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
95,180 Commits 8,304 Branches 150 Tags 1.5 GiB
369f2d6951
Commit Graph

263 Commits

Author SHA1 Message Date
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
Natalia Gimelshein
37c6087334 Add split-K control to cuBLAS reduced-precision settings (#164766) 
## Summary
- add a CuBLASReductionOption enum so the CUDA context can track reduced-precision and split-K options
- extend the Python bindings, backend helpers, and docs to accept an optional allow_splitk argument for fp16/bf16 matmul controls
- update cuBLAS/cuBLASLt call sites plus dynamo guards and tests to respect the new combinations

## Testing
- python test/test_cuda.py TestCuda.test_cublas_allow_fp16_reduced_precision_reduction_get_set -v *(fails: ModuleNotFoundError: No module named 'psutil')*

------
https://chatgpt.com/codex/tasks/task_e_68e404623178832f8a3e1d34e1e175da

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164766
Approved by: https://github.com/malfet, https://github.com/albanD
 2025-10-08 18:48:45 +00:00
Yuanyuan Chen
5103ecc5d8 [1/N] Fix clang-tidy readability checks (#164561) 
Check all `.cpp` files except `jit` files for readability thoroughly.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164561
Approved by: https://github.com/Skylion007
 2025-10-04 09:40:38 +00:00
Lakshay Garg
f006aee601 Speed up FP precision lookup (#164044) 
This commit simplifies the precision lookup and setting logic
by reducing the number of branches and using a custom hash
function. Fixes #161822. The issue described in #163709 still
persists. This is meant as a short term fix.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164044
Approved by: https://github.com/ngimel, https://github.com/eqy
 2025-10-03 21:35:20 +00:00
PyTorch MergeBot
2a7c486750 Revert "Speed up FP precision lookup (#164044)" 
This reverts commit 723ba21393.

Reverted https://github.com/pytorch/pytorch/pull/164044 on behalf of https://github.com/yangw-dev due to broke internal build In file included from xplat/caffe2/aten/src/ATen/DeviceAccelerator.cpp:1: xplat/caffe2/aten/src/ATen/Context.h:502:38: error: shift count >= width of type [-Werror,-Wshift-count-overflow] 502 | return std::hash<size_t>{}((k1 << 32) | k2); ([comment](https://github.com/pytorch/pytorch/pull/164044#issuecomment-3363016702))
 2025-10-02 21:00:44 +00:00
Lakshay Garg
723ba21393 Speed up FP precision lookup (#164044) 
This commit simplifies the precision lookup and setting logic
by reducing the number of branches and using a custom hash
function. Fixes #161822. The issue described in #163709 still
persists. This is meant as a short term fix.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/164044
Approved by: https://github.com/ngimel, https://github.com/eqy
 2025-10-02 00:59:19 +00:00
Animesh Jain
991e3d0d16 [dynamo][guards] Revert introduction of different types of lambda_guards (#163385) 
With
https://fb.workplace.com/groups/260102303573409/permalink/787294574187510/
issue, it might be a better idea to just speedup _realize_dict and keep
the changes very local. So reverting this PR as well, to return to clean
slate.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/163385
Approved by: https://github.com/jansel
 2025-09-27 18:20:48 +00:00
Karhou Tam
39df24fe04 [Code Clean] Replace std::runtime_error with TORCH_CHECK (#163610) 
Including:
- `torch/csrc/instruction_counter`
- `torch/csrc/lazy`
- `torch/csrc/monitor`
- `torch/csrc/profiler`
- `torch/csrc/dynamo`

Fixes part of #148114

Personal mistake about (PR #163317), this PR does the same thing **and PR #163317 has already been approved by @albanD.**

This is a personal mistake on my part, and I'm so sorry about that. Hope you won't mind @albanD. 🥹

Pull Request resolved: https://github.com/pytorch/pytorch/pull/163610
Approved by: https://github.com/albanD, https://github.com/Skylion007
 2025-09-26 04:52:48 +00:00
PyTorch MergeBot
32ad29b72a Revert "[dynamo][guards] Fail on an unknown framelocals to dict conversion (#162695)" 
This reverts commit a8432bcaad.

