https://api.github.com/users/pianpwk/following{/other_user}",
"gists_url": "https:... | [
"ciflow/inductor",
"release notes: distributed (dtensor)",
"ciflow/torchtitan"
] | open | false | [] | null | 1 | 2026-03-10T05:21:27 | 2026-03-10T08:24:02 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176992.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176992",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176992.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176992"
} | Stack from [ghstack](https://github.com/ezyang/ghstack/tree/0.13.0) (oldest at bottom):
* __->__ #176992
* #176991
* #176955
* #176821
* #176820
* #176819
| null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176992/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176992/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176991 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176991/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176991/comments | https://api.github.com/repos/pytorch/pytorch/issues/176991/events | https://github.com/pytorch/pytorch/pull/176991 | 4,049,474,804 | PR_kwDOA-j9z87JRTe1 | 176,991 | [shard prop] interpolate | {
"avatar_url": "https://avatars.githubusercontent.com/u/32071673?v=4",
"events_url": "https://api.github.com/users/pianpwk/events{/privacy}",
"followers_url": "https://api.github.com/users/pianpwk/followers",
"following_url": "https://api.github.com/users/pianpwk/following{/other_user}",
"gists_url": "https:... | [
"ciflow/inductor",
"release notes: distributed (dtensor)",
"ciflow/torchtitan"
] | open | false | [] | null | 1 | 2026-03-10T05:21:23 | 2026-03-10T08:24:14 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176991.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176991",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176991.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176991"
} | Stack from [ghstack](https://github.com/ezyang/ghstack/tree/0.13.0) (oldest at bottom):
* #176992
* __->__ #176991
* #176955
* #176821
* #176820
* #176819
| null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176991/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176991/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176989 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176989/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176989/comments | https://api.github.com/repos/pytorch/pytorch/issues/176989/events | https://github.com/pytorch/pytorch/pull/176989 | 4,049,366,622 | PR_kwDOA-j9z87JQ9Zy | 176,989 | Add supports for itertools.batched and itertools.combinations_with_re… | {
"avatar_url": "https://avatars.githubusercontent.com/u/6831019?v=4",
"events_url": "https://api.github.com/users/Kingwl/events{/privacy}",
"followers_url": "https://api.github.com/users/Kingwl/followers",
"following_url": "https://api.github.com/users/Kingwl/following{/other_user}",
"gists_url": "https://ap... | [
"open source",
"topic: not user facing",
"module: dynamo"
] | open | false | [] | null | 2 | 2026-03-10T04:45:25 | 2026-03-10T05:17:27 | null | null | NONE | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176989.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176989",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176989.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176989"
} | Adds Dynamo polyfill support for itertools.batched and itertools.combinations_with_replacement in torch/_dynamo/polyfills/itertools.py
#175869
- .venv/bin/python test/dynamo/test_functions.py FunctionTests.test_itertools_batched FunctionTests.test_itertools_combinations_with_replacement: Ok
- PYTORCH_TEST_WITH_... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176989/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176989/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176988 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176988/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176988/comments | https://api.github.com/repos/pytorch/pytorch/issues/176988/events | https://github.com/pytorch/pytorch/pull/176988 | 4,049,358,850 | PR_kwDOA-j9z87JQ7yA | 176,988 | [ROCm][inductor] Unload CUDA/HIP modules in static launcher to prevent GPU driver module table exhaustion | {
"avatar_url": "https://avatars.githubusercontent.com/u/165712832?v=4",
"events_url": "https://api.github.com/users/naromero77amd/events{/privacy}",
"followers_url": "https://api.github.com/users/naromero77amd/followers",
"following_url": "https://api.github.com/users/naromero77amd/following{/other_user}",
"... | [
"module: rocm",
"open source",
"module: inductor",
"ciflow/inductor",
"release notes: inductor",
"ciflow/rocm-mi300",
"ciflow/torchtitan"
] | open | false | [] | null | 2 | 2026-03-10T04:42:48 | 2026-03-10T08:05:25 | null | null | COLLABORATOR | null | null | 1 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176988.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176988",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176988.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176988"
} | This PR resolves GPU driver module table exhaustion encountered during exhaustive tuning of templated GEMMs on ROCm. In principle, the same problem can occur on CUDA. For transparency, Cursor was used to explore several potential fixes and this was the most expedient approach that resolved the issue
Without explicit... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176988/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176988/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176987 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176987/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176987/comments | https://api.github.com/repos/pytorch/pytorch/issues/176987/events | https://github.com/pytorch/pytorch/pull/176987 | 4,049,276,149 | PR_kwDOA-j9z87JQrKT | 176,987 | Auto-record structured trace logs for cache replay | {
"avatar_url": "https://avatars.githubusercontent.com/u/13564?v=4",
"events_url": "https://api.github.com/users/ezyang/events{/privacy}",
"followers_url": "https://api.github.com/users/ezyang/followers",
"following_url": "https://api.github.com/users/ezyang/following{/other_user}",
"gists_url": "https://api.... | [
"module: inductor",
"module: dynamo",
"ciflow/inductor",
"ciflow/torchtitan"
] | open | false | [] | null | 2 | 2026-03-10T04:12:30 | 2026-03-10T07:30:10 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176987.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176987",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176987.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176987"
} | Stack from [ghstack](https://github.com/ezyang/ghstack/tree/0.12.0) (oldest at bottom):
* __->__ #176987
Instead of hardcoding which structured trace artifacts to replay on cache
hit, automatically record trace_structured calls during cache-miss
compilation and replay them on cache hit. This uses a thread-local
Struct... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176987/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176987/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176986 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176986/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176986/comments | https://api.github.com/repos/pytorch/pytorch/issues/176986/events | https://github.com/pytorch/pytorch/issues/176986 | 4,049,156,112 | I_kwDOA-j9z87xWTgQ | 176,986 | Inductor failure with `view_as_complex` in fused Multi-Head Attention subgraph | {
"avatar_url": "https://avatars.githubusercontent.com/u/259246027?v=4",
"events_url": "https://api.github.com/users/psiwho/events{/privacy}",
"followers_url": "https://api.github.com/users/psiwho/followers",
"following_url": "https://api.github.com/users/psiwho/following{/other_user}",
"gists_url": "https://... | [
"triaged",
"module: complex",
"oncall: pt2",
"module: inductor",
"bot-triaged"
] | open | false | [] | null | 0 | 2026-03-10T03:31:39 | 2026-03-10T03:34:27 | null | null | NONE | null | null | null | null | ### 🐛 Describe the bug
## Summary
`torch.compile` with the Inductor backend raises a `RuntimeError: Tensor must have a stride divisible by 2 for all but last dimension` when attempting to fuse a subgraph containing Multi-Head Rotary Positional Embeddings (RoPE) and a subsequent `matmul`.
