• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** GPU Hierarchy ***

It's a great post. Is anyone aware of a similar chart for laptop gpus/cpus? Been considering getting one, but wondering what level I would need to get a similar spec to my desktop (on sig).

I'd guess 4080, but given the varying wattages, seems a very tricky game to play.

Edit to point out genuinely no pun intended.
 
Last edited:

Computerbase updated their list:

OljsaHa.png


PCFyQAZ.png


wQReBJv.png


ZDO7nIr.png
 
I have an EVGA 3070 which is currently tier E. Passmark's web site scores as 22,415 in its database. Because my card is manufacture-overclocked though, Passmark (on my local PC) scored it as 24,651 which is just a cat's whisker underneath a 3080. This moves it up to a tier D. No doubt it will drop to like tier P when the 5000 series come out :D

Also interesting to know that my 5 year old laptop with a 1660 Ti still scrapes through as entry level.
 
Do cards have the same ranking running in Linux? Not absolute performance, which I expect to be different. But the ranks ie 1st running in Windows still 1st in Linux for a given task.

I'm interested in getting a card for AI in linux.

I ask because stable diffusion performance is much higher for AMD on linux.
 
Do cards have the same ranking running in Linux? Not absolute performance, which I expect to be different. But the ranks ie 1st running in Windows still 1st in Linux for a given task.

I'm interested in getting a card for AI in linux.

I ask because stable diffusion performance is much higher for AMD on linux.
No this is purely compiled based on gaming performance.

Linux performance should be similar, and you'd expect ai performance to be similar in that more powerful cards should be better at it, but there are a few intricacies that may affect ai, e.g. which Cuda version is supported for Nvidia cards (e.g. old high end cards might not even be able to run it, as don't support a new enough version) and amount of vram (e.g. new low end cards might not have enough vram for the models you require)
 
Back
Top Bottom