frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity
86,972 Commits 5,420 Branches 1,235 Tags
2c5c793085ba1de70dd0f9bcd2ddad3d84f07153
Commit Graph

21 Commits

Author SHA1 Message Date
FFFrog
2c5c793085 [Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404) 
As the title stated

**Changes:**
- Add **record**, **query** and **enable_timing** check
- Add related tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151404
Approved by: https://github.com/albanD
 2025-04-25 20:15:04 +00:00
PyTorch MergeBot
67f75244ea Revert "[Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404)" 
This reverts commit c91acad73a11825c366c51fb1e91d7e1a47d3f9e.

Reverted https://github.com/pytorch/pytorch/pull/151404 on behalf of https://github.com/ZainRizvi due to Sorry but this is breaking internally. @albanD can you please help it get relanded? To validate the fixes internally, you can follow the instructions here: https://fburl.com/fixing-ghfirst-reverts ([comment](https://github.com/pytorch/pytorch/pull/151404#issuecomment-2830829368))
 2025-04-25 16:08:27 +00:00
Yu, Guangye
33c75cae0a Add torch.accelerator.device_index as accelerator's device switch context (#148864) 
# Motivation
We propose adding support for the Python with statement on `torch.accelerator.device_index` to enable device switching functionality. This enhancement would simplify writing device-agnostic code and provide benefits across all accelerators. Its device-specific counterparts include [`torch.cuda.device`](00199acdb8/torch/cuda/__init__.py (L482)) and  [`torch.cuda._DeviceGuard`](00199acdb8/torch/cuda/__init__.py (L469)).

**Design Philosophy**
It accepts either an `Int` or `None` as input. When `None` is passed, no device switch is performed. Supporting `None` is important for compatibility, as it's possible to encounter `None` values from `torch.device.index`.

