• 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.

AMD Navi 23 ‘NVIDIA Killer’ GPU Rumored to Support Hardware Ray Tracing, Coming Next Year

Status
Not open for further replies.
Polaris and RDNA1 were all good midrange cards and very competitive on the price/performance battle; it's a shame we never got any big die variants to compete at high end.

Hopefully we get a full product stack this time and at least an attempt to compete at the higher end.
 
Why did Vega have such a high TFlop capability but performed like nothing of the sort?


As far as I know it the architecture was designed to cover both gaming and the professional space and a lot of the design went into compute capability. This meant a lot of the capability wasn't really geared to gaming from the start which meant it wasn't as efficient per TFlop as a lot were expecting.

Considering the shoestring budget AMD have worked on the last 5 years or so it's impressive they made one as competitive as they did in my opinion.
 
As far as I know it the architecture was designed to cover both gaming and the professional space and a lot of the design went into compute capability. This meant a lot of the capability wasn't really geared to gaming from the start which meant it wasn't as efficient per TFlop as a lot were expecting.

Considering the shoestring budget AMD have worked on the last 5 years or so it's impressive they made one as competitive as they did in my opinion.
It was a decent card, no doubt. I had one but I found it too hot and noisy. Made my bedroom like a sauna in the summer.
 
Why did Vega have such a high TFlop capability but performed like nothing of the sort?
As far as I know it the architecture was designed to cover both gaming and