OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity
95,310 Commits 8,304 Branches 150 Tags 1.5 GiB
master
Commit Graph

95310 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Frost Mitchell
8221ee6db9 [xpu] Fix type annotation for ProcessGroupXCCL (#166418) 
After #163049, this PR fixes the type annotations to match the actual implementation for ProcessGroupXCCL::Options.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/166418
Approved by: https://github.com/guangyey, https://github.com/ezyang
 2025-10-30 19:29:06 +00:00
Austin Morton
b939de26d1 Avoid writing temporary modules to disk (#157713) 
In some cases the warning from #147744 still gets emitted because [atexit hooks aren't called](https://github.com/python/cpython/pull/114279).

Even in those cases, if the atexit hooks _were_ called you could end up with issues due to the directory being deleted in one process, but still being used elsewhere.

It's better all round to load these modules entirely in-memory.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/157713
Approved by: https://github.com/xush6528
 2025-10-30 19:11:16 +00:00
Yuanyuan Chen
694db5f549 Use 'is' in callable comparisons (#166624) 
Just like we use `is/is not` for class comparisons, it is generally advised to use `is/is not` for comparisons against torch functions.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166624
Approved by: https://github.com/Lucaskabela, https://github.com/Skylion007
 2025-10-30 19:00:09 +00:00
Scott Wolchok
639a0b1239 Remove torch.distributed.tensor.OpSchema.has_symints (#163667) 
It appears to be unused based on `cd torch; rg has_symints`.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/163667
Approved by: https://github.com/xmfan, https://github.com/azahed98, https://github.com/albanD
ghstack dependencies: #162990
 2025-10-30 18:57:17 +00:00
FFFrog
398775a43e [CodeClean] Replace std::runtime_error with TORCH_CHECK (#165119) 
As the title stated.

**Changes**:
- torch/csrc/inductor(Part 2)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/165119
Approved by: https://github.com/janeyx99
ghstack dependencies: #165139
 2025-10-30 18:43:58 +00:00
FFFrog
fcd5f8c352 [CodeClean] Remove the Unused MACRO for AOT Inductor Runtime (#165139) 
As the title stated.

- AOTI_TORCH_CHECK depend on TORCH_CHECK_MSG which located in c10/util/Exception.h, which maybe break BC
- AOTI_TORCH_CHECK is not used everywhere
- STD_TORCH_CHECK have ABI check tests.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/165139
Approved by: https://github.com/Skylion007, https://github.com/janeyx99
 2025-10-30 18:43:58 +00:00
Edward Z. Yang
4acc66f119 Make PT2 compile backprop through custom op without autograd key a hard error (#166367) 
Signed-off-by: Edward Z. Yang <ezyang@meta.com>

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166367
Approved by: https://github.com/bdhirsh
 2025-10-30 18:43:07 +00:00
PyTorch MergeBot
8f40a0c634 Revert "address DDE in matmul decomp (#166541)" 
This reverts commit 90519402c2.

Reverted https://github.com/pytorch/pytorch/pull/166541 on behalf of https://github.com/atalman due to breaks internal test ([comment](https://github.com/pytorch/pytorch/pull/166541#issuecomment-3469382334))
 2025-10-30 18:11:33 +00:00
Nicolas De Carli
a5c3c08d10 [Pytorch] Use exp_u20 for aarch64's erf (#166594) 
Summary:
After a precision study, we concluded it is ok to use ACL's exp function on f32's erf()
We can keep erf inline this way.

Benchmarks show about 91% higher throughput when processing a tensor of 1M elements, compiling with clang-19:

Before:
f32 erf: 2539.179us
After:
f32 erf: 1329.063us

Test Plan:
Correctness:

buck2 test mode/opt //caffe2/test:test_ops
buck2 test mode/opt //caffe2/test:torch

Performance:

buck2 run mode/opt //caffe2/benchmarks/operator_benchmark/fb:operator_benchmark_test

Differential Revision: D85730452

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166594
Approved by: https://github.com/mcfi, https://github.com/fadara01
 2025-10-30 18:09:05 +00:00
Aaron Orenstein
a553ea9ea4 Fix missing symbol when printing guards (#165723) 
Fixes #165177

When converting guards to sources if we were unable to get the expected symbol from symbol_to_source then try to get it from var_to_sources.