Reverted https://github.com/pytorch/pytorch/pull/162695 on behalf of https://github.com/anijain2305 due to internal failure at https://fburl.com/workplace/qiitdlp6 ([comment](https://github.com/pytorch/pytorch/pull/162695#issuecomment-3310757225))
 2025-09-19 06:18:27 +00:00
PyTorch MergeBot
1302637a23 Revert "[dynamo][guards] Do not construct entire framelocals dict for LAMBDA_GUARD (#162525)" 
This reverts commit 5f630d28d7.

Reverted https://github.com/pytorch/pytorch/pull/162525 on behalf of https://github.com/anijain2305 due to internal tests fail ([comment](https://github.com/pytorch/pytorch/pull/162525#issuecomment-3310748980))
 2025-09-19 06:15:28 +00:00
joshuamarkovic
559e8d1c20 [doc]: Small typos (#162982) 
Small typo fixes

Pull Request resolved: https://github.com/pytorch/pytorch/pull/162982
Approved by: https://github.com/ezyang, https://github.com/zou3519
 2025-09-16 17:42:19 +00:00
Isuru Fernando
79d2418b5a [inductor] Add FLOAT_IS_NAN and COMPLEX_IS_NAN guards (#162537) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/162537
Approved by: https://github.com/anijain2305, https://github.com/mlazos
ghstack dependencies: #162528
 2025-09-12 04:32:46 +00:00
Isuru Fernando
5dd84559a5 [dynamo] Add DUAL_LEVEL_MATCH C++ guard (#162528) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/162528
Approved by: https://github.com/anijain2305
 2025-09-12 04:32:46 +00:00
Animesh Jain
a8432bcaad [dynamo][guards] Fail on an unknown framelocals to dict conversion (#162695) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/162695
Approved by: https://github.com/williamwen42
ghstack dependencies: #162694
 2025-09-11 15:01:00 +00:00
Animesh Jain
a3a40cb741 [dynamo][guards] Do not consturct framelocals to dict on GlobalsGuardAccessor (#162694) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/162694
Approved by: https://github.com/williamwen42
 2025-09-11 15:01:00 +00:00
Animesh Jain
5f630d28d7 [dynamo][guards] Do not construct entire framelocals dict for LAMBDA_GUARD (#162525) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/162525
Approved by: https://github.com/williamwen42
ghstack dependencies: #162509
 2025-09-10 18:52:15 +00:00
Animesh Jain
a67e798cb7 [dynamo][guards] Prevent framelocals to dict conversion for not required LAMBDA_GUARD (#162509) 
This is a smaller PR to reduce framelocals to dict conversion.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/162509
Approved by: https://github.com/williamwen42
 2025-09-10 18:52:15 +00:00
Animesh Jain
4d5b3f2d5a [dynamo][guards] Install dict watchers for recrusive dict tag optimization (#159796) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159796
Approved by: https://github.com/jansel
 2025-08-12 09:49:11 +00:00
Markus Hoehnerbach
e167c7d0f3 [inductor] allocate non-blocking copy destinations in pinned memory (#155121) (#158758) 
Fixes #155121

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158758
Approved by: https://github.com/EikanWang, https://github.com/eellison
 2025-08-07 17:07:26 +00:00
PyTorch MergeBot
83ba3f1101 Revert "[inductor] allocate non-blocking copy destinations in pinned memory (#155121) (#158758)" 
This reverts commit 6085bf7565.

Reverted https://github.com/pytorch/pytorch/pull/158758 on behalf of https://github.com/davidberard98 due to I need to revert #158462 (it causes device-side asserts), and this PR causes a merge conflict in the test file. Sorry about that! ([comment](https://github.com/pytorch/pytorch/pull/158758#issuecomment-3152490371))
 2025-08-04 21:47:11 +00:00
Markus Hoehnerbach
6085bf7565 [inductor] allocate non-blocking copy destinations in pinned memory (#155121) (#158758) 
Fixes #155121