Therefore, with this PR, we can do like this

```python
src = 0
dst = 1
# Set src to current device
torch.accelerator.set_device_index(src)
with torch.accelerator.device_index(dst):
    # Inside with statement, we set dst to current device
    assert torch.accelerator.get_device_index() == dst
# Here the current device should be src
assert torch.accelerator.get_device_index() == src
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/148864
Approved by: https://github.com/albanD
 2025-04-25 09:45:25 +00:00
FFFrog
c91acad73a [Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404) 
As the title stated

**Changes:**
- Add **record**, **query** and **enable_timing** check
- Add related tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151404
Approved by: https://github.com/albanD
 2025-04-24 01:28:09 +00:00
PyTorch MergeBot
9374064483 Revert "[Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404)" 
This reverts commit 783be8f93248ca3af24b968bdf84188f5a3257d1.

Reverted https://github.com/pytorch/pytorch/pull/151404 on behalf of https://github.com/malfet due to suspected of breaking linux builds and breaks internal tests as well ([comment](https://github.com/pytorch/pytorch/pull/151404#issuecomment-2819041756))
 2025-04-21 17:11:53 +00:00
PyTorch MergeBot
48761e9737 Revert "[Easy] Fix the function signature of torch.Event (#151221)" 
This reverts commit 92baeecbdd3fb717880485e529df4efb02627c9d.

Reverted https://github.com/pytorch/pytorch/pull/151221 on behalf of https://github.com/malfet due to This broke rocm tests, see 92baeecbdd (40818271233-box) ([comment](https://github.com/pytorch/pytorch/pull/151221#issuecomment-2816883409))
 2025-04-19 22:06:24 +00:00
FFFrog
92baeecbdd [Easy] Fix the function signature of torch.Event (#151221) 
As the title stated.

The difference between declaration and implemention.
declaration:
d5a19e4525/torch/_C/__init__.pyi.in (L157-L162)

Implementation:
d5a19e4525/torch/csrc/Event.cpp (L30-L32)

**Question**: Which one should we choose?
- Change enable_timing to False to be consistent with torch.cuda.Event
- Change enable_timing to True to avoid BC-break
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151221
Approved by: https://github.com/albanD
ghstack dependencies: #151226
 2025-04-19 11:56:37 +00:00
FFFrog
783be8f932 [Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404) 
As the title stated

**Changes:**
- Add **record**, **query** and **enable_timing** check
- Add related tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151404
Approved by: https://github.com/albanD
 2025-04-18 15:26:13 +00:00
PyTorch MergeBot
1ce7969e81 Revert "[Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404)" 
This reverts commit 90c5b86cd8fcbbe6ee7c46ad17a05767f884794b.

Reverted https://github.com/pytorch/pytorch/pull/151404 on behalf of https://github.com/clee2000 due to broke a cpp extension test? test_cpp_extensions_stream_and_event.py::TestCppExtensionStreamAndEvent::test_stream_event [GH job link](https://github.com/pytorch/pytorch/actions/runs/14519277500/job/40736981315) [HUD commit link](90c5b86cd8), bad TD ([comment](https://github.com/pytorch/pytorch/pull/151404#issuecomment-2813649667))
 2025-04-17 17:45:41 +00:00
FFFrog
90c5b86cd8 [Easy] Add more check for elapsedTime of torch.xxx.Event and torch.Event (#151404) 
As the title stated

**Changes:**
- Add **record**, **query** and **enable_timing** check
- Add related tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/151404
Approved by: https://github.com/albanD
 2025-04-17 15:30:12 +00:00
Catherine Lee
9092dd2e82 [CI] Disable some tests that are failing in periodic (#150059) 
Disabling some tests to restore periodic

nogpu avx512 timeout:
59f14d19ae (38492953496-box)

profiler failure: 7ae0ce6360 (38461255009-box)

test_accelerator failure:
87bfd66c3c (39476723746-box)
origin: 146098

test_overrides failure:
bf752c36da (39484562957-box)
origin: 146098

inductor cpu repro:
bb9c426024 (38447525659-box)

functorch eager transforms:
8f858e226b (39488068620-box)
f2cea01f71 (39555064878)
b5281a4a18 (39599355600)
either 148288 or 148261?

2ec9aceaeb/1

Pull Request resolved: https://github.com/pytorch/pytorch/pull/150059
Approved by: https://github.com/ZainRizvi, https://github.com/atalman, https://github.com/malfet
 2025-03-28 20:31:32 +00:00
Yu, Guangye
719938c77f Generalize pin memory logic for accelerator when non blocking copy happened (#143783) 
# Motivation
fix https://github.com/pytorch/pytorch/issues/143641
Generalize pin memory logic for accelerator when non-blocking copy happened. Each accelerator has its implementation on `empty_strided`. The accelerator which doesn't have pin memory mechanism could ignore or mimic when pin_out is True.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/143783
Approved by: https://github.com/EikanWang, https://github.com/albanD
ghstack dependencies: #144959
 2025-01-23 03:43:05 +00:00
Yu, Guangye
6de110b862 Support with statement on torch.Stream (#140138) 
# Motivation
We propose to support Python with statement on `torch.Stream`. This is a benefit for all accelerators when writing device-agnostic code. The device-specific stream will also be supported because they are generally derived from `torch.Stream`.

With this PR, we can do like this
```python
s1= torch.Stream()
# Set s1 to the current stream
torch.accelerator.set_stream(s1)
with torch.Stream() as s2:
    # Inside with statement, we set s2 to the current stream
    assert torch.accelerator.current_stream() == s2
# Here the current stream should be s1
assert torch.accelerator.current_stream() == s1
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/140138
Approved by: https://github.com/albanD
 2025-01-10 02:05:19 +00:00
Dmitry Rogozhkin
7314cf44ae torch/accelerator: fix device type comparison (#143541) 
This was failing without the fix:
```
python -c 'import torch; d=torch.device("xpu:0"); torch.accelerator.current_stream(d)'
```
with:
```
ValueError: xpu doesn't match the current accelerator xpu.
```

CC: @guangyey, @EikanWang

Pull Request resolved: https://github.com/pytorch/pytorch/pull/143541
Approved by: https://github.com/guangyey, https://github.com/albanD
 2024-12-23 10:54:53 +00:00
Yu, Guangye
45ac4ebf15 [RELAND] Add UTs for accelerator device-agnostic runtime APIs (#133572) 
# Motivation
This PR intends to add UTs for accelerator device-agnostic APIs.

