misk@sopuli.xyz to Hardware@lemmy.world · 2 days agoNvidia's data center Blackwell GPUs reportedly overheat, require rack redesigns and cause delays for customerswww.tomshardware.comexternal-linkmessage-square7fedilinkarrow-up134arrow-down10cross-posted to: technology@lemmy.worldnvidia@lemmy.zip
arrow-up134arrow-down1external-linkNvidia's data center Blackwell GPUs reportedly overheat, require rack redesigns and cause delays for customerswww.tomshardware.commisk@sopuli.xyz to Hardware@lemmy.world · 2 days agomessage-square7fedilinkcross-posted to: technology@lemmy.worldnvidia@lemmy.zip
minus-squareRekall Incorporated@lemm.eelinkfedilinkarrow-up6·2 days ago 120kW per rack I knew GPU compute took a lot of energy, but I didn’t realize it was 120 kW per rack. That is a stupid amount.
I knew GPU compute took a lot of energy, but I didn’t realize it was 120 kW per rack. That is a stupid amount.