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158758
Approved by: https://github.com/EikanWang, https://github.com/eellison
 2025-08-04 21:22:11 +00:00
Animesh Jain
53e47af0f7 [dynamo][guards] Read the attr name from GetAttrGuardAccessor (#159754) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159754
Approved by: https://github.com/jansel
ghstack dependencies: #159752
 2025-08-04 16:51:27 +00:00
Animesh Jain
66ad881fc7 [dynamo][guards][refactor] Simplify type extraction from GuardManager (#159752) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159752
Approved by: https://github.com/jansel
 2025-08-04 16:51:27 +00:00
Animesh Jain
64cbaa876c [dynamo][guards] Make class members go through obj.__class__.__dict__ (#159534) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159534
Approved by: https://github.com/jansel
 2025-08-04 05:12:44 +00:00
Animesh Jain
4516c59f5f [dynamo][source] Add special source for __code__ and __closure__ (#159722) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159722
Approved by: https://github.com/jansel
 2025-08-04 05:02:05 +00:00
PyTorch MergeBot
805a102beb Revert "[dynamo][guards] Make class members go through obj.__class__.__dict__ (#159534)" 
This reverts commit 1616777cd2.

Reverted https://github.com/pytorch/pytorch/pull/159534 on behalf of https://github.com/malfet due to Broke some inductor test and lint among other things, see 9c18901bfd/1 ([comment](https://github.com/pytorch/pytorch/pull/159534#issuecomment-3146983186))
 2025-08-03 04:58:32 +00:00
Animesh Jain
1616777cd2 [dynamo][guards] Make class members go through obj.__class__.__dict__ (#159534) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159534
Approved by: https://github.com/jansel
ghstack dependencies: #159186
 2025-08-02 18:04:35 +00:00
Animesh Jain
7eb5fdb358 [dynamo][guards] Recursive dict tag optimization (#159183) 
Design doc here - https://docs.google.com/document/d/1W29DrWID5miGWlZXspsQVN5U0zydE3kjZpziOXrhuaY/edit?tab=t.0#bookmark=id.sba04iw9sp68

Pull Request resolved: https://github.com/pytorch/pytorch/pull/159183
Approved by: https://github.com/jansel
 2025-07-30 06:01:32 +00:00
Animesh Jain
f7d6e9f500 [dynamo][guards] More small guard optimizations (#159345) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159345
Approved by: https://github.com/williamwen42
ghstack dependencies: #159288
 2025-07-29 18:36:49 +00:00
anwang
c55e72bea1 [Re-land][Inductor] Support native Inductor as backend for MTIA (#159211) 
The previous [diff/PR] (https://github.com/pytorch/pytorch/pull/158526) was reverted due to this docstring lint error:
<img width="1736" height="722" alt="image" src="https://github.com/user-attachments/assets/216b1720-4002-48da-b5f3-32b5d48aaa54" />
I didn't add the docstring cause I thought I'm not supposed to add docstring for an EXISTING function.

So this diff/PR is an exactly copy of the previous one, except for adding the docstring.

-------------
This diff/PR includes the changes to support native Inductor integration for MTIA. The goal is to support `torch.compile(backend="inductor")` for MTIA. Inductor should generate code(triton kernel + python wrapper code) similar to CUDA. And the triton kernels can be launched eagerly.

The changes include:
- Add MTIA device interfaces used by Dynamo and Inductor, including APIs on device, stream, event, etc.
- Add required torch.mtia APIs, like is_bf16_supported, memory_allocated, set_stream_by_id, etc.
- MTIA specific codegen logic, for example, loading MTIA dynamic_library.
- Other necessary changes to integrate with Inductor codegn, following other devices like CUDA, XPU.
- Integrate with the [empty_strided_mtia](https://www.internalfb.com/code/fbsource/[0d017d3a4a1bdff7253f9c66a9f38e77bd62166b]/fbcode/caffe2/aten/src/ATen/native/mtia/EmptyTensor.cpp?lines=49%2C63%2C71%2C74%2C78) API that we’ve added for the new MTIA ATen backend.
- A change in Inductor runtime to avoid re-initialize MTIADriver.
- BUCK changes to include ATen-mtia in Inductor, and to use -USE_MTIA preprocessor flag.
- Update `test_mnist_e2e.py` to cover native Inductor as backend, using the `--use_native_inductor` flag.
- Add a personal script(`scripts/anwang/run_native_inductor_script.py`) for testing purpose.

