Reducing the Cost of Dropout in Flash-Attention by Hiding RNG with GEMM

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ma, Haiyue, Liu, Jian, Krashinsky, Ronny
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916824666341376
author Ma, Haiyue
Liu, Jian
Krashinsky, Ronny
author_facet Ma, Haiyue
Liu, Jian
Krashinsky, Ronny
contents Dropout, a network operator, when enabled is likely to dramatically impact the performance of Flash-Attention, which in turn increases the end-to-end training time of Large-Language-Models (LLMs). The main contributor to such performance degradation is the Random Number Generation (RNG) phase. The state-of-the-art optimization is to fuse RNG into the Flash-Attention kernel. However, while RNG and Attention do not compete on compute or memory resources, they are bounded on the same lower-level architecture bottlenecks. Fusion can hardly hide RNG latency within the Attention kernel. We propose overlapping RNG with previous GEMM layers in the network to hide RNG latency and improve end-to-end performance. RNG and GEMM have distinct resource requirements and hardware bottlenecks, so they can run together without compromising each other's performance. We propose a fine-grained analytical performance model that analyzes low-level architecture resource utilization to evaluate RNG-GEMM overlapping performance benefits. This model, cross-validated by silicon results, shows 1.26x speedup for overlapping RNG and GEMM layers over a sequential implementation on one Transformer Block (one LLM layer including multi-head attention and feed-forward layers), and 1.22x over state-of-the-art fusion implementation, for Llama3 on GH100 GPUs with FP8 precision. Because the kernel patterns are regular, the findings of the shared bottlenecks, as well as the achievable performance benefits, can be generalized to different model architectures, software implementations and hardware configurations.
format Preprint
id arxiv_https___arxiv_org_abs_2410_07531
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Reducing the Cost of Dropout in Flash-Attention by Hiding RNG with GEMM
Ma, Haiyue
Liu, Jian
Krashinsky, Ronny
Hardware Architecture
Artificial Intelligence
Dropout, a network operator, when enabled is likely to dramatically impact the performance of Flash-Attention, which in turn increases the end-to-end training time of Large-Language-Models (LLMs). The main contributor to such performance degradation is the Random Number Generation (RNG) phase. The state-of-the-art optimization is to fuse RNG into the Flash-Attention kernel. However, while RNG and Attention do not compete on compute or memory resources, they are bounded on the same lower-level architecture bottlenecks. Fusion can hardly hide RNG latency within the Attention kernel. We propose overlapping RNG with previous GEMM layers in the network to hide RNG latency and improve end-to-end performance. RNG and GEMM have distinct resource requirements and hardware bottlenecks, so they can run together without compromising each other's performance. We propose a fine-grained analytical performance model that analyzes low-level architecture resource utilization to evaluate RNG-GEMM overlapping performance benefits. This model, cross-validated by silicon results, shows 1.26x speedup for overlapping RNG and GEMM layers over a sequential implementation on one Transformer Block (one LLM layer including multi-head attention and feed-forward layers), and 1.22x over state-of-the-art fusion implementation, for Llama3 on GH100 GPUs with FP8 precision. Because the kernel patterns are regular, the findings of the shared bottlenecks, as well as the achievable performance benefits, can be generalized to different model architectures, software implementations and hardware configurations.
title Reducing the Cost of Dropout in Flash-Attention by Hiding RNG with GEMM
topic Hardware Architecture
Artificial Intelligence
url https://arxiv.org/abs/2410.07531