Mase/NewComputeBench/PLENA/../ repo Separation#292
Open
Aaron-Zhao123 wants to merge 1 commit intomainfrom
Open
Mase/NewComputeBench/PLENA/../ repo Separation#292Aaron-Zhao123 wants to merge 1 commit intomainfrom
Aaron-Zhao123 wants to merge 1 commit intomainfrom
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Repo Separation - Task List
Updated Goal (2 Feb 2026)
We will create five independent GitHub repositories, each responsible for a minimal, well-defined functional unit, with no redundant or overlapping functionality between them.
1. MASE: Provides core compiler functionality, including quantization, PLENA-based passes, and other necessary passes.
2. MASE Acceleration Gen (could be a different name): Imports
mase.passesand independently maintains its own pass set, effectivelymase.passes + own.passes.3. PLENA: Defines the single-core, full-stack NPU design flow. Imports
mase.passesis necessary.4. PLENA-Dist (could be a different name): Implements system-level multi-chip interconnection simulation based on PLENA.
5. NewComputeBench: Imports
mase.passesand updates the previous fine-tuning workflow to a pre-training workflow.MASE Acceleration Gen (new repo build @firemountain154B ):
import maseif dependentMASE (integrate quantization @JiaYi):
MASE (remove overlap @Binglei):
PLENA (@GeorgeWu1204 @firemountain154B ):
import masePLENA-Dist (@przemekff ):
NewComputeBench (@JeffreyWong20 @ChengZhang-98 @przemekff @firemountain154B @GeorgeWu1204 @Binglei ):
Follow the previous task goal below
Previous Goal (1 Nov 2025)
We would like to add all NewComputeBench related code change in a single PR.
Core Module Transformation Passes
For each module-level transformation pass on New Compute Paradigm, complete the following:
SNN pass (SpikeZIP-TF @JeffreyWong20 ):
docs/tutorials/newcompute)docs/source/modules/documentation)SNN pass (STA @GeorgeWu1204 ):
docs/tutorials/newcompute)docs/source/modules/documentation)ONN pass (STA @ChengZhang-98):
docs/tutorials/newcompute)docs/source/modules/documentation)PIM pass (PIM @firemountain154B ):
docs/tutorials/newcompute)docs/source/modules/documentation)Documentation & Testing Review (@Binglei @przemekff)
Optimizations (To Consider, @przemekff)