Multithreading and work stealing

Use every core: parallel loops, work-stealing schedulers, task granularity and load balancing.

795proven techniques

41pitfalls: tried, then slower

352repositories

129from papers

Languages

C++416
Rust87
C76
Python40
Go31
CUDA26

Hardware

CPU452
x86119
NVIDIA GPU57
GPU34
ARM33

Fields

general232
data & compression145
graphics132
math102
GPU73

Where these come from

Repositories with the most techniques on this topic in the database:

blender/blender · 13OSGeo/grass · 9PaddlePaddle/FlyCV · 9blender/blender · 8blender/blender · 8sdcb/SimdPaddleOCR · 8succinctlabs/sp1 · 7rui314/mold · 7blender/blender · 7llvm/llvm-project · 7

Ask Claude for the full answer

Each technique comes with its bottleneck, measured speedup, trade-offs, conditions and the commit or paper behind it, served to Claude through MCP.

claude mcp add --transport http adtomic https://www.adtomic.co/mcp

Connect Claude Pricing