Make Channel API accept buffer structs rather than raw pointers. #212
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
+ Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. [ghstack-poisoned]
This was referenced Sep 9, 2020
Closed
|
Note: Preliminary discussion for this happened here. |
beauby
added a commit
to beauby/tensorpipe
that referenced
this pull request
Sep 9, 2020
+ Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. ghstack-source-id: 132990d Pull Request resolved: pytorch#212
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. [ghstack-poisoned]
lw
reviewed
Sep 10, 2020
There was a problem hiding this comment.
There's a couple of naming discussions I'd like to get resolved. And, as you know, I'd like to see a WIP PR up against PyTorch to see how this new API would look like when used from there.
Also, I found several instances where you used Channel<CpuTensor> instead of CpuChannel: did you not find&replace to substitute them? Could you do so now to make sure I didn't miss any?
|
PR against pytorch: pytorch/pytorch#41475 |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
lw
reviewed
Sep 11, 2020
…nters." + Introduce tensor.h defining the tensor struct(s). The `CpuTensor` struct is always defined, while the `CudaTensor` struct is defined only when `TENSORPIPE_HAS_CUDA` is true. + Update all channels to take a `CpuTensor` or `CudaTensor` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TTensor`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
…nters." + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: [D23598033](https://our.internmc.facebook.com/intern/diff/D23598033) [ghstack-poisoned]
|
@beauby has updated the pull request. You must reimport the pull request before landing. |
lw
approved these changes
Sep 15, 2020
beauby
added a commit
to beauby/tensorpipe
that referenced
this pull request
Sep 17, 2020
…orch#212) Summary: Pull Request resolved: pytorch#212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: 1c1de3cfb8c492e3e1497db17d7d29fbb1db9161
This was referenced Sep 17, 2020
beauby
added a commit
to beauby/tensorpipe
that referenced
this pull request
Sep 19, 2020
…014) Summary: Pull Request resolved: pytorch/pytorch#45014 Pull Request resolved: pytorch#219 Pull Request resolved: pytorch#212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: a18ca76e1453f2b5d091a2d0103362818eafa028
beauby
added a commit
to beauby/tensorpipe
that referenced
this pull request
Sep 21, 2020
…014) Summary: Pull Request resolved: pytorch/pytorch#45014 Pull Request resolved: pytorch#219 Pull Request resolved: pytorch#212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: 8bd83a8a7e7c2e7d96b4c6025f1238546cdd6222
beauby
added a commit
to beauby/tensorpipe
that referenced
this pull request
Sep 21, 2020
…014) Summary: Pull Request resolved: pytorch/pytorch#45014 Pull Request resolved: pytorch#219 Pull Request resolved: pytorch#212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: 6e66ec1184cd568e99f6d0096d3f88e18e9e5a91
beauby
added a commit
that referenced
this pull request
Sep 21, 2020
…014) Summary: Pull Request resolved: pytorch/pytorch#45014 Pull Request resolved: #219 Pull Request resolved: #212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: a18ca76e1453f2b5d091a2d0103362818eafa028
beauby
added a commit
that referenced
this pull request
Sep 21, 2020
…014) Summary: Pull Request resolved: pytorch/pytorch#45014 Pull Request resolved: #219 Pull Request resolved: #212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: a18ca76e1453f2b5d091a2d0103362818eafa028
beauby
added a commit
to beauby/pytorch
that referenced
this pull request
Sep 21, 2020
…orch#45014) Summary: Pull Request resolved: pytorch#45014 Pull Request resolved: pytorch/tensorpipe#219 Pull Request resolved: pytorch/tensorpipe#212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: 7bd0ddbb23b063bc0c436f09c1fac2a875a4f62e
facebook-github-bot
pushed a commit
to pytorch/pytorch
that referenced
this pull request
Sep 21, 2020
) Summary: Pull Request resolved: #45014 Pull Request resolved: pytorch/tensorpipe#219 Pull Request resolved: pytorch/tensorpipe#212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: 1d6c3f91e288420858835cd5e7962e8da051b44b
lw
pushed a commit
that referenced
this pull request
Sep 20, 2021
…014) Summary: Pull Request resolved: pytorch/pytorch#45014 Pull Request resolved: #219 Pull Request resolved: #212 + Introduce buffer.h defining the buffer struct(s). The `CpuBuffer` struct is always defined, while the `CudaBuffer` struct is defined only when `TENSORPIPE_SUPPORTS_CUDA` is true. + Update all channels to take a `CpuBuffer` or `CudaBuffer` for `send`/`recv` rather than a raw pointer and a length. + Make the base `Channel`/`Context` classes templated on `TBuffer`, effectively creating two channel hierarchies (one for CPU channels, one for CUDA channels). + Update the Pipe and the generic channel tests to use the new API. So far, generic channel tests are CPU only, and tests for the CUDA IPC channel are (temporarily) disabled. A subsequent PR will take care of refactoring tests so that generic tests work for CUDA channels. An other PR will add support for CUDA tensors in the Pipe. Differential Revision: D23598033 Test Plan: Imported from OSS Reviewed By: lw Pulled By: beauby fbshipit-source-id: 1d6c3f91e288420858835cd5e7962e8da051b44b
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Stack from ghstack:
channelName()in channel test helper. #213 Get rid ofchannelName()in channel test helper.channel::Channel::tochannel::. #210 Move channel types fromchannel::Channel::tochannel::.CpuBufferstruct is always defined, while the
CudaBufferstruct is definedonly when
TENSORPIPE_SUPPORTS_CUDAis true.CpuBufferorCudaBufferforsend/recvrather than a raw pointer and a length.Channel/Contextclasses templated onTBuffer,effectively creating two channel hierarchies (one for CPU channels,
one for CUDA channels).
far, generic channel tests are CPU only, and tests for the CUDA IPC
channel are (temporarily) disabled. A subsequent PR will take care of
refactoring tests so that generic tests work for CUDA channels. An
other PR will add support for CUDA tensors in the Pipe.
Differential Revision: D23598033