[Relax] Memory planning for "partially dynamic" shapes #16466
Merged
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
This PR improves the memory planning to plan for tensors with shape such as `(m, n)` where only `m` has an integer upper bound (say `m <= 20`). Prior to this PR, the tensor is not planned even though `m` has an upper bound. This is not as ideal since it does not fully leverage the upper bound info. This PR enhances the planning, so that a memory chunk of `20 * n` will be allocated first, and the tensor will then be allocated out from this chunk. This enables us to fully leverage the upper bound info. A unit test is provided.
Member
|
This idea is pretty much similar to the "unified memory pool" in S-LoRA. It's not as feature-complete as S-LoRA, so please take a look whenever you have time |
junrushao
approved these changes
Jan 25, 2024
masahi
pushed a commit
to masahi/tvm
that referenced
this pull request
Feb 20, 2024
This PR improves the memory planning to plan for tensors with shape such as `(m, n)` where only `m` has an integer upper bound (say `m <= 20`). Prior to this PR, the tensor is not planned even though `m` has an upper bound. This is not as ideal since it does not fully leverage the upper bound info. This PR enhances the planning, so that a memory chunk of `20 * n` will be allocated first, and the tensor will then be allocated out from this chunk. This enables us to fully leverage the upper bound info. A unit test is provided.
elvin-n
pushed a commit
to Deelvin/tvm
that referenced
this pull request
Mar 19, 2024
This PR improves the memory planning to plan for tensors with shape such as `(m, n)` where only `m` has an integer upper bound (say `m <= 20`). Prior to this PR, the tensor is not planned even though `m` has an upper bound. This is not as ideal since it does not fully leverage the upper bound info. This PR enhances the planning, so that a memory chunk of `20 * n` will be allocated first, and the tensor will then be allocated out from this chunk. This enables us to fully leverage the upper bound info. A unit test is provided.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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 PR improves the memory planning to plan for tensors with shape such as
(m, n)where onlymhas an integer upper bound (saym <= 20).Prior to this PR, the tensor is not planned even though
mhas an upper bound. This is not as ideal since it does not fully leverage the upper bound info.This PR enhances the planning, so that a memory chunk of
20 * nwill be allocated first, and the tensor will then be allocated out from this chunk. This enables us to fully leverage the upper bound info.A unit test is provided.