I was unable to make a simpler repro than what was described in the issue (which relies on llama3 - so inappropriate for a unit test).

Pull Request resolved: https://github.com/pytorch/pytorch/pull/165723
Approved by: https://github.com/bobrenjc93
 2025-10-30 18:03:51 +00:00
fduwjj
ba71e9ca9a [DeviceMesh] Isolate pg creation logic in Device Mesh into a separate func _init_one_process_group (#166614) 
To makes pg cache change easier and code modularization, we isolate the logic of process group creation into a separate function named `_init_one_process_group`.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166614
Approved by: https://github.com/lw
 2025-10-30 17:57:41 +00:00
PyTorch MergeBot
694d205143 Revert "shrink_group implementation to expose ncclCommShrink API (#164518)" 
This reverts commit 311ea0dec0.

Reverted https://github.com/pytorch/pytorch/pull/164518 on behalf of https://github.com/atalman due to breaks internal builds Error: from logging_utils import ( ModuleNotFoundError: No module named 'logging_utils' ([comment](https://github.com/pytorch/pytorch/pull/164518#issuecomment-3469308568))
 2025-10-30 17:52:29 +00:00
eellison
629293f568 bucket all reduce (#166528) 
Bucket all reduce in bucketer, thanks to @IvanKobzarev's earlier pr.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166528
Approved by: https://github.com/IvanKobzarev
ghstack dependencies: #166527
 2025-10-30 17:12:34 +00:00
eellison
c37802a8c4 use multi-dtype bucketing (#166527) 
Make the bucketer use multi-dtype bucketing for all gathers.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166527
Approved by: https://github.com/IvanKobzarev, https://github.com/ezyang
 2025-10-30 16:54:49 +00:00
PyTorch MergeBot
0a3ac47c0a Revert "[user-streams] Fix stream graph output semantics (#164819)" 
This reverts commit f5cb9a4c68.

Reverted https://github.com/pytorch/pytorch/pull/164819 on behalf of https://github.com/atalman due to breaks CI ([comment](https://github.com/pytorch/pytorch/pull/164819#issuecomment-3469018283))
 2025-10-30 16:53:32 +00:00
Maggie Moss
e83be7042e Fix pyrefly errors on main (#166548) 
Fixes existing errors to keep noise from lintrunner to a minimum

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166548
Approved by: https://github.com/Lucaskabela, https://github.com/mlazos
 2025-10-30 16:47:27 +00:00
Simon Layton
fb545fb068 Add MXFP4 grouped gemm support via. FBGEMM kernels (#166530) 
Summary:

* Extend `_scaled_grouped_mm_v2` to include MXFP4 support
* Add testing to existing grouped routines

Test Plan:

```
pytest -svv -k "mxfp4 and group" test/test_scaled_matmul_cuda.py
```

Reviewers:

Subscribers:

Tasks:

Tags:
Signed-off-by: Simon Layton <simonlayton@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/166530
Approved by: https://github.com/drisspg
 2025-10-30 16:46:11 +00:00
Bin Bao
2df2c316e2 [devx] Fix invalid symbol definition emitted in fx_graph_runnable.py (#166529) 
Summary: When emitting symbolic variable definition in fx_graph_runnable.py, we need to check if a SymNode is actually an expression, so that we won't generate something like "s27*s53**2 = 36".

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166529
Approved by: https://github.com/mlazos
ghstack dependencies: #166432
 2025-10-30 16:40:12 +00:00
Bin Bao
08b0a8f11a [Inductor] Fix an inductor_provenance bug (#166432) 
Summary: Fix an inductor_provenance related error seen when running TORCH_COMPILE_DEBUG generated fx_graph_runnable.py.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166432
Approved by: https://github.com/mlazos
 2025-10-30 16:40:12 +00:00
PyTorch MergeBot
3f1824742c Revert "Fix comparing inductor actual strides vs bw graph for activations should not throw DDE. (#166277)" 
This reverts commit b2a0f90501.

