Add exporting with argument device affinities #300
Merged
Conversation
This file contains 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
|
This requires iree-org/iree-turbine#231. |
sogartar
changed the title
rsuderman
approved these changes
Oct 22, 2024
rsuderman
requested changes
Oct 22, 2024
sharktank/sharktank/export.py
Outdated
| module_fn_with_flat_signature, | ||
| *transitive_args, | ||
| args=flat_args, | ||
| argument_device_affinities=argument_device_affinities, |
There was a problem hiding this comment.
Wait, where does this actually plumb through? I don't see this as part of the PR
There was a problem hiding this comment.
I see, the changes are dependent on https://github.com/iree-org/iree-turbine/pull/231/files
Lets focus on that and shelve these changes. Once export has the right parameter we won't need to make the sharktank changes.
sogartar
force-pushed
the
export-device-affinities
branch
3 times, most recently
from
c9404c0
to
0eb437f
Compare
rsuderman
approved these changes
Oct 23, 2024
sogartar
force-pushed
the
export-device-affinities
branch
from
0eb437f
to
9d35b3d
Compare
We don't have handling of device affinties for our sharded tensor types when they are the arguments of exported functions. Although, they can be specified explicitly when exporting, this change adds the ability to hide this and deduces the affinties from the tensor types.
sogartar
force-pushed
the
export-device-affinities
branch
from
9d35b3d
to
abbabed
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
We don't have handling of device affinties for our sharded tensor types when they are the arguments of exported functions. Although, they can be specified explicitly when exporting, this change adds the ability to hide this and deduces the affinties from the tensor types.