Cache locality and data reordering

Lay data out the way it is read: structure-of-arrays, tiling, blocking, sorted access and hot/cold splitting.

897proven techniques

5pitfalls: tried, then slower

391repositories

376from papers

Languages

C++297
C110
CUDA93
Python30
Rust27
Go10

Hardware

CPU253
NVIDIA GPU187
GPU102
x8688
ARM28

Fields

GPU180
graphics168
data & compression144
general96
math93

Where these come from

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

flashinfer-ai/flashinfer · 5google/tcmalloc · 5zeux/meshoptimizer · 5ceres-solver/ceres-solver · 4torvalds/linux · 4h2oai/h2o-3 · 4NVIDIAGameWorks/dxvk-remix · 4flame/blis · 4cdcseacave/openMVS · 4vllm-project/vllm · 3

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