Skip to content

L LabsResearch

White Papers.

Read the blog

A closer look at the runtime: architecture, execution, and the machine underneath.

-jl 5 min read

RUNTIME / SIMDFeatured

SIMD: making every cycle count.

SIMD is a discipline, not a wider instruction. Masks instead of branches, in-place partition, and why correct parallel code can still be slow.

-jl 4 min read

RUNTIME / COMPRESSION

Compute on compressed.

Compression usually means storing less. In l it also means computing differently: running primitives on the encoding itself.

-jl 4 min read

RUNTIME / COMPRESSION

Compression: breaking the memory wall.

Modern analytical performance is bound by bytes, not arithmetic. How l treats compression as execution layout.