Tag
This article delves into the differences between C in the Linux kernel and ordinary userspace C, covering core techniques such as resource management, error handling, concurrency, logging, static analysis, and extensively using GNU C extensions and kernel-specific patterns.
Excited to share cuTile Rust, bringing Rust's fearless concurrency to GPU kernel programming. Their paper 'Fearless Concurrency on the GPU' is now on arXiv.
Discussion about rewriting parallelism to improve kernel performance using CuTe DSL and tile programming models for the FA4 (FlashAttention 4) kernel.