lemmydividebyzero@reddthat.com to Technology@lemmy.worldEnglish · 2 days agoDatacenters in space are a terrible, horrible, no good idea.taranis.ieexternal-linkmessage-square285linkfedilinkarrow-up1666arrow-down113cross-posted to: pulse_of_truth@infosec.pub
arrow-up1653arrow-down1external-linkDatacenters in space are a terrible, horrible, no good idea.taranis.ielemmydividebyzero@reddthat.com to Technology@lemmy.worldEnglish · 2 days agomessage-square285linkfedilinkcross-posted to: pulse_of_truth@infosec.pub
minus-squarebrucethemoose@lemmy.worldlinkfedilinkEnglisharrow-up1·edit-21 day ago So from the inference / customer computation side of things, it isn’t a problem. Not necessarily. There are inference schemes where spreading MoE models across 40+ GPUs with a fast interconnect yields better efficiency. looks like blackwells can run sunstained at 88c The coolant still needs to remain relatively cool to hold that silicon temperature, though. Practically it can’t be like 60C.
minus-squareNotMyOldRedditName@lemmy.worldlinkfedilinkEnglisharrow-up1·edit-21 day ago The coolant still needs to remain relatively cool to hold that silicon temperature, though. Practically it can’t be like 60C. Ah, ya that makes sense, whatever the numbers the the chip can be the coolant will be less.
Not necessarily. There are inference schemes where spreading MoE models across 40+ GPUs with a fast interconnect yields better efficiency.
The coolant still needs to remain relatively cool to hold that silicon temperature, though. Practically it can’t be like 60C.
Ah, ya that makes sense, whatever the numbers the the chip can be the coolant will be less.