Reverted https://github.com/pytorch/pytorch/pull/166277 on behalf of https://github.com/atalman due to Breaks internal executorch tests ([comment](https://github.com/pytorch/pytorch/pull/166277#issuecomment-3468696623))
 2025-10-30 15:49:23 +00:00
Isuru Fernando
bbb7d2270b [inductor] print 0.0 as 0 for triton (#164291) 
Fixes https://github.com/pytorch/pytorch/issues/164157
Fixes https://github.com/pytorch/pytorch/issues/164086

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164291
Approved by: https://github.com/bobrenjc93, https://github.com/mlazos
 2025-10-30 15:15:25 +00:00
Scott Wolchok
6a5a436624 DTensor: C++ compute_global_tensor_info (#162990) 
compute_global_tensor_info is on the hot path for DTensor.{from,to}_local. More incremental progress toward C++.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/162990
Approved by: https://github.com/ezyang
 2025-10-30 15:10:54 +00:00
Shawn Xu
ad559072db [triton][sigmoid] Fix kernel cache and serialization issue for triton sigmoid + CUDA kernel bug (#166568) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/166568
Approved by: https://github.com/minjang
 2025-10-30 14:54:54 +00:00
PyTorch MergeBot
ad02bd13df Revert "[user-streams] Add current stream source (#165211)" 
This reverts commit 79aee77381.

Reverted https://github.com/pytorch/pytorch/pull/165211 on behalf of https://github.com/atalman due to failure: test/test_python_dispatch.py::TestPythonDispatch::test_return_stream [GH job link](https://github.com/pytorch/pytorch/actions/runs/18942517662/job/54086481693) [HUD commit link](7563f61cc8) ([comment](https://github.com/pytorch/pytorch/pull/165211#issuecomment-3468332362))
 2025-10-30 14:34:43 +00:00
eellison
7563f61cc8 Make bucketing aware of collective LIFO semantics (#166324) 
In the initial pr for overlapping preserving bucketing, for a graph like:

```
def foo(...):
     ag = all_gather(...)
     hiding_compute = mm(...)
     wait(ag)
```

We would add dependencies from mm -> ag, and wait from wait -> hiding_compute, to prevent bucketing reordering these collectives so that overlap no long occurred. however, there is an additional way for bucketing to prevent overlap.

If we were to reorder another collective so the graph looked like:

```
def foo(...):
     ag = all_gather(...)
     ar = all_reduce(...)
     wait(ar)
     hiding_compute = mm(...)
     wait(ag)
```

Overlap would not occur, because the wait for the all reduce would also force realization of every collective enqueued on the same stream prior to the all reduce. NCCL uses a single stream per process group.

To model, we set a set a strict ordering of all collective starts, waits, and hiding compute initially when bucketing. Then, when trying to add a collective to a bucket, we will see if we interfere with overlap for all of the following possible bucketings:

[move collective start to bucket start, move bucket start to collective start] x [move collective wait to bucket wait x move bucket wait to collective wait].

For any of these positions, we check if overlap would have been interfered with because of stream queue semantics. Then, if not, we remove the moving start and wait from the constrained ordering of collectives, and see if it's topologically valid to merge the nodes.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166324
Approved by: https://github.com/IvanKobzarev
ghstack dependencies: #166309
 2025-10-30 13:37:00 +00:00
PyTorch MergeBot
fa8e073a4e Revert "[triton][sigmoid] Fix kernel cache and serialization issue for triton sigmoid + CUDA kernel bug (#166568)" 
This reverts commit d46d8d6f54.

Reverted https://github.com/pytorch/pytorch/pull/166568 on behalf of https://github.com/atalman due to Failed test/test_extension_utils.py::TestExtensionUtils::test_external_module_register_with_renamed_backend [GH job link](https://github.com/pytorch/pytorch/actions/runs/18931754443/job/54050880312) [HUD commit link](d46d8d6f54) ([comment](https://github.com/pytorch/pytorch/pull/166568#issuecomment-3468008894))
 2025-10-30 13:31:47 +00:00
PyTorch MergeBot
95b5534773 Revert "[user-streams] Track symbolic current stream (#165212)" 
This reverts commit a5335263d3.

