Stories by matt_d

TLX: Triton-Like Simplicity, a Clear Path to Peak Performance [video]

Automating Inference Optimizations with NVIDIA TensorRT LLM AutoDeploy

SMTLIB as a Compiler IR I

Equality Saturation Meets ML: The Next Step for Smarter Optimizing Compilers [video]

Silicon Photonics in the Data Center: What a CMOS Exec Needs to Know

Verifying Distributed Protocols in Veil

Project Pterodactyl: Incremental Architecture

European Lisp Symposium 2025: Talks

AutoOverlap: Enabling Fine-Grained Overlap of Computation and Communication

Axe: A Simple Unified Layout Abstraction for Machine Learning Compilers

35th ACM SIGPLAN International Conference on Compiler Construction (CC 2026)

VFlatten: Selective Value-Object Flattening Using Hybrid Static&Dynamic Analysis [pdf]

Agentic Proof-Oriented Programming

MLIR-Tutor: Exercises for Learning MLIR (Originally Written for PPoPP 2026)

Fast Autoscheduling for Sparse ML Frameworks

Replicate Forwards, Partial Backwards

Frontier-CS 1.0 Release

uops-again.info: corner-case behaviours of port assignment on Intel processors

When magic meets multicore: OCaml and its elegant era of parallelism [video]

FlashAttention-T: Towards Tensorized Attention

1
2
3
4
5
6
7
8
9
10