Skip to content
#

instruction-level-parallelism

Here are 18 public repositories matching this topic...

IAA03_fast_math is a single-header math kernel(Atan2 only for now) designed to eliminate the "Trigonometry Tax" in high-throughput systems (Physics Engines, Audio DSP, and ML Pre-processing). Branchless ,ILP and SIMD (AVX2/SSE4.1), it achieves up to a ~186x per-element throughput speedup over std::atan2 while being IEEE 754 compliant

  • Updated Jan 14, 2026
  • C++

Superscalar CPU simulation models how modern processors execute instructions in parallel and manage hazards, analyzing and scheduling Python function instructions using advanced CPU features.

  • Updated Jun 25, 2025
  • Python

Un simulador moderno para aprender arquitectura de computadores. Tiene editor con resaltado de sintaxis, cronograma interactivo del pipeline, registros y memoria en tiempo real, y soporte para instrucciones enteras, de punto flotante y vectoriales.

  • Updated Oct 24, 2026
  • TypeScript

Add this topic to your repo

To associate your repository with the instruction-level-parallelism topic, visit your repo's landing page and select "manage topics."

Learn more