Note:
- This approach(option 3) aims to provide a pytorch native approach of Inductor integration for MTIA, minimizing the onboarding overhead. The downside of this approach is that it doesn't leverage MTIA specific graph optimization, and is limited to eagerly launch overhead.
- MTIA will support another approach(option 2) to provide best performance, based on WrapperFxCodegen. We should be able to reuse the fundamental changes of this diff for option 2, like the device interfaces, steam/event APIs, etc, especially as WrapperFxCodegen inherits PythonWrapperCodegen.

Internal:
References:
- [post for context](https://fb.workplace.com/groups/mtiasw/permalink/1718377262384606/)
- [Inductor integration discussion(option 1/2/3)](https://docs.google.com/document/d/1p6363OXtVIRv1hPoaKlRSK3j-iir3QIbDd5bjyqCNig/edit?tab=t.0#heading=h.7s4ns6wcnhmb)
- [Project design doc(option 3)](https://docs.google.com/document/d/1jXUmhgoV9WvkMf-bcY3Od_kK9K_RDOdgHdt1LoQ5Tc4/edit?tab=t.0#heading=h.y43gwdqlv46w)
- [early prototying diff](https://www.internalfb.com/diff/D75110196)
- [MPS integration PR](https://github.com/pytorch/pytorch/pull/153959)
- [empty_strided_xpu PR](https://github.com/pytorch/pytorch/pull/126678)

Differential Revision: [D79040806](https://our.internmc.facebook.com/intern/diff/D79040806/)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/159211
Approved by: https://github.com/eellison, https://github.com/blaine-rister, https://github.com/jansel
 2025-07-29 17:03:24 +00:00
PyTorch MergeBot
fe0ff12dab Revert "[Inductor] Support native Inductor as backend for MTIA (#158526)" 
This reverts commit cd68559d04.

Reverted https://github.com/pytorch/pytorch/pull/158526 on behalf of https://github.com/facebook-github-bot due to Diff reverted internally ([comment](https://github.com/pytorch/pytorch/pull/158526#issuecomment-3122186057))
 2025-07-26 17:58:00 +00:00
anwang
cd68559d04 [Inductor] Support native Inductor as backend for MTIA (#158526) 
This diff/PR includes the changes to support native Inductor integration for MTIA. The goal is to support `torch.compile(backend="inductor")` for MTIA. Inductor should generate code(triton kernel + python wrapper code) similar to CUDA. And the triton kernels can be launched eagerly.

The changes include:
- Add MTIA device interfaces used by Dynamo and Inductor, including APIs on device, stream, event, etc.
- Add required torch.mtia APIs, like is_bf16_supported, memory_allocated, set_stream_by_id, etc.
- MTIA specific codegen logic, for example, loading MTIA dynamic_library.
- Other necessary changes to integrate with Inductor codegn, following other devices like CUDA, XPU.
- Integrate with the [empty_strided_mtia](https://www.internalfb.com/code/fbsource/[0d017d3a4a1bdff7253f9c66a9f38e77bd62166b]/fbcode/caffe2/aten/src/ATen/native/mtia/EmptyTensor.cpp?lines=49%2C63%2C71%2C74%2C78) API that we’ve added for the new MTIA ATen backend.
- A change in Inductor runtime to avoid re-initialize MTIADriver.
- BUCK changes to include ATen-mtia in Inductor, and to use -USE_MTIA preprocessor flag.
- Update `test_mnist_e2e.py` to cover native Inductor as backend, using the `--use_native_inductor` flag.
- Add a personal script(`scripts/anwang/run_native_inductor_script.py`) for testing purpose.

Note:
- This approach(option 3) aims to provide a pytorch native approach of Inductor integration for MTIA, minimizing the onboarding overhead. The downside of this approach is that it doesn't leverage MTIA specific graph optimization, and is limited to eagerly launch overhead.
- MTIA will support another approach(option 2) to provide best performance, based on WrapperFxCodegen. We should be able to reuse the fundamental changes of this diff for option 2, like the device interfaces, steam/event APIs, etc, especially as WrapperFxCodegen inherits PythonWrapperCodegen.