Reverted https://github.com/pytorch/pytorch/pull/165212 on behalf of https://github.com/atalman due to test/test_rename_privateuse1_to_existing_device.py::TestRenamePrivateuseoneToExistingBackend::test_external_module_register_with_existing_backend [GH job link](https://github.com/pytorch/pytorch/actions/runs/18930365446/job/54046768884) [HUD commit link](a5335263d3) ([comment](https://github.com/pytorch/pytorch/pull/165212#issuecomment-3467968796))
 2025-10-30 13:24:56 +00:00
PyTorch MergeBot
9ee1afbf66 Revert "[user-streams] Handle returning the current stream with/without device index (#165356)" 
This reverts commit f1af679270.

Reverted https://github.com/pytorch/pytorch/pull/165356 on behalf of https://github.com/atalman due to test/test_rename_privateuse1_to_existing_device.py::TestRenamePrivateuseoneToExistingBackend::test_external_module_register_with_existing_backend [GH job link](https://github.com/pytorch/pytorch/actions/runs/18930365446/job/54046768884) [HUD commit link](a5335263d3) ([comment](https://github.com/pytorch/pytorch/pull/165356#issuecomment-3467967061))
 2025-10-30 13:22:24 +00:00
PyTorch MergeBot
f60751024e Revert "[2/N] Add strict parameter to Python zip calls (#166257)" 
This reverts commit 39e5cdddf7.

Reverted https://github.com/pytorch/pytorch/pull/166257 on behalf of https://github.com/atalman due to Failing: test/distributed/fsdp/test_fsdp_mixed_precision.py::TestFSDPTrainEval::test_train_ema_eval_flow [GH job link](https://github.com/pytorch/pytorch/actions/runs/18934047991/job/54057218160) [HUD commit link](39e5cdddf7) ([comment](https://github.com/pytorch/pytorch/pull/166257#issuecomment-3467955332))
 2025-10-30 13:20:00 +00:00
Yuanyuan Chen
2de4cf2102 [1/N] Remove unused loop variables (#166258) 
This PR removes unused loop variables.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166258
Approved by: https://github.com/Lucaskabela, https://github.com/mlazos
 2025-10-30 12:22:25 +00:00
linhaifeng
369f2d6951 [3/N] fix typo in other folders (#166606) 
fix typo in other folders