# Additional Context
This PR is relanded. It is reverted because `torch.Event` doesn't support mps backend. We have fixed it in https://github.com/pytorch/pytorch/pull/142468. The previous commit is 952514f0c8

Pull Request resolved: https://github.com/pytorch/pytorch/pull/133572
Approved by: https://github.com/EikanWang, https://github.com/albanD
ghstack dependencies: #143171
 2024-12-16 02:18:41 +00:00
PyTorch MergeBot
1b3f8b7589 Revert "[RELAND] Add UTs for accelerator device-agnostic runtime APIs (#133572)" 
This reverts commit 209119424922b135fef39aba1f25da3b67f5879a.

Reverted https://github.com/pytorch/pytorch/pull/133572 on behalf of https://github.com/huydhn due to Sorry for reverting your change but the new test is still very flaky on MacOS even when it does not segfault anymore ([comment](https://github.com/pytorch/pytorch/pull/133572#issuecomment-2537256522))
 2024-12-11 21:47:18 +00:00
PyTorch MergeBot
cd50bd8477 Revert "[BE][accelerator] formalize API name {current,set}_device_{idx => index} (#140542)" 
This reverts commit fb02b40d27737213e0547dec0e30977dfc50f2f3.

Reverted https://github.com/pytorch/pytorch/pull/140542 on behalf of https://github.com/huydhn due to Sorry for reverting your change, but I need to revert this in order to revert https://github.com/pytorch/pytorch/pull/133572#issuecomment-2537204202 due to a conflict ([comment](https://github.com/pytorch/pytorch/pull/140542#issuecomment-2537253665))
 2024-12-11 21:44:23 +00:00
Xuehai Pan
fb02b40d27 [BE][accelerator] formalize API name {current,set}_device_{idx => index} (#140542) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/140542
Approved by: https://github.com/guangyey, https://github.com/albanD
 2024-12-11 17:57:56 +00:00
Yu, Guangye
2091194249 [RELAND] Add UTs for accelerator device-agnostic runtime APIs (#133572) 
# Motivation
This PR intends to add UTs for accelerator device-agnostic APIs.

# Additional Context
This PR is relanded. It is reverted because `torch.Event` doesn't support mps backend. We have fixed it in https://github.com/pytorch/pytorch/pull/142468. The previous commit is 952514f0c8

Pull Request resolved: https://github.com/pytorch/pytorch/pull/133572
Approved by: https://github.com/EikanWang, https://github.com/albanD
ghstack dependencies: #142468
 2024-12-11 02:04:52 +00:00
PyTorch MergeBot
a1c6cf7e9f Revert "Add UTs for accelerator device-agnostic runtime APIs (#133572)" 
This reverts commit 952514f0c8d8ff2e1719e0ca82b0d178a5c5ff45.

Reverted https://github.com/pytorch/pytorch/pull/133572 on behalf of https://github.com/malfet due to Sorry for reverting your PR, but it segfaults on MacOS ([comment](https://github.com/pytorch/pytorch/pull/133572#issuecomment-2530354401))
 2024-12-10 04:42:55 +00:00
Yu, Guangye
952514f0c8 Add UTs for accelerator device-agnostic runtime APIs (#133572) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/133572
Approved by: https://github.com/EikanWang, https://github.com/albanD
 2024-12-07 13:14:10 +00:00