Internal:
References:
- [post for context](https://fb.workplace.com/groups/mtiasw/permalink/1718377262384606/)
- [Inductor integration discussion(option 1/2/3)](https://docs.google.com/document/d/1p6363OXtVIRv1hPoaKlRSK3j-iir3QIbDd5bjyqCNig/edit?tab=t.0#heading=h.7s4ns6wcnhmb)
- [Project design doc(option 3)](https://docs.google.com/document/d/1jXUmhgoV9WvkMf-bcY3Od_kK9K_RDOdgHdt1LoQ5Tc4/edit?tab=t.0#heading=h.y43gwdqlv46w)
- [early prototying diff](https://www.internalfb.com/diff/D75110196)
- [MPS integration PR](https://github.com/pytorch/pytorch/pull/153959)
- [empty_strided_xpu PR](https://github.com/pytorch/pytorch/pull/126678)

Differential Revision: [D78458745](https://our.internmc.facebook.com/intern/diff/D78458745/)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158526
Approved by: https://github.com/blaine-rister, https://github.com/jansel, https://github.com/eellison
 2025-07-26 08:16:34 +00:00
Animesh Jain
659f8fb115 [dynamo][guards] Add some relational guard helpers (#159077) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/159077
Approved by: https://github.com/jansel
ghstack dependencies: #158995
 2025-07-25 06:28:10 +00:00
Animesh Jain
05a748d287 [dynamo][guards] Expand is_immutable_object to have None (#158995) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158995
Approved by: https://github.com/Lucaskabela, https://github.com/jansel
 2025-07-25 06:12:05 +00:00
Animesh Jain
1b456c580d [dynamo][guards] Add type info of the guarded value in guard managers (#158765) 
tlparse looks like this

<img width="1165" height="226" alt="image" src="https://github.com/user-attachments/assets/04c4e6b1-34a3-4d9d-8304-6eb6d9a94980" />

This will aid in reading guards.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158765
Approved by: https://github.com/Lucaskabela, https://github.com/StrongerXi
 2025-07-23 16:59:15 +00:00
cyy
1b91954b9f Suppress volatile type error (#158435) 
Fixes
```
/var/lib/jenkins/workspace/torch/csrc/dynamo/guards.cpp:5320:10:
error: compound assignment to object of volatile-qualified type 'volatile char' is deprecated [-Werror,-Wdeprecated-volatile]
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158435
Approved by: https://github.com/janeyx99
 2025-07-17 22:21:04 +00:00
Animesh Jain
2179afd714 [easy][guards] Add developer comment for posterity (#158471) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/158471
Approved by: https://github.com/StrongerXi
 2025-07-17 01:17:04 +00:00
Animesh Jain
cc0faeb80f [dynamo][guards] Instruction count for guard eval for development work (#158214) 
Its turned off  by default. Even the code is hidden before of the define preprocessing flag. It will be used only for development work.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158214
Approved by: https://github.com/StrongerXi
ghstack dependencies: #158215
 2025-07-15 20:29:23 +00:00
Guilherme Leobas
e7167dbacf [Set] Support sets in VariableBuilder (#153150) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/153150
Approved by: https://github.com/zou3519
 2025-07-04 00:45:03 +00:00
Animesh Jain
bb476310a4 [dynamo][guards] Stash root guard manager pointer in the LeafGuard (#157325) 
Preparing to simplify the recompilation reason codebase. This PR was 95% done by using AI tools.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/157325
Approved by: https://github.com/jansel
 2025-07-02 00:42:43 +00:00
zhxchen17
0f9c1b374f [dynamo] Ensure global state guard is preserved across serialization. (#157285) 
Currently, every time we construct a GLOBAL_STATE guard, we always create a fresh guard based on the current global state. For precompile, we want to create a GLOBAL_STATE guard always based on some external sources, e.g. serialized global states. This can also be applied with the normal case where we just pass in the global state guard from Python.

Differential Revision: [D77400988](https://our.internmc.facebook.com/intern/diff/D77400988/)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/157285
Approved by: https://github.com/jansel
 2025-07-01 15:46:34 +00:00
haozhe.zhu
53e0b9c393 refine fp32 precision api (#125888) 
Based on the [conversation](https://github.com/pytorch/pytorch/issues/121791), we plan to drop the "highest, high, medium" to represent fp32  internal computation data types . Instead, we will directly use the algorithm to represent it.

### Design Choice: Directly use algorithms name like "TF32", "BF16".
#### Pros
 - The names are more informative. 'tf32' is more informative than a simple "high".
 - Easier to extend new algorithm like `tf32x3`
#### Cons
 - "HIGHEST, HIGH, MEDIUM" indicated the relative precision between different algorithms. However, we can have more documents to discuss them.

### We provide a layered structure for backends/operators.
('f32' is short for 'fp32_precision')
![image](https://github.com/user-attachments/assets/f89143e5-d6a1-4865-9351-9a50439f5067)

### We provide 3 fp32 compute precision can be set:
 - **"ieee"**: Not allowed to use any other internal computation data types .
 - **"tf32"**: Allowed to use tf32 as internal computation data types.
 - **"bf16"**: Allowed to use bf16 as internal computation data types.
 - **"none"**:  Precision's are not set. Can be override by its father node.

### Overriding Precision Settings
Child node can be override by its father node if it is set to default.
For current default settings:
```
backend = generic, op = all, precision setting = none
    backend = cuda, op = all, precision setting = none
        backend = cuda, op = conv, precision setting = tf32
        backend = cuda, op = rnn, precision setting = tf32
        backend = cuda, op = matmul, precision setting = none
    backend = matmul, op = all, precision setting = none
        backend = matmul, op = conv, precision setting = none
        backend = matmul, op = rnn, precision setting = none
        backend = matmul, op = matmul, precision setting = none
```
 - If the user set `torch.backends.mkldnn.fp32_precision="bf16"`, his child nodes `torch.backends.mkldnn.matmul.fp32_precision` / `torch.backends.mkldnn.conv.fp32_precision` / `torch.backends.mkldnn.rnn.fp32_precision` will also be override to "bf16".
 - If the user set `torch.backends.fp32_precision="bf16"`,  `torch.backends.mkldnn.fp32_precision` and his child nodes will also we override to "bf16".

### Backward Compatible
Since new API allow user to have more fine-grained control. There will be some conflict. For example, previous `torch.backends.cudnn.allow_tf32` are not enough to represent the status for `torch.backends.cudnn.rnn.fp32_precision="ieee"` and `torch.backends.cudnn.conv.fp32_precision="tf32"`. Therefore, our goal for backward compatible is
 - If the user only uses previous APIs, it will work as previous expectations.
 - If the user use **new** API to change the status to an **un-representable** status for old API, and try to access the status by **old** API. We will raise Runtime Error and point the document for user.

### Test Plan
```
