Leveraging the bfloat16 Artificial Intelligence Datatype For Higher-Precision Computations · HackerPapers discussionshttps://hackerpaper.kaipfister.de/p/1904.06376/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMTLeveraging the bfloat16 Artificial Intelligence Datatype For Higher-Precision Computations — mentioned by kolbusahttps://news.ycombinator.com/item?id=29482176urn:hackerpapers:mention:1904.06376:29482176Wed, 08 Dec 2021 07:38:51 GMT<p>I don&#x27;t think so. The new ISA that is in the SPR is mostly about deep learning: it supports int8 and bfloat16 (https://fuse.wikichip.org/news/3600/the-x86-advanced-matrix-...). You can emulate higher precision using bfloat16 (https://arxiv.org/abs/1904.06376), but I have not seen this used in the wild.</p><p><a href="https://hackerpaper.kaipfister.de/p/1904.06376/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.MS