The error specifically occu... | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176986/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176986/timeline | null | null | {
"completed": 0,
"percent_completed": 0,
"total": 0
} | {
"blocked_by": 0,
"blocking": 0,
"total_blocked_by": 0,
"total_blocking": 0
} | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176985 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176985/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176985/comments | https://api.github.com/repos/pytorch/pytorch/issues/176985/events | https://github.com/pytorch/pytorch/pull/176985 | 4,049,038,441 | PR_kwDOA-j9z87JP6wk | 176,985 | [MPS] Error out if one tries to create `torch.cdouble` tensor | {
"avatar_url": "https://avatars.githubusercontent.com/u/2453524?v=4",
"events_url": "https://api.github.com/users/malfet/events{/privacy}",
"followers_url": "https://api.github.com/users/malfet/followers",
"following_url": "https://api.github.com/users/malfet/following{/other_user}",
"gists_url": "https://ap... | [
"topic: bug fixes",
"release notes: mps",
"ciflow/mps"
] | open | false | [] | null | 1 | 2026-03-10T02:51:18 | 2026-03-10T05:17:35 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176985.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176985",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176985.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176985"
} | Which is not supported, same as `double`
Fixes https://github.com/pytorch/pytorch/issues/176981 | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176985/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176985/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176984 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176984/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176984/comments | https://api.github.com/repos/pytorch/pytorch/issues/176984/events | https://github.com/pytorch/pytorch/pull/176984 | 4,049,021,053 | PR_kwDOA-j9z87JP3Lb | 176,984 | [dynamo] Suppot requires_grad_ and then backward call | {
"avatar_url": "https://avatars.githubusercontent.com/u/13822661?v=4",
"events_url": "https://api.github.com/users/anijain2305/events{/privacy}",
"followers_url": "https://api.github.com/users/anijain2305/followers",
"following_url": "https://api.github.com/users/anijain2305/following{/other_user}",
"gists_u... | [
"module: dynamo",
"ciflow/inductor",
"ciflow/torchtitan"
] | open | false | [] | null | 2 | 2026-03-10T02:45:04 | 2026-03-10T09:47:04 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176984.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176984",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176984.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176984"
} | Stack from [ghstack](https://github.com/ezyang/ghstack) (oldest at bottom):
* __->__ #176984
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @kadeng @chauhang @amjames @Lucaskabela @jataylo | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176984/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176984/timeline | null | null | null | null | null | null |
https://api.github.com/repos/pytorch/pytorch/issues/176983 | https://api.github.com/repos/pytorch/pytorch | https://api.github.com/repos/pytorch/pytorch/issues/176983/labels{/name} | https://api.github.com/repos/pytorch/pytorch/issues/176983/comments | https://api.github.com/repos/pytorch/pytorch/issues/176983/events | https://github.com/pytorch/pytorch/pull/176983 | 4,048,970,622 | PR_kwDOA-j9z87JPs8w | 176,983 | [DO NOT MERGE][ROCm][CI] Test different mi300 label | {
"avatar_url": "https://avatars.githubusercontent.com/u/107946068?v=4",
"events_url": "https://api.github.com/users/amdfaa/events{/privacy}",
"followers_url": "https://api.github.com/users/amdfaa/followers",
"following_url": "https://api.github.com/users/amdfaa/following{/other_user}",
"gists_url": "https://... | [
"module: rocm",
"open source",
"topic: not user facing",
"keep-going",
"ci-no-td",
"ciflow/rocm-mi300",
"ciflow/periodic-rocm-mi300",
"ciflow/inductor-rocm-mi300"
] | open | false | [] | null | 1 | 2026-03-10T02:27:23 | 2026-03-10T06:49:34 | null | null | CONTRIBUTOR | null | null | 0 | {
"diff_url": "https://github.com/pytorch/pytorch/pull/176983.diff",
"html_url": "https://github.com/pytorch/pytorch/pull/176983",
"merged_at": null,
"patch_url": "https://github.com/pytorch/pytorch/pull/176983.patch",
"url": "https://api.github.com/repos/pytorch/pytorch/pulls/176983"
} | cc @jeffdaily @sunway513 @jithunnair-amd @pruthvistony @ROCmSupport @jataylo @hongxiayang @naromero77amd @pragupta @jerrymannil @xinyazhang | null | {
"+1": 0,
"-1": 0,
"confused": 0,
"eyes": 0,
"heart": 0,
"hooray": 0,
"laugh": 0,
"rocket": 0,
"total_count": 0,
"url": "https://api.github.com/repos/pytorch/pytorch/issues/176983/reactions"
} | https://api.github.com/repos/pytorch/pytorch/issues/176983/timeline | null | null | null | null | null | null |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 5