Thomas Viehmann
d3d8da7a8e
Enable CUDA Fuser for ROCm ( #45965 )
...
Summary:
This enables the cuda fuser on ROCm and enables tests for them.
Part of this patch is based on work of Rohith Nallamaddi, thank you.
Errors are my own, of course.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45965
Reviewed By: seemethere
Differential Revision: D24170457
Pulled By: walterddr
fbshipit-source-id: 3dd25b3501a41d2f00acba3ce8642ce51c49c9a6
2020-10-08 10:41:56 -07:00
Rong Rong
986af53be2
type check for torch.testing._internalcodegen:* ( #45368 )
...
Summary:
part of `torch.testing._internal.*` effort
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45368
Reviewed By: malfet
Differential Revision: D23950512
Pulled By: walterddr
fbshipit-source-id: 399f712d12cdd9795b0136328f512c3f86a15f24
2020-09-28 14:04:52 -07:00
Akihiro Nitta
f17d7a5556
Fix exception chaining in torch/ ( #43836 )
...
Summary:
## Motivation
Fixes https://github.com/pytorch/pytorch/issues/43770 .
## Description of the change
This PR fixes exception chaining only in files under `torch/` where appropriate.
To fix exception chaining, I used either:
1. `raise new_exception from old_exception` where `new_exception` itself seems not descriptive enough to debug or `old_exception` delivers valuable information.
2. `raise new_exception from None` where raising both of `new_exception` and `old_exception` seems a bit noisy and redundant.
I subjectively chose which one to use from the above options.
## List of lines containing raise in except clause:
I wrote [this simple script](https://gist.github.com/akihironitta/4223c1b32404b36c1b349d70c4c93b4d ) using [ast](https://docs.python.org/3.8/library/ast.html#module-ast ) to list lines where `raise`ing in `except` clause.
- [x] 000739c31a/torch/jit/annotations.py (L35)000739c31a/torch/jit/annotations.py (L150)000739c31a/torch/jit/annotations.py (L158)000739c31a/torch/jit/annotations.py (L231)000739c31a/torch/jit/_trace.py (L432)000739c31a/torch/nn/utils/prune.py (L192)000739c31a/torch/cuda/nvtx.py (L7)000739c31a/torch/utils/cpp_extension.py (L1537)000739c31a/torch/utils/tensorboard/_pytorch_graph.py (L292)000739c31a/torch/utils/data/dataloader.py (L835)000739c31a/torch/utils/data/dataloader.py (L849)000739c31a/torch/utils/data/dataloader.py (L856)000739c31a/torch/testing/_internal/common_utils.py (L186)000739c31a/torch/testing/_internal/common_utils.py (L189)000739c31a/torch/testing/_internal/common_utils.py (L424)000739c31a/torch/testing/_internal/common_utils.py (L1279)000739c31a/torch/testing/_internal/common_utils.py (L1283)000739c31a/torch/testing/_internal/common_utils.py (L1356)000739c31a/torch/testing/_internal/common_utils.py (L1388)000739c31a/torch/testing/_internal/common_utils.py (L1391)000739c31a/torch/testing/_internal/common_utils.py (L1412)000739c31a/torch/testing/_internal/codegen/random_topo_test.py (L310)000739c31a/torch/testing/_internal/codegen/random_topo_test.py (L329)000739c31a/torch/testing/_internal/codegen/random_topo_test.py (L332)000739c31a/torch/testing/_internal/jit_utils.py (L183)000739c31a/torch/testing/_internal/common_nn.py (L4789)000739c31a/torch/onnx/utils.py (L367)000739c31a/torch/onnx/utils.py (L659)000739c31a/torch/onnx/utils.py (L892)000739c31a/torch/onnx/utils.py (L897)000739c31a/torch/serialization.py (L108)000739c31a/torch/serialization.py (L754)000739c31a/torch/distributed/rpc/_testing/faulty_agent_backend_registry.py (L76)000739c31a/torch/distributed/rpc/backend_registry.py (L260)000739c31a/torch/distributed/distributed_c10d.py (L184)000739c31a/torch/_utils_internal.py (L57)000739c31a/torch/hub.py (L494)000739c31a/torch/contrib/_tensorboard_vis.py (L16)000739c31a/torch/distributions/lowrank_multivariate_normal.py (L100)000739c31a/torch/distributions/constraint_registry.py (L142)https://github.com/pytorch/pytorch/pull/43836
Reviewed By: ailzhang
Differential Revision: D23431212
Pulled By: malfet
fbshipit-source-id: 5f7f41b391164a5ad0efc06e55cd58c23408a921
2020-08-31 20:26:23 -07:00
Stanislau Hlebik
b774ce54f8
remediation of S205607
...
fbshipit-source-id: 798decc90db4f13770e97cdce3c0df7d5421b2a3
2020-07-17 17:19:47 -07:00
Stanislau Hlebik
8fdea489af
remediation of S205607
...
fbshipit-source-id: 5113fe0c527595e4227ff827253b7414abbdf7ac
2020-07-17 17:17:03 -07:00
Jie
07518e120b
[nvFuser] add torch.jit.fuser context manager ( #38993 )
...
Summary:
1. `torch.jit.fuser(str)` context manager facilitates switch between backend fusers:
str - 'fuser0' enables only legacy fuser;
str - 'fuser1' enables only NNC;
str - 'fuser2' enables only nvFuser;
2. cleanup updated python tests.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38993
Reviewed By: nairbv, pbelevich
Differential Revision: D21800620
Pulled By: soumith
fbshipit-source-id: 7fe855f5a5b97368e5e84c98c28d04b2e1276c85
2020-06-01 10:52:40 -07:00
Christian Sarofeen
8e69c3be17
[nvFuser] Reduction support in codegen, fp16 support ( #38627 )
...
Summary:
Adds reduction support for the code generator. Reductions are fully supported with split/merge/reorder/rfactor/computeAt/unroll operators. There is also cross thread (intra-block) reduction support.
The two remaining pieces missing for reduction support is:
- Safety: If cross thread reduction was used, child operators shouldn't be able to bind that thread dim anymore
- Cross block reduction: we will want inter-block reduction support to match parity with tensor iterator
PR also provides FP16 support for fusions now. We insert casts on FP16 inputs to FP32, and we insert casts to FP16 on FP16 outputs.
Also working towards reductions and shape inference for reductions in the fusion pass.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38627
Reviewed By: albanD
Differential Revision: D21663196
Pulled By: soumith
fbshipit-source-id: 3ff2df563f86c39cd5821ab9c1148149e5172a9e
2020-05-21 17:18:39 -07:00
