# A Dataflow Compiler for Efficient LLM Inference using Custom Microscaling Formats

arXiv: https://arxiv.org/abs/2307.15517 · HTML: https://mlsanity.com/paper/2307.15517/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in KV cache quantization, one of 12 directions in Efficient LLM inference and serving.
- Builds on 31 of its 71 references that are in the map.
- First posted on arXiv 28 Jul 2023.

## Placed in

- Efficient LLM inference and serving / [KV cache quantization](https://mlsanity.com/inference/kv-cache-quantization/)

## Builds on (most cited first, 30 of 31)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://mlsanity.com/paper/1606.05250/) (2016)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions](https://mlsanity.com/paper/1905.10044/) (2019)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [What Makes Good In-Context Examples for GPT-3?](https://mlsanity.com/paper/2101.06804/) (2021)
- [LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale](https://mlsanity.com/paper/2208.07339/) (2022)
- [Language-agnostic BERT Sentence Embedding](https://mlsanity.com/paper/2007.01852/) (2020)
- [ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformers](https://mlsanity.com/paper/2206.01861/) (2022)
- [Quantizing deep convolutional networks for efficient inference: A whitepaper](https://mlsanity.com/paper/1806.08342/) (2018)
- [HAWQ: Hessian AWare Quantization of Neural Networks with Mixed-Precision](https://mlsanity.com/paper/1905.03696/) (2019)
- [LQ-Nets: Learned Quantization for Highly Accurate and Compact Deep Neural Networks](https://mlsanity.com/paper/1807.10029/) (2018)
- [FINN: A Framework for Fast, Scalable Binarized Neural Network Inference](https://mlsanity.com/paper/1612.07119/) (2016)
- [Towards Accurate Binary Convolutional Neural Network](https://mlsanity.com/paper/1711.11294/) (2017)
- [GOBO: Quantizing Attention-Based NLP Models for Low Latency and Energy Efficient Inference](https://mlsanity.com/paper/2005.03842/) (2020)
- [Convolutional Neural Networks using Logarithmic Data Representation](https://mlsanity.com/paper/1603.01025/) (2016)
- [Training and Inference with Integers in Deep Neural Networks](https://mlsanity.com/paper/1802.04680/) (2018)
- [VS-Quant: Per-vector Scaled Quantization for Accurate Low-Precision Neural Network Inference](https://mlsanity.com/paper/2102.04503/) (2021)
- [hls4ml: An Open-Source Codesign Workflow to Empower Scientific Low-Power Machine Learning Devices](https://mlsanity.com/paper/2103.05579/) (2021)
- [FTRANS: Energy-Efficient Acceleration of Transformers using FPGA](https://mlsanity.com/paper/2007.08563/) (2020)
- [DFX: A Low-latency Multi-FPGA Appliance for Accelerating Transformer-based Text Generation](https://mlsanity.com/paper/2209.10797/) (2022)
- [Torch.fx: Practical Program Capture and Transformation for Deep Learning in Python](https://mlsanity.com/paper/2112.08429/) (2021)
- [FLAT: An Optimized Dataflow for Mitigating Attention Bottlenecks](https://mlsanity.com/paper/2107.06419/) (2021)
- [Adaptable Butterfly Accelerator for Attention-based NNs via Hardware and Algorithm Co-design](https://mlsanity.com/paper/2209.09570/) (2022)
- [ScaleHLS: A New Scalable High-Level Synthesis Framework on Multi-Level Intermediate Representation](https://mlsanity.com/paper/2107.11673/) (2021)
- [Revet: A Language and Compiler for Dataflow Threads](https://mlsanity.com/paper/2302.06124/) (2023)
- [Automating Constraint-Aware Datapath Optimization using E-Graphs](https://mlsanity.com/paper/2303.01839/) (2023)

## Built on by (most cited first, 2 of 2)

- [hls4ml: A Flexible, Open-Source Platform for Deep Learning Acceleration on Reconfigurable Hardware](https://mlsanity.com/paper/2512.01463/) (2025)
- [Mixed-Precision Quantization for Language Models: Techniques and Prospects](https://mlsanity.com/paper/2510.16805/) (2025)

## Abstract (arXiv)

Model quantization represents both parameters (weights) and intermediate values (activations) in a more compact format, thereby directly reducing both computational and memory cost in hardware. The quantization of recent large language models (LLMs) faces challenges to achieve competitive memory density compared to other models such as convolutional neural networks, since values in LLMs require larger dynamic ranges. Current hardware can expedite computation for LLMs using compact numerical formats such as low-bitwidth integers or floating-point numbers. Each has advantages: integer operations simplify circuit design, whereas floating-point calculations can enhance accuracy when a wider dynamic range is required. In this work, we seek an efficient data format that combines the best of both worlds: Microscaling (MX) formats. MX formats are efficient data formats that achieve both large dynamic ranges and high memory density. In this paper, we propose a compiler named MASE for exploring mixed-precision MX formats on dataflow hardware accelerators for LLM inference. Our main contributions are twofold. First, we propose a novel orchestration abstraction to explore both software and hardware optimizations with new data formats. Second, MASE achieves LLM inference at an average precision of 4-bits, with minimal to no accuracy degradation. To our knowledge, MASE represents the first effort to harness fine-grain multi-precision MX formats in the design of LLM hardware accelerators. Over a range of LLMs and datasets, MASE achieves an average improvement of 24% in Delta accuracy with an overhead of only 3% in energy efficiency compared to designs using 8-bit fixed-point numbers.