# Fast and Efficient 2-bit LLM Inference on GPU: 2/4/16-bit in a Weight Matrix with Asynchronous Dequantization

arXiv: https://arxiv.org/abs/2311.16442 · HTML: https://mlsanity.com/paper/2311.16442/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 40 references that are in the map.
- First posted on arXiv 28 Nov 2023.

## Builds on (most cited first, 22 of 22)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation](https://mlsanity.com/paper/2305.01210/) (2023)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot](https://mlsanity.com/paper/2301.00774/) (2023)
- [A White Paper on Neural Network Quantization](https://mlsanity.com/paper/2106.08295/) (2021)
- [Recent Advances in Natural Language Processing via Large Pre-Trained Language Models: A Survey](https://mlsanity.com/paper/2111.01243/) (2021)
- [SpQR: A Sparse-Quantized Representation for Near-Lossless LLM Weight Compression](https://mlsanity.com/paper/2306.03078/) (2023)
- [Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT](https://mlsanity.com/paper/1909.05840/) (2019)
- [Sign Language Transformers: Joint End-to-end Sign Language Recognition and Translation](https://mlsanity.com/paper/2003.13830/) (2020)
- [LEVER: Learning to Verify Language-to-Code Generation with Execution](https://mlsanity.com/paper/2302.08468/) (2023)
- [GOBO: Quantizing Attention-Based NLP Models for Low Latency and Energy Efficient Inference](https://mlsanity.com/paper/2005.03842/) (2020)
- [Shortcut Learning of Large Language Models in Natural Language Understanding](https://mlsanity.com/paper/2208.11857/) (2022)
- [ChatGPT and a New Academic Reality: Artificial Intelligence-Written Research Papers and the Ethics of the Large Language Models in Scholarly Publishing](https://mlsanity.com/paper/2303.13367/) (2023)

## Built on by (most cited first, 14 of 14)

- [Parameter-Efficient Fine-Tuning Methods for Pretrained Language Models: A Critical Review and Assessment](https://mlsanity.com/paper/2312.12148/) (2023)
- [Designing Large Foundation Models for Efficient Training and Inference: A Survey](https://mlsanity.com/paper/2409.01990/) (2024)
- [Towards Message Brokers for Generative AI: Survey, Challenges, and Opportunities](https://mlsanity.com/paper/2312.14647/) (2023)
- [Mixed-Precision Quantization for Language Models: Techniques and Prospects](https://mlsanity.com/paper/2510.16805/) (2025)
- [SoftmAP: Software-Hardware Co-design for Integer-Only Softmax on Associative Processors](https://mlsanity.com/paper/2411.17847/) (2024)
- [FlashEdit: Decoupling Speed, Structure, and Semantics for Precise Image Editing](https://mlsanity.com/paper/2509.22244/) (2025)
- [ScaleBITS: Scalable Bitwidth Search for Hardware-Aligned Mixed-Precision LLMs](https://mlsanity.com/paper/2602.17698/) (2026)
- [OASIS: Outlier-Aware LUT-Based GEMM with Dual-Side Quantization for LLM Inference Acceleration](https://mlsanity.com/paper/2507.23035/) (2025)
- [CLAQ: Pushing the Limits of Low-Bit Post-Training Quantization for LLMs](https://mlsanity.com/paper/2405.17233/) (2024)
- [SFMP: Fine-Grained, Hardware-Friendly and Search-Free Mixed-Precision Quantization for Large Language Models](https://mlsanity.com/paper/2602.01027/) (2026)
- [Extreme Low-Bit Inference in Reasoning Models: Failure Modes and Targeted Recovery](https://mlsanity.com/paper/2606.02011/) (2026)
- [Breaking the 1.58-bit Barrier for Ternary LLMs](https://mlsanity.com/paper/2609.16338/) (2026)
- [SimReg: Achieving Higher Performance in the Pretraining via Embedding Similarity Regularization](https://mlsanity.com/paper/2605.08809/) (2026)
- [FlashQuant: Sparse-Dense Fusion for Memory-Efficient Outlier-Aware LLM Inference](https://mlsanity.com/paper/2608.15531/) (2026)

## Abstract (arXiv)

Large language models (LLMs) have demonstrated impressive abilities in various domains while the inference cost is expensive. Many previous studies exploit quantization methods to reduce LLM inference cost by reducing latency and memory consumption. Applying 2-bit single-precision weight quantization brings >3% accuracy loss, so the state-of-the-art methods use mixed-precision methods for LLMs (e.g. Llama2-7b, etc.) to improve the accuracy. However, challenges still exist: (1) Uneven distribution in weight matrix. (2) Large speed degradation by adding sparse outliers. (3) Time-consuming dequantization operations on GPUs. To tackle these challenges and enable fast and efficient LLM inference on GPUs, we propose the following techniques in this paper. (1) Intra-weight mixed-precision quantization. (2) Exclusive 2-bit sparse outlier with minimum speed degradation. (3) Asynchronous dequantization. We conduct extensive experiments on different model families (e.g. Llama3, etc.) and model sizes. We achieve 2.91-bit for each weight considering all scales/zeros for different models with negligible loss. As a result, with our 2/4/16 mixed-precision quantization for each weight matrix and asynchronous dequantization during inference, our design achieves an end-to-end speedup for Llama2-7b is 1.74x over the original model, and we reduce both runtime cost and total cost by up to 2.53x and 2.29x with less GPU requirements.