New compiler DCC optimizes ML kernels for Processing-In-Memory architectures

Researchers have developed DCC, a novel data-centric compiler designed to optimize machine learning kernels for Processing-In-Memory (PIM) architectures. This compiler addresses the challenges of data rearrangement and compute code optimization by jointly optimizing these interdependent processes. DCC supports multiple PIM backends through a multi-layer abstraction and has demonstrated significant speedups, achieving up to 7.68x on HBM-PIM and 13.17x on AttAcc PIM compared to GPU-only execution. For end-to-end LLM inference, DCC on AttAcc accelerated GPT-3 and LLaMA-2 by an average of 4.52x. AI

IMPACT Enables significant acceleration for LLM inference and other ML workloads on specialized Processing-In-Memory hardware.

RANK_REASON The cluster contains an academic paper detailing a new compiler for optimizing machine learning kernels on specific hardware architectures. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• AttAcc PIM
• DCC
• GPT-3
• HBM-PIM
• LLaMA-2
• Machine Learning
• Processing-In-Memory

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →