frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-21 05:34:18 +08:00
Code Issues Packages Projects Releases Wiki Activity
94,736 Commits 5,442 Branches 1,295 Tags
main
Commit Graph

4 Commits

Author SHA1 Message Date
Sherlock Huang
fb067de550 [NativeRT] Remove device_ member from OpKernel base class (#158944) 
Summary:
In general, device_ is not very useful in OpKernel.  Remove it to avoid misuse.

Also, the meaning of `device_` is also ambiguous in the OpKernel.
For StaticDispatch kernels, we always call cpu kernel.
For C10Kernel, we rely on input tensor's device and dispatcher to determine which device to run on.
For ops involves multiple device, e.g. aten._to_copy(device), the meaning of device is ill-defined.

Test Plan:
CI

Rollback Plan:

Reviewed By: henryoier, dolpm, kqfu, zhxchen17

Differential Revision: D78704840

Pull Request resolved: https://github.com/pytorch/pytorch/pull/158944
Approved by: https://github.com/dolpm
 2025-07-24 09:21:37 +00:00
Dylan Maloy
81759afed4 [nativert] clean up some migration side-effects (#156919) 
Summary: explicit torch::nativert namespace usage + // manual declarations

Test Plan:
ci

Rollback Plan:

Differential Revision: D77328855

Pull Request resolved: https://github.com/pytorch/pytorch/pull/156919
Approved by: https://github.com/zhxchen17
 2025-06-26 20:28:32 +00:00
Georgia Phillips
24063ad109 Fix native static dispatch kernels (#156331) 
Summary: Fix for native static dispatch kernels not taking effect

Test Plan:
```
buck2 test //sigmoid/backend/test:static_kernels_ops_test

buck2 run mode/opt caffe2/torch/fb/model_transform/fx2trt/packaging:load_net_predictor -- --loadMode=BenchmarkByOp --inputNetFile=/data/users/$USER/models/${MODEL_ENTITY_ID}/${SNAPSHOT_ID}/${MODEL_ENTITY_ID}_${SNAPSHOT_ID}${SUFFIX} --moduleName=${MODULE} --submodToDevice "" --pytorch_predictor_sigmoid_static_dispatch_enable=true --pytorch_predictor_sigmoid_graph_passes_enable=true --benchmarkEnableProfiling=true --load_lowered_merge=3 --using_aoti_lowering_allowlist=false --requestFilePath=/data/users/georgiaphillips/replayer/inputs/742055223/0/mix/742055223_0_mix.inputs.recordio --benchmarkNumIterations=2
```

Rollback Plan:

Reviewed By: dolpm

Differential Revision: D76559764

Pull Request resolved: https://github.com/pytorch/pytorch/pull/156331
Approved by: https://github.com/Skylion007, https://github.com/jingsh
 2025-06-24 00:05:49 +00:00
Yiming Zhou
f2d70898c6 [nativert] Move OpKernel to PyTorch core (#156011) 
Summary:
Moves OpKernel base class to PyTorch core. It is an abstract interface representing a kernel, which is responsible for executing a single Node in the graph.

Torch Native Runtime RFC: pytorch/rfcs#72

Test Plan:
buck2 run mode/dev-nosan caffe2/test/cpp/nativert:op_kernel_test

Rollback Plan:

Differential Revision: D76525939

Pull Request resolved: https://github.com/pytorch/pytorch/pull/156011
Approved by: https://github.com/zhxchen17
 2025-06-16 22:53:10 +00:00