[python][utils] MemRef Manager #20
Open
+218
−0
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
Contributor
Author
|
The example could be refactored into a proper test once we settle down on a testing strategy. |
adam-smnk
force-pushed
the
mlir-memref-manager
branch
from
994a567 to
127b122
Compare
rolfmorel
reviewed
Nov 19, 2025
Contributor
rolfmorel
left a comment
rolfmorel
left a comment
There was a problem hiding this comment.
Looks interesting! Here's a first pass.
adam-smnk
force-pushed
the
mlir-memref-manager
branch
from
127b122 to
2e0acb9
Compare
adam-smnk
force-pushed
the
mlir-memref-manager
branch
from
2e0acb9 to
08df086
Compare
Adds a utility for manual memory management of memref buffers across
Python and jitted MLIR modules.
Explicit memory management becomes required when an MLIR function
returns a newly allocated buffer e.g., results of a computation.
This can become a complex task due to difference in memory models
between Python and the MLIR runtime allocators.
By default, returned MLIR buffers' lifetime cannot be automatically
managed by the Python environment.
The Python memref manager aims to address the following challenges:
- use of the same runtime allocators as a jitted MLIR module for
consistent memory management
- lean abstraction using memref descriptors directly
- buffers usable both by Python and jitted MLIR modules
Current implementation assumes that memref allocation ops are lowered
to standard C functions, like 'malloc' and 'free', which are preloaded
together with the Python process.
adam-smnk
force-pushed
the
mlir-memref-manager
branch
from
08df086 to
62cce0d
Compare
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.
Adds a utility for manual memory management of memref buffers across Python and jitted MLIR modules.
Explicit memory management becomes required when an MLIR function returns a newly allocated buffer e.g., results of a computation. This can become a complex task due to difference in memory models between Python and the MLIR runtime allocators.
By default, returned MLIR buffers' lifetime cannot be automatically managed by the Python environment.
The Python memref manager aims to address the following challenges:
Current implementation assumes that memref allocation ops are lowered to standard C functions, like 'malloc' and 'free', which are preloaded together with the Python process.