LLama inference implementation in Scala#82

Open

szymon-rd wants to merge 11 commits intomainfrom

Commits on Jan 31, 2026

LLama GPU Implementation
szymon-rd
committed

Commits on Feb 2, 2026

perf(llama): Multi-row matmul optimization (NUM_ROWS=2) - 3.5x speedup
szymon-rd
committed
perf(llama): Optimize F16OutputVec4Program with NUM_ROWS=4
szymon-rd
committed
perf(runtime): Smarter barrier insertion for parallel dispatch execution
szymon-rd
committed

Commits on Feb 6, 2026