python test/test_cuda.py -k test_fp32_precision_with_tf32
python test/test_cuda.py -k test_fp32_precision_with_float32_matmul_precision
python test/test_cuda.py -k test_invalid_status_for_legacy_api
python test/test_mkldnn.py -k test_mlkdnn_get_set
python test/test_mkldnn.py -k test_generic_precision
python test/test_mkldnn.py -k test_invalid
python test/test_mkldnn.py -k test_default_use_parent
```

Pull Request resolved: https://github.com/pytorch/pytorch/pull/125888
Approved by: https://github.com/jgong5, https://github.com/albanD

Co-authored-by: Jiang, Yanbing <yanbing.jiang@intel.com>
 2025-06-26 10:32:20 +00:00
Yuanyuan Chen
07bb097698 Fix clang-tidy bugprone* warnings (#148529) 
Fixes #ISSUE_NUMBER

Pull Request resolved: https://github.com/pytorch/pytorch/pull/148529
Approved by: https://github.com/ezyang
 2025-06-23 23:09:56 +00:00
Xuehai Pan
5b210bb3a6 [BE][9/16] fix typos in torch/ (torch/csrc/) (#156319) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/156319
Approved by: https://github.com/albanD
ghstack dependencies: #156313, #156314, #156315, #156316, #156317
 2025-06-23 02:57:50 +00:00
PyTorch MergeBot
1d3bca40ed Revert "[BE][9/16] fix typos in torch/ (torch/csrc/) (#156319)" 
This reverts commit a23ccaa847.

Reverted https://github.com/pytorch/pytorch/pull/156319 on behalf of https://github.com/atalman due to export/test_torchbind.py::TestCompileTorchbind::test_compile_error_on_input_aliasing_contents_backend_aot_eager [GH job link](https://github.com/pytorch/pytorch/actions/runs/15804799771/job/44548489912) [HUD commit link](c95f7fa874) ([comment](https://github.com/pytorch/pytorch/pull/156313#issuecomment-2994171213))
 2025-06-22 12:31:56 +00:00
Xuehai Pan
a23ccaa847 [BE][9/16] fix typos in torch/ (torch/csrc/) (#156319) 
Pull Request resolved: https://github.com/pytorch/pytorch/pull/156319
Approved by: https://github.com/albanD
ghstack dependencies: #156313, #156314, #156315, #156316, #156317
 2025-06-22 08:43:49 +00:00
karthickai
10c3e6ec43 [inductor][dynamo] Include operator name in size/stride/alignment assertion (#152353) 
Fixes #151930

This PR updates the `assert_size_stride` and `assert_alignment` functions in [guards.cpp](https://github.com/pytorch/pytorch/blob/main/torch/csrc/dynamo/guards.cpp) to accept an optional `op_name` argument and includes it in the error messages.

The corresponding type stubs in [guards.pyi](https://github.com/pytorch/pytorch/blob/main/torch/_C/_dynamo/guards.pyi) are updated to match the new function arg.

In [inductor/ir.py](https://github.com/pytorch/pytorch/blob/main/torch/_inductor/ir.py) extracts the operator name from the FX graph and passes it into the `codegen_size_asserts` and `codegen_alignment_asserts` functions, so that generated assertions in Triton code include the op name for better debugging.

Added unit tests inside [test_torchinductor.py](https://github.com/pytorch/pytorch/blob/main/test/inductor/test_torchinductor.py).
- Verified both successful and failing assertion cases include the operator name.
- Verified that generated Triton code contains the op name inside the asserts.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/152353
Approved by: https://github.com/jansel, https://github.com/shunting314
 2025-06-03 19:21:15 +00:00
Animesh Jain
635b73e697 [dynamo][guards] Flush cache to more accurately measure guard overhead (#154764) 
We observed that guard overhead at runtime using profiler traces was
higher than reported in this profiling function at the compile time.
After investigation, we found that f_locals are already in cache and
that was causing the guard overhead to be way smaller while profiling
during the compilation. To be more realistic, we flush the cache here.

Profiling the guard overhead during compilation (in addition to at
runtime) allows faster iteration time, and logging in tlparse and
internal databases.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/154764
Approved by: https://github.com/zou3519, https://github.com/jansel, https://github.com/StrongerXi
 2025-06-03 11:50:57 +00:00
PyTorch MergeBot
b86aaaae0b Revert "[dynamo][guards] Flush cache to more accurately measure guard overhead (#154764)" 
This reverts commit 7dee899130.

Reverted https://github.com/pytorch/pytorch/pull/154764 on behalf of https://github.com/seemethere due to This fails internal tests see [fburl.com/diff/67gyp7gp](https://fburl.com/diff/67gyp7gp) ([comment](https://github.com/pytorch/pytorch/pull/154769#issuecomment-2933629894))
 2025-06-03 06:13:49 +00:00
Animesh Jain
7dee899130 [dynamo][guards] Flush cache to more accurately measure guard overhead (#154764) 
We observed that guard overhead at runtime using profiler traces was
higher than reported in this profiling function at the compile time.
After investigation, we found that f_locals are already in cache and
that was causing the guard overhead to be way smaller while profiling
during the compilation. To be more realistic, we flush the cache here.

Profiling the guard overhead during compilation (in addition to at
runtime) allows faster iteration time, and logging in tlparse and
internal databases.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/154764
Approved by: https://github.com/zou3519, https://github.com/jansel, https://github.com/StrongerXi
ghstack dependencies: #154769
 2025-06-02 23:01:58 +00:00