SQ-format: A Unified Sparse-Quantized Hardware-friendly Data Format for LLMs

TL;DR

The paper introduces SQ-format, a unified sparse-quantized data format designed to enhance the efficiency and accuracy of large language model quantization, compatible with existing hardware and future accelerators.

Contribution

It proposes a novel unified data format that combines sparsification and quantization, improving performance and accuracy in LLM deployment.

Findings

Achieves state-of-the-art PTQ performance with SQ-format.

Supports hardware acceleration for sparse and low-precision matrix multiplication.

Provides design insights for next-generation AI accelerators.

Abstract

Post-training quantization (PTQ) plays a crucial role in the democratization of large language models (LLMs). However, existing low-bit quantization and sparsification techniques are difficult to balance accuracy and efficiency due to the limited hardware support. For example, W4A8 can only achieve the same peak TOPS as W8A8 whereas the GPU-supported sparse data format (2:4 semi-structure sparse) is seldomly adopted due to the loss of accuracy. To bridge this gap, in this paper, we propose the Sparse-Quantized Format (SQ-format), which is a unified data format for quantization and sparsification potentially easily supported by new hardware and existing GPUs. SQ-format makes use of the fact that sparse matrix can be accelerated in high-precision, and low-precision matrix multiplication can also be accelerated accordingly. As such, SQ-format is proposed to achieve Pareto improvement between performance and throughput. This format is particularly suitable for activations with outlier inequality status and makes their static compression possible. We show the state-of-the-art PTQ performance with SQ-format, propose the hardware required to support it, and further offer the design exploration and insights for the next-generation AI accelerators.