Stories by matt_d

Megatron via shard_map

Cloud-Hardware Co-Design for Memory Bandwidth-Bound HPC Workloads: Azure HBv5

CuTile on Blackwell: NVIDIA's Compiler Moat Is Already Built

Compiling Classical Sequent Calculus to Stock Hardware: Duality of Compilation [video]

Computing Sharding with Einsum

What Is Control Flow Analysis for Lambda Calculus? [audio]

Introduction to Coinduction in Agda Part 1: Coinductive Programming

Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in CLIs

Multi-Modal Program Verification in Velvet

Python, Is It Being Killed by Incremental Improvements?

Terabyte-Scale Analytics in the Blink of an Eye

What Is Control Flow Analysis for Lambda Calculus? – Iowa Type Theory Commute

Benchmarking a Baseline Fully-in-Place Functional Language Compiler [pdf]

Trends in Functional Programming (TFP) 2026

Categorical Foundations for CuTe Layouts

StackWarp: Exploiting Stack Layout Vulnerabilities in Modern Processors

Cloud RAM

Triton Linear Layout: Examples

When XLA Isn't Enough: From Pallas to VLIW with Splash Attention on TPU

Warp Specialization in Triton: Design and Roadmap

1
2
3
4
5
6
7
8
9
10