[tmva][sofie] Fix the allocation of intermediate tensors in the memory pool #19730
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
sanjibansg
requested changes
Aug 22, 2025
Test Results 19 files 19 suites 3d 11h 32m 14s ⏱️ For more details on these failures, see this check. Results for commit 5e1d1e2. ♻️ This comment has been updated with latest results. |
lmoneta
force-pushed
the
tmva_sofie_fix_memory_pool
branch
from
59f9ea2 to
096e3ad
Compare
When broadcasting from scalar tensor the tensor size is 0 and shape.front() is undefined. Add then the check on size before calling shape.front()
Add counters in RModel to monitor allocations of Constant, Weights, initermidiates and other types of tensor sizes allocated at code generation
By adding the conv temporary tensors in input lists they will be flushed afterwards and their memory can be reused by next operator
sanjibansg
approved these changes
Aug 25, 2025
Fix an issue of merging free chunks in the list of available_stack memory This will make easier to re-use more efficiently the memory In addition order the optput tensor by decreasing sizes Add also debug of the current chunk allocated and avaialable during the process
lmoneta
force-pushed
the
tmva_sofie_fix_memory_pool
branch
from
096e3ad to
5e1d1e2
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Outdated suggestions cannot be applied.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.
This PR fixes an issue in merging the free chunks of memory which are used in the memory pool for the
allocation of the intermediate tensors.
This PR provides a significant (x2) improvement in total memory usage for the intermediate tensors