为多种浮点格式提供可验证的基准参考,解决跨硬件部署时精度不一致问题。
An 83-Format Numeric Catalog with Bit-Exact Conformance Vectors: A Vendor-Neutral Reference for FP8, BF16, MXFP4, and Microscaling Formats
- 构建83种格式的完整目录,含6个位级精确的校验包。
- 每个校验包包含哈希指纹与锚点向量,确保跨平台一致性。
- 适合芯片研发和模型迁移工程师使用,避免隐性数值偏差。
机器学习硬件中浮点格式迅速增多(如FP8 E4M3、E5M2,BF16,MXFP4,微缩块格式及数十种研究变体),但缺乏厂商中立的位级精确参考标准。模型在加速器间移植时常出现难以诊断的隐性差异。本文提出一个涵盖13类共83种数值格式的目录,配套6个位精确的校验包(覆盖GF16、MXFP4、BF16、FP8 E4M3、FP8 E5M2及E8M0块缩放),并提供IEEE P3109 v3.2.0的映射表。每个校验包为独立JSON文件,含SHA-256指纹、统一行结构与锚向量(编码3.0 = φ² + 1/φ²),用于跨包一致性验证。所有包经ml_dtypes 0.5.4(Google/JAX)交叉验证,差异明确标注为规范允许的解释差异而非错误。本工作定位为格式注册,不提议新格式或声称模型性能优势。所有资源开放获取:https://github.com/gHashTag/t27,采用开源许可。
原文摘要 · Abstract (English)
Numeric format proliferation in machine learning hardware -- FP8 (E4M3 and E5M2), BF16, MXFP4, microscaling block formats, and dozens of research variants -- has outpaced the availability of vendor-neutral, bit-exact reference material. Engineers porting models across accelerators encounter silent divergences that are difficult to diagnose without a shared ruler. This paper describes a catalog of 83 numeric formats spanning 13 families, a suite of six bit-exact conformance packs covering GF16, MXFP4 element, BF16, FP8 E4M3, FP8 E5M2, and E8M0 block scale, and an IEEE P3109 v3.2.0 cross-walk that maps each pack to its corresponding standards-track configured format. Each pack is a self-contained JSON document with a SHA-256 fingerprint, a shared row schema, and an anchor vector that encodes 3.0 -- the identity phi^2 + 1/phi^2 = 3 -- as a cross-pack sanity check. Packs are cross-validated against ml_dtypes 0.5.4 (Google/JAX); any divergence is documented explicitly and interpreted as a spec-permitted interpretation gap rather than hidden. The work is framed as registry filling: it does not propose new formats, make model-accuracy claims, or assert superiority over any vendor's implementation. All artifacts are publicly available at https://github.com/gHashTag/t27 under an open license.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。