themachinestops@lemmy.dbzer0.com to Technology@lemmy.worldEnglish · 20 hours agoAMD: Ryzen Desktop CPUs Reportedly See a Price Increase of Over 50% in Some Countriesin.ign.comexternal-linkmessage-square9fedilinkarrow-up1149
arrow-up1149external-linkAMD: Ryzen Desktop CPUs Reportedly See a Price Increase of Over 50% in Some Countriesin.ign.comthemachinestops@lemmy.dbzer0.com to Technology@lemmy.worldEnglish · 20 hours agomessage-square9fedilink
minus-squareNotMyOldRedditName@lemmy.worldlinkfedilinkEnglisharrow-up3·16 hours agoModels can split loads across a discrete GPU and CPU/RAM. Its not as fast as if you can load it all in the GPU, but it gives you more options. Its been quite common for a long time.
minus-squareJesus_666@lemmy.worldlinkfedilinkEnglisharrow-up3·15 hours agoYeah, that’s what I refer to with offloading. Depending on the model and runtime it might be a bit fiddly but it usually works fine.
minus-squareNotMyOldRedditName@lemmy.worldlinkfedilinkEnglisharrow-up3·15 hours agoIm apparently just bad at reading the whole message.
Models can split loads across a discrete GPU and CPU/RAM.
Its not as fast as if you can load it all in the GPU, but it gives you more options. Its been quite common for a long time.
Yeah, that’s what I refer to with offloading. Depending on the model and runtime it might be a bit fiddly but it usually works fine.
Im apparently just bad at reading the whole message.