performance
Everything on Ground Truth tagged “performance” — 3 items.
Amdahl’s law: why a faster AI kernel barely changes the waiting time Lesson
Amdahl’s law limits an application’s speedup to the share of its original running time affected by an optimization, explaining why faster AI kernels may produce modest user-visible gains.
GigaToken Claims a 1,000x Tokenizer Speedup, But the Real Serving Win Is Single-Digit News
A new tokenizer called GigaToken reports roughly 1,000 times the offline throughput of Hugging Face Tokenizers, but its benchmark is not apples-to-apples and its real inference payoff is a single-digit latency improvement.
GigaToken Tool
An open-source native BPE tokenizer optimized with SIMD byte scanning and instruction-level parallelism, best used for fast offline corpus preparation and bulk token counting rather than end-to-end serving speedups.