Stories by matt_d

Scaling GPU-to-CPU Migration for Efficient Distributed Execution on CPU Clusters

MetaAttention: A Unified&Performant Attention Framework across Hardware Backends

Ga68, a GNU Algol 68 Compiler

LLMs versus the Halting Problem: Revisiting Program Termination Prediction

DTensor Erasure

Let the Barbarians In: How AI Can Accelerate Systems Performance Research

Triton Bespoke Layouts

AMD64 Bit Matrix Multiply and Bit Reversal Instructions

Demystifying ARM SME to Optimize General Matrix Multiplications

Evolving the OCaml programming language – CSE Bytes: K C Sivaramakrishnan [video]

Magellan: Autonomous Discovery of Compiler Optimization Heuristics w/AlphaEvolve

Automatic Data Enumeration for Fast Collections

An MLIR Lowering Pipeline for Stencils at Wafer-Scale

The JAX sharding type system

AutoSP: Unlocking Long-Context LLM Training via Compiler-Based SP (ICLR 2026)

Disentangling unification and implicit coercion (subtyping interaction problem)

Global vs. Local SPMD

ACM SIGPLAN Symposium on Principles of Programming Languages (POPL) 2026 talks

Long branches in compilers, assemblers, and linkers

1
2
3
4
5
6
7
8
9
10