#166374
#166126

_typos.toml
```bash
[files]
extend-exclude = ["tools/linter/dictionary.txt"]
[default.extend-words]
nd = "nd"
arange = "arange"
Nd = "Nd"
GLOBALs = "GLOBALs"
hte = "hte"
iy = "iy"
PN = "PN"
Dout = "Dout"
optin = "optin"
gam = "gam"
PTD = "PTD"
Sur = "Sur"
nin = "nin"
tme = "tme"
inpt = "inpt"
mis = "mis"
Raison = "Raison"
ouput = "ouput"
nto = "nto"
Onwer = "Onwer"
callibrate = "callibrate"
ser = "ser"
Metdata = "Metdata"
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166606
Approved by: https://github.com/ezyang
 2025-10-30 10:30:40 +00:00
Zhang, Jianyi
32920926f0 [xpu][fix] [Inductor] Avoid using tl.sqrt_rn on XPU before triton is ready (#165740) 
Fixes #165738

Pull Request resolved: https://github.com/pytorch/pytorch/pull/165740
Approved by: https://github.com/etaf, https://github.com/EikanWang, https://github.com/chuanqi129, https://github.com/desertfire
 2025-10-30 09:24:24 +00:00
Yuanyuan Chen
39e5cdddf7 [2/N] Add strict parameter to Python zip calls (#166257) 
This PR adds `strict=True/False` to zip calls in test utils. strict=True is passed when possible.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166257
Approved by: https://github.com/janeyx99
 2025-10-30 08:10:10 +00:00
libohao1201
2829d48bd1 [xpu][test][1/N] Port 3 fsdp distributed test cases to Intel GPU (#161476) 
For https://github.com/pytorch/pytorch/issues/114850, we will port 3 distributed tests to Intel GPU.
We could enable Intel GPU with the following methods and try the best to keep the original code styles:

- use "torch.accelerator.current_accelerator()" to determine the accelerator backend
- use "requires_accelerator_dist_backend" to enable "xccl"
- enabled XPU for some test path
- skip some test cases that Intel GPU does not support

Pull Request resolved: https://github.com/pytorch/pytorch/pull/161476
Approved by: https://github.com/weifengpy, https://github.com/guangyey
 2025-10-30 07:30:04 +00:00
Michael Lazos
f1af679270 [user-streams] Handle returning the current stream with/without device index (#165356) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/165356
Approved by: https://github.com/anijain2305
ghstack dependencies: #164304, #164522, #164819, #165211, #165212
 2025-10-30 07:20:25 +00:00
Shawn Xu
d46d8d6f54 [triton][sigmoid] Fix kernel cache and serialization issue for triton sigmoid + CUDA kernel bug (#166568) 
Differential Revision: D85792537

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166568
Approved by: https://github.com/minjang
 2025-10-30 06:17:39 +00:00
Michael Lazos
a5335263d3 [user-streams] Track symbolic current stream (#165212) 
merge into stream tests

Pull Request resolved: https://github.com/pytorch/pytorch/pull/165212
Approved by: https://github.com/anijain2305
ghstack dependencies: #164304, #164522, #164819, #165211
 2025-10-30 04:58:53 +00:00
Michael Lazos
79aee77381 [user-streams] Add current stream source (#165211) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/165211
Approved by: https://github.com/anijain2305
ghstack dependencies: #164304, #164522, #164819
 2025-10-30 04:58:53 +00:00
Michael Lazos
f5cb9a4c68 [user-streams] Fix stream graph output semantics (#164819) 
Preivously, we would stash a single stream value we constructed at trace time in a global and return the same value from repeated calls to the graph.

With this PR, we construct the stream value in advance, reference the constructed value in the graph via the lookup table, and if that value is returned as an output, read the value from the lookup table and return it (in bytecode, not as a graph output, since we don't support arbitrary stream outputs).

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164819
Approved by: https://github.com/anijain2305
ghstack dependencies: #164304, #164522
 2025-10-30 04:58:46 +00:00
PyTorch UpdateBot
f20bf77874 [audio hash update] update the pinned audio hash (#166597) 
This PR is auto-generated nightly by [this action](https://github.com/pytorch/pytorch/blob/main/.github/workflows/nightly.yml).
Update the pinned audio hash.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/166597
Approved by: https://github.com/pytorchbot
 2025-10-30 04:28:30 +00:00
Nicolas Macchioni
75f798e05b [inductor][mi350] add tech specs for MI350 (#166576) 
Summary:
was digging through matmul padding for other work, and I noticed that the compute bound checking won't work on MI350 since we haven't supplied the tech specs yet.

I added MI350 specs following the predefined format

Test Plan: CI

Differential Revision: D85804980

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166576
Approved by: https://github.com/leitian
 2025-10-30 03:46:52 +00:00
