ML Systems

Feeds to Scour
SubscribedAll
Scoured 31 posts in 16.3 ms

APEX4: Efficient Pure W4A4 LLM Inference via Intra-SM Compute Rebalancing

 💬LLMs  Content type: Academic
arxiv.org·

If Claude Fable stops helping you, you’ll never know

 💬LLMs

Gerrymandering the Warp: Non-Control-Data Attacks on CUDA Collective Decision

 📐Scaling Laws  Content type: Academic
arxiv.org·

LLM-Based Porting of Optimized C++ to CUDA Through Deoptimization and Reoptimization

 📐Scaling Laws  Content type: Academic
arxiv.org·

WarpGuard: Protected-Site Control-Flow Integrity for CUDA SASS Binaries

 🤖AI Agents  Content type: Academic
arxiv.org·

SET: Stream-Event-Triggered Scheduling for Efficient CUDA Graph Pipelines

 🔥PyTorch  Content type: Academic
arxiv.org·

TileFuse: A Fused Mixed-Precision Kernel Library for Efficient Quantized LLM Inference on AMD NPUs

 📉Deep Learning  Content type: Academic
arxiv.org·

AutoMegaKernel: A Statically-Checked Agent Harness for Self-Retargeting Megakernel Synthesis

 💬LLMs  Content type: Academic
arxiv.org··Hacker News

Joint Structural Pruning and Mixed-Precision Quantization for LLM Compression

 💬LLMs  Content type: Academic
arxiv.org·

Learned Subspace Compression for Communication-Efficient Pipeline Parallelism

 🧠AI Research  Content type: Academic
arxiv.org·

Characterizing Software Aging in GPU-Based LLM Serving Systems

 💬LLMs  Content type: Academic
arxiv.org·

Enhancing AI Interpretability and Safety through Localised Architectures

 🔍Interpretability  Content type: Academic
arxiv.org·

INFRAMIND: Infrastructure-Aware Multi-Agent Orchestration

 🎮Reinforcement Learning  Content type: Academic
arxiv.org·

Resource-aware Computation-Communication Overlap for multi-GPU ML Workloads

 📉Deep Learning  Content type: Academic
arxiv.org·

Demystifying NVSHMEM: A System-Level Analysis on Symmetric Memory and Device-Initiated Operations in GPU Communication

 📐Scaling Laws  Content type: Academic
arxiv.org·

Bergson: An Open Source Library for Data Attribution

 ⚙️Model Training  Content type: Academic
arxiv.org·

Breaking the Bubble: Asynchronous Pipeline Parallel Training with Bounded Weight Inconsistency

 ⚙️Model Training  Content type: Academic
arxiv.org·

A Scalable PyTorch Abstraction for Multi-GPU Gaussian Splatting

 🔥PyTorch  Content type: Academic
arxiv.org·

On GPU Implementation for Multi-Precision Integer Division

 📐Scaling Laws  Content type: Academic
arxiv.org·

PALUTE: Processing-In-Memory Acceleration via Lookup Table for Edge LLM Inference

 💬LLMs  Content type: Academic
arxiv.org·

Keyboard Shortcuts

Navigation

Next / previous item
j/k
Open post
oorEnter
Preview post
v

Post Actions

Love post
a
Like post
l
Dislike post
d
Undo reaction
u
Save / unsave
s

Recommendations

Add interest / feed
Enter
Not interested
x

Go to

Home
gh
Interests
gi
Feeds
gf
Likes
gl
History
gy
Changelog
gc
Settings
gs
Browse
gb
Search
/

General

Show this help
?
Submit feedback
!
Close modal / unfocus
Esc

Press ? anytime to show this help