Repository navigation
Conversation
Signed-off-by: Robert Washbourne <rawashbourne@gmail.com>
Signed-off-by: Robert Washbourne <rawashbourne@gmail.com>
Signed-off-by: Robert Washbourne <rawashbourne@gmail.com>
rawsh
force-pushed
the
feature/adv-metrics-packing
branch
from
October 21, 2026 09:58
07f7a0b to
51b3e05
Compare
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Coalesces small tensor frames through the existing packed codec, with the existing row decoder and original payload bytes preserved.
136 allocated byte/socket controls pass; matched 48,384-row 4.262 GB field PUT+GET measured 7.739 s before and 8.033 s after, with identical input hashes, CPU affinity and threads. No combined gain; experiment is closed and excluded from adoption.
Full advantage/metrics target remains open; this single-field measurement establishes no full-path or production speedup. AI-assisted implementation; hosted release-branch baseline type/checkpoint failures remain explicit.