Performance Analysis

Modern GPU DSLs: Performance and Portability
Benchmarking RMSNorm and Flash Attention across Helion, Triton, Gluon, and CuTe on four NVIDIA GPU generations.
Mar 17, 2026 2 min read
PDF Code
Peeling Back the Layers: An Empirical Analysis of Modern ARM Primitives
A performance characterization of the Rockchip RK3588S SoC, exploring the gap between theoretical hardware limits and observed system behavior.
Jan 23, 2026 2 min read