Stories by matt_d

NanoTag: Systems Support for Efficient Byte-Granular Overflow Detection on Arm

InferenceBench: A Benchmark for Open-Ended Inference Optimization by AI Agents

Tracking Capabilities for Safer Agents

Scalable Packed Layouts for Vector-Length-Agnostic ML Code Generation

Verifying EDA and compiler optimizations once and for all

StepStone: LLM-Based GPU Kernel Driver Fuzzing via User-Space Libraries [pdf]

Graded Modal Types for Memory and Communication Safety

Systems Are Changing: The Architect's Role in the Era of Agentic Co-Design

Code-Specify-Test-Debug-Prove: Flexibly Integrating Separation Logic [pdf]

Detecting Relaxed Memory Concurrency Bugs in C and C++ Compilers

The downgrading semantics of memory safety (Extended version)

Direction-Preserving Number Representations

On the Unreasonable Effectiveness of PBT for Validating Formal Specifications

Understanding, Analyzing, and Optimizing Agentic AI: A CPU-Centric Perspective

Getting Confidence in (Agentic) Code

Compute Optimal Tokenization: Scaling Laws for Data Compression in LLMs

Nitsum: Serving Tiered LLM Requests with Adaptive Tensor Parallelism

SuperInfer: SLO-Aware Rotary Scheduling and Memory Management for LLM Inference

KV cache is becoming the memory hierarchy of inference

Ada-MK: Adaptive MegaKernel Optimization via DAG-Based Search for LLM Inference

3 4 5 6 7 8 9 10 11 12