Introduce LegacyTHDispatcher for dispatching to TH functions. #14708
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
There was a problem hiding this comment.
@gchanan has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
ezyang
reviewed
Dec 3, 2018
|
|
||
| namespace at { | ||
|
|
||
| // TODO: This could be bad juju if someone calls globalContext() in the |
There was a problem hiding this comment.
is it? This is the same comment as in globalLegacyTypeDispatch -- is it out of date there? I'll admit I don't know what the comment is referring to.
ezyang
reviewed
Dec 3, 2018
|
|
||
| // LegacyTHDispatcher is the legacy mechanism for dispatching directly | ||
| // to TH/THNN/THC/THCUNN functions in ATen, which is essentially a giant virtual | ||
| // dispatch table for every TH function we support dynamically dispatching over. |
There was a problem hiding this comment.
I think what you want to add to the comment here is:
- Why you need another dispatcher, as opposed to using the existing LegacyTypeDispatch (you say that it's similar to the LegacyTypeDispatch with some simplifications, but that doesn't explain the "why")
- Under what situation we can delete this (when all the TH functions go away right)
ezyang
approved these changes
Dec 3, 2018
There was a problem hiding this comment.
I'm OK with this because we've discussed it in person, but I think you need to make sure @dzhulgakov and @smessmer understand what's going on here.
There was a problem hiding this comment.
@gchanan has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
gchanan
force-pushed
the
legacy_th_dispatcher
branch
from
58be3c0 to
f0dcc50
Compare
gchanan
force-pushed
the
legacy_th_dispatcher
branch
from
f0dcc50 to
a7e6f4d
Compare
…h#14708) Summary: This isn't hooked up to anything yet, this is just putting the skeleton in place. The idea here is that the functions generated via Declarations.cwrap and nn.yaml are not actually operators, they are implementation details of operators, and thus don't need to participate in VariableType, JIT dispatch generation. So, we will split these functions out from the usual Type/operator hierarchy; for now the dispatch will be done by a Type-like class called LegacyTHDispatcher. Once this is done this probably means we can collapse Type to be backend-specific, not Type/ScalarType specific, because all the ScalarType specific code will live in the LegacyTHDispatcher. Pull Request resolved: pytorch#14708 Reviewed By: ezyang Differential Revision: D13304654 Pulled By: ezyang fbshipit-source-id: dac7c7fba8b58d52836f48466e986487c58e78b0
gchanan
force-pushed
the
legacy_th_dispatcher
branch
from
a7e6f4d to
65defe1
Compare
zdevito
pushed a commit
to zdevito/ATen
that referenced
this pull request
Dec 4, 2018
Summary: This isn't hooked up to anything yet, this is just putting the skeleton in place. The idea here is that the functions generated via Declarations.cwrap and nn.yaml are not actually operators, they are implementation details of operators, and thus don't need to participate in VariableType, JIT dispatch generation. So, we will split these functions out from the usual Type/operator hierarchy; for now the dispatch will be done by a Type-like class called LegacyTHDispatcher. Once this is done this probably means we can collapse Type to be backend-specific, not Type/ScalarType specific, because all the ScalarType specific code will live in the LegacyTHDispatcher. Pull Request resolved: pytorch/pytorch#14708 Reviewed By: ezyang Differential Revision: D13304654 Pulled By: gchanan fbshipit-source-id: cfe3e1a28adcc355f67fe143495ee7e5c5118606
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.
This isn't hooked up to anything yet, this is just putting the skeleton in place.
The idea here is that the functions generated via Declarations.cwrap and nn.yaml are not actually operators, they are implementation details of operators, and thus don't need to participate in VariableType, JIT dispatch generation.
So, we will split these functions out from the usual Type/operator hierarchy; for now the dispatch will be done by a Type-like class called LegacyTHDispatcher. Once this is done this probably means we can collapse Type to be backend-specific, not Type/ScalarType specific, because all the ScalarType specific code will live in the LegacyTHDispatcher.