Angel Li
476b149a00 bwd pass (#164504) 
**Summary**
This implements the backward pass for the Varlen API and registers `_varlen_attn()` as a custom op.

**Benchmarking**

To benchmark, we compare runtime and TFLOPs against the current SDPA approach with padding.

Settings:

- 1 H100 machine
- `batch_size=8`, `max_seq_len=2048`, `embed_dim=1024`, `num_heads=16`
- dtype `torch.bfloat16`
- `is_causal=False`
- for variable length, we set sequences to be random multiples of 64 up to `max_seq_len`
- 100 runs

|        | Variable Length API | SDPA     |
|--------|--------------------|----------|
| Runtime | 0.8189142608642578 ms       | 3.263883056640625 ms  |
| TFLOPs | 268.652       | 158.731  |

We can see that runtime for Varlen is >3x faster

**Testing**

Run `python test/test_varlen_attention.py` for unit tests where we verify basic functionality and confirm numerical match between varlen gradients vs SDPA.

For custom op testing, `test_custom_op_registration` uses logging mode to verify that `_varlen_attn()` was called and tests with `torch.compile`. `test_custom_op_compliances` uses `torch.library.opcheck()` to verify.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164504
Approved by: https://github.com/drisspg
 2025-10-30 03:46:37 +00:00
Justin Chu
845da9c817 [ONNX] Ignore pyrefly errors in torchlib (#166588) 
Fixes #166475

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166588
Approved by: https://github.com/titaiwangms
 2025-10-30 03:43:52 +00:00
xinan.lin
0918bf321c [xpu][test] Reuse native_mm and mix_order_reduction for Intel GPU. (#166384) 
This PR reused native_mm and mix_order_reduction for Intel GPU and enabled the corresonding test.
Fixes #165370

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166384
Approved by: https://github.com/jansel
 2025-10-30 03:38:35 +00:00
Laith Sakka
90519402c2 address DDE in matmul decomp (#166541) 
Address https://github.com/pytorch/pytorch/issues/165081
Pull Request resolved: https://github.com/pytorch/pytorch/pull/166541
Approved by: https://github.com/mlazos
 2025-10-30 03:19:29 +00:00
Dzmitry Huba
791ca80d3a Enable local tensor mode for DTensor attention and convolution tests (#166406) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/166406
Approved by: https://github.com/ezyang
 2025-10-30 02:48:02 +00:00
Yuanyuan Chen
5cbdade914 Fix a syntactic error in test_indexing.py (#166390) 
This PR fixes a syntactic error in test_indexing.py by a misplaced `if else` expression.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166390
Approved by: https://github.com/jerryzh168
 2025-10-30 02:28:01 +00:00
amdfaa
0187db88d4 [ROCm][CI] Create periodic-rocm-mi200.yml (#166544) 
* We are separating out the rocm jobs of the periodic workflow
* We are introducing a new label `ciflow/periodic-rocm-mi200` to allow us to run distributed tests only on ROCm runners, without triggering many other jobs on the `periodic.yml` workflow (via `ciflow/periodic`)
* This new workflow will also be triggered via the `ciflow/periodic`, thus maintaining the old status quo.
* We are reverting to the `linux.rocm.gpu.4` label since it targets a lot more CI nodes at this point than the K8s/ARC-based `linux.rocm.gpu.mi250.4` label, as that is still having some network/scaling issues.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/166544
Approved by: https://github.com/jeffdaily

Co-authored-by: Jeff Daily <jeff.daily@amd.com>
 2025-10-30 02:08:07 +00:00
Bruce Chang
311ea0dec0 shrink_group implementation to expose ncclCommShrink API (#164518) 
Closes #164529

To expose the new [ncclCommShrink](https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/api/comms.html#ncclcommshrink) API to PyTorch.

This is useful when you need to exclude certain GPUs or nodes from a collective operation, for example in fault tolerance scenarios or when dynamically adjusting resource utilization.

For more info:  [Shrinking a communicator](https://docs.nvidia.com/deeplearning/nccl/user-guide/docs/usage/communicators.html#shrinking-a-communicator)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164518
Approved by: https://github.com/kwen2501
 2025-10-30 01:50:54 +00:00
dependabot[bot]
cf7756da38
Bump uv from 0.9.5 to 0.9.6 in /.ci/lumen_cli (#166578) 
Bumps [uv](https://github.com/astral-sh/uv) from 0.9.5 to 0.9.6.
- [Release notes](https://github.com/astral-sh/uv/releases)
- [Changelog](https://github.com/astral-sh/uv/blob/main/CHANGELOG.md)
- [Commits](https://github.com/astral-sh/uv/compare/0.9.5...0.9.6)

---
updated-dependencies:
- dependency-name: uv
  dependency-version: 0.9.6
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
 2025-10-29 18:28:14 -07:00