CORTEXA
← Browse

Om Mohite

1 paper indexed

arxivcs.ARcs.DCcs.LGcs.PF2026-07-24

FusionML: Prefill, Not Decode - Mechanism and Boundaries of CPU+GPU Co-Execution on Unified-Memory Apple Silicon

Om Mohite

Apple-Silicon SoCs share CPU, GPU, and Neural Engine over one unified memory system, raising the question of whether transformer inference can be accelerated by splitting single operators across units. Prior attempts, including our own, failed or produced precision-confounded win…

View free PDFSource page