commander@lemmy.world to Hardware@lemmy.worldEnglish · 2 days agoHuawei unveils new Atlas 350 AI accelerator with 1.56 PFLOPS of FP4 compute and up to 112GB of HBM — claims 2.8x more performance than Nvidia's H20www.tomshardware.comexternal-linkmessage-square4linkfedilinkarrow-up113arrow-down11
arrow-up112arrow-down1external-linkHuawei unveils new Atlas 350 AI accelerator with 1.56 PFLOPS of FP4 compute and up to 112GB of HBM — claims 2.8x more performance than Nvidia's H20www.tomshardware.comcommander@lemmy.world to Hardware@lemmy.worldEnglish · 2 days agomessage-square4linkfedilink
minus-squarefuckwit_mcbumcrumble@lemmy.dbzer0.comlinkfedilinkEnglisharrow-up5·2 days agoAi work is usually very low precision. FP8 being the lowest you could go for a while meant it’s been the standard.
minus-squareBuffalox@lemmy.worldlinkfedilinkEnglisharrow-up4·2 days ago Ai work is usually very low precision. Apparently 😋
Ai work is usually very low precision. FP8 being the lowest you could go for a while meant it’s been the standard.
Apparently 😋