mirage-project/mirage

[Feature Request] - Optimizing Shared Memory Usage in norm_linear

Open

#451 opened on Aug 11, 2025

View on GitHub
 (1 comment) (0 reactions) (1 assignee)Cuda (232 forks)auto 404
good first issue

Repository metrics

Stars
 (2,381 stars)
PR merge metrics
 (PR metrics pending)

Description

Feature Description

Our norm_linear task cannot process large number of tokens. We should reduce shared memory usage in this task, which would allow us to run prefill using larger batch sizes.

Contributor guide

Research direction
Analyze the norm linear kernel to identify current shared memory usage per block. Look for opportunities to reduce shared memory such as using registers or global memory with coalesced access, or splitting kernel to use less shared memory per block. Check kernel launch parameters and occupancy.
Tech stack
ccpp
Domain
backend
Issue type
Performance
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
CUDA programmingShared memory optimization
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
40