Eh this is to be expected seeing how nvidia handles their architecture. For the last few generations nvidia has designed their architecture around the enterprise world’s needs with gaming being an after thought. The fact is their gaming GPUs are defective enterprises GPUs not to say that’s a bad thing but nvidia does not sell any GeForce GPUs that utilizes the full die. Simply put the enterprise world is all in on AI workloads and ROPs do not benefit those workloads so why would Nvidia dedicate precious silicon to an after thought?
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: [email protected]
No game suggestions, friend requests, surveys, or begging.
No Let’s Plays, streams, highlight reels/montages, random videos or shorts.
No off-topic posts/comments, within reason.
Use the original source, no clickbait titles, no duplicates.
(Submissions should be from the original source if possible, unless from paywalled or non-english sources.
If the title is clickbait or lacks context you may lightly edit the title.)
Clickbait. Even the author admits that this doesn’t mean anything because the architecture is different and rops usually aren’t a bottleneck
And the first part is wrong too, top 2 dies’ Cuda counts are going up, rest are the same or going down.
Eh this is to be expected seeing how nvidia handles their architecture. For the last few generations nvidia has designed their architecture around the enterprise world’s needs with gaming being an after thought. The fact is their gaming GPUs are defective enterprises GPUs not to say that’s a bad thing but nvidia does not sell any GeForce GPUs that utilizes the full die. Simply put the enterprise world is all in on AI workloads and ROPs do not benefit those workloads so why would Nvidia dedicate precious silicon to an after thought?