Metraipolis › archive › dc_31f877b2a262
MXCode: Vector Quantization for MX-Native Inference at Scale — arXiv:2611.00403
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.00403
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T18:35:10.202Z
- SHA-256 of the archived text
43a3230d9c4d0e6575cdfd61c6b344380d8b6840404ea9e4788c7b54eb5f034f- Stable URL of this page
https://metraipolis.com/documento/dc_31f877b2a262
How to cite
MXCode: Vector Quantization for MX-Native Inference at Scale — arXiv:2611.00403 — arXiv (en), https://arxiv.org/abs/2611.00403 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_31f877b2a262 (Metraipolis archive dc_31f877b2a262, sha256 43a3230d9c4d0e6575cdfd61c6b344380d8b6840404ea9e4788c7b54eb5f034f)
BibTeX
@misc{dc31f877b2a262,
title = {MXCode: Vector Quantization for MX-Native Inference at Scale — arXiv:2611.00403},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.00403},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_31f877b2a262; sha256 43a3230d9c4d0e6575cdfd61c6b344380d8b6840404ea9e4788c7b54eb5f034f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.00403. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Memory capacity and bandwidth are major bottlenecks in LLM inference, motivating weight compression beyond conventional 4-bit quantization. Vector quantization (VQ) enables approximately 2-bit weight storage but often relies on codebooks incompatible with native FP4 computation. We introduce MXCode (MX-Native Codebook Quantization), a framework bridging ultra-low-bit VQ compression with MicroXcaling (MX)-native FP4 computation. MXCode proposes three strategies: PostScale, which quantizes learned codebooks to MXFP4; PreScale, which extracts MX scales before codebook learning; and PreScale+, which further compresses these scales using VQ. These strategies explore tradeoffs among quantization quality, scale overhead, and codebook granularity. All reconstruct MX-native W4A4 operands without materializing intermediate high-precision weights. We evaluate MXCode on Qwen3 (235B), DeepSeek-R1 (671B), and GLM-5.2 (753B) using the Microsoft MX Emulation Library. MXCode achieves approximately 2 bits per weight with competitive, workload-dependent model quality. On GB300 hardware, experiments with Qwen3 (235B) show approximately 50% lower model-load memory than Native FP4 while maintaining near-Native-FP4 serving throughput. Under memory-constrained prefix reuse, P90 TTFT improves by up to 94.5%. These results demonstrate software-based reconstruction on existing FP4 hardware and inform the development of future MX-native inference systems. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.00403 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-11-03T18:12:54.414Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).