# FP8-Flow-MoE: A Casting-Free FP8 Recipe without Double Quantization Error

arXiv: https://arxiv.org/abs/2511.02302 · HTML: https://mlsanity.com/paper/2511.02302/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Efficient MoE training and inference systems, one of 12 directions in Mixture-of-experts.
- Builds on 13 of its 20 references that are in the map.
- First posted on arXiv 4 Nov 2025.

## Placed in

- Mixture-of-experts / [Efficient MoE training and inference systems](https://mlsanity.com/moe/efficient-moe-training-and-inference-systems/)

## Builds on (most cited first, 13 of 13)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Sparks of Artificial General Intelligence: Early experiments with GPT-4](https://mlsanity.com/paper/2303.12712/) (2023)
- [GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding](https://mlsanity.com/paper/2006.16668/) (2020)
- [Mixed Precision Training](https://mlsanity.com/paper/1710.03740/) (2017)
- [DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scale](https://mlsanity.com/paper/2201.05596/) (2022)
- [FP8 Formats for Deep Learning](https://mlsanity.com/paper/2209.05433/) (2022)
- [Training Deep Neural Networks with 8-bit Floating Point Numbers](https://mlsanity.com/paper/1812.08011/) (2018)
- [FP8-LM: Training FP8 Large Language Models](https://mlsanity.com/paper/2310.18313/) (2023)
- [Insights into DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architectures](https://mlsanity.com/paper/2505.09343/) (2025)

## Built on by (most cited first, 1 of 1)

- [Practical FP4 Training for Large-Scale MoE Models on Hopper GPUs](https://mlsanity.com/paper/2603.02731/) (2026)

## Abstract (arXiv)

Training large Mixture-of-Experts (MoE) models remains computationally prohibitive due to their extreme compute and memory demands. Although low-precision training promises to accelerate computation and reduce memory footprint, existing implementations still rely on BF16-dominated dataflows with frequent quantize-dequantize (Q/DQ) conversions. These redundant casts erode much of FP8's theoretical efficiency. However, naively removing these casts by keeping dataflows entirely in FP8 introduces double quantization error: tensors quantized along different dimensions accumulate inconsistent scaling factors, degrading numerical stability. We propose FP8-Flow-MoE, an FP8 training recipe featuring a quantization-consistent FP8-centric dataflow with a scaling-aware transpose and fused FP8 operators that streamline computation and eliminate explicit cast operations from 12 to 2. Evaluations on a 671B-parameter MoE model demonstrate up to 21% higher throughput and 16.5 GB lower memory usage per GPU compared to BF16 and na"ive FP8 baselines, while maintaining stable convergence. We provide a plug-and-play FP8 recipe compatible with TransformerEngine and Megatron-LM, which will be open-sourced soon.