paper-with-me

홈 › Papers

DFX: A Low-latency Multi-FPGA Appliance for Accelerating Transformer-based Text Generation

2022-09-22 · Seongmin Hong, Seungjae Moon, Junsoo Kim, Sungjae Lee, Minsub Kim, Dongsoo Lee, Joo-Young Kim

Transformer is a deep learning language model widely used for natural language processing (NLP) services in datacenters. Among transformer models, Generative Pre-trained Transformer (GPT) has achieved remarkable performance in text generation, or natural language generation (NLG), which needs the processing of a large input context in the summarization stage, followed by the generation stage that produces a single word at a time. The conventional platforms such as GPU are specialized for the parallel processing of large inputs in the summarization stage, but their performance significantly degrades in the generation stage due to its sequential characteristic. Therefore, an efficient hardware platform is required to address the high latency caused by the sequential characteristic of text generation. In this paper, we present DFX, a multi-FPGA acceleration appliance that executes GPT-2 model inference end-to-end with low latency and high throughput in both summarization and generation stages. DFX uses model parallelism and optimized dataflow that is model-and-hardware-aware for fast simultaneous workload execution among devices. Its compute cores operate on custom instructions and provide GPT-2 operations end-to-end. We implement the proposed hardware architecture on four Xilinx Alveo U280 FPGAs and utilize all of the channels of the high bandwidth memory (HBM) and the maximum number of compute resources for high hardware efficiency. DFX achieves 5.58x speedup and 3.99x energy efficiency over four NVIDIA V100 GPUs on the modern GPT-2 model. DFX is also 8.21x more cost-effective than the GPU appliance, suggesting that it is a promising solution for text generation workloads in cloud datacenters.

📄 PDF Abstract BibTeX arXiv:2209.10797

Code (0)

등록된 구현이 없습니다.

Tasks

GPULanguage ModellingText Generation

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Cosine Annealing Cosine Annealing is a type of learning rate schedule that has the effect of starting with a large learning rate that is relatively rapidly decreased to a minimum value before…
Linear Warmup With Cosine Annealing Linear Warmup With Cosine Annealing is a learning rate schedule where we increase the learning rate linearly for $n$ updates and then anneal according to a cosine schedule…
Weight Decay 설명 없음
Attention Dropout Attention Dropout is a type of dropout used in attention-based architectures, where elements are randomly dropped out of the…
Discriminative Fine-Tuning Discriminative Fine-Tuning is a fine-tuning strategy that is used for ULMFiT type models. Instead of using the same learning rate…
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…

Similar Papers 제목 키워드 기반

Full-stack Optimization for Accelerating CNNs with FPGA Validation

2019-05-01 · Bradley McDanel, Sai Qian Zhang, H. T. Kung, Xin Dong

We present a full-stack optimization framework for accelerating inference of CNNs (Convolutional Neural Networks) and validate the approach with field-programmable gate arrays (FPGA) implementations. By jointly optimizin…

Accelerating Recurrent Neural Networks for Gravitational Wave Experiments

2021-06-26 · Zhiqiang Que, Erwei Wang, Umar Marikar, Eric Moreno 외

This paper presents novel reconfigurable architectures for reducing the latency of recurrent neural networks (RNNs) that are used for detecting gravitational waves. Gravitational interferometers such as the LIGO detector…

High-Level SynthesisTime SeriesTime Series Analysis

Systolic-CNN: An OpenCL-defined Scalable Run-time-flexible FPGA Accelerator Architecture for Accelerating Convolutional Neural Network Inference in Cloud/Edge Computing

2020-12-06 · Akshay Dua, Yixing Li, Fengbo Ren

This paper presents Systolic-CNN, an OpenCL-defined scalable, run-time-flexible FPGA accelerator architecture, optimized for accelerating the inference of various convolutional neural networks (CNNs) in multi-tenancy clo…

Edge-computing

Low Latency Transformer Inference on FPGAs for Physics Applications with hls4ml

2024-09-08 · Zhixing Jiang, Dennis Yin, Yihui Chen, Elham E Khoda 외

This study presents an efficient implementation of transformer architectures in Field-Programmable Gate Arrays(FPGAs) using hls4ml. We demonstrate the strategy for implementing the multi-head attention, softmax, and norm…

Design Flow of Accelerating Hybrid Extremely Low Bit-width Neural Network in Embedded FPGA

2018-07-31 · Junsong Wang, Qiuwen Lou, Xiaofan Zhang, Chao Zhu 외

Neural network accelerators with low latency and low energy consumption are desirable for edge computing. To create such accelerators, we propose a design flow for accelerating the extremely low bit-width neural network …

Edge-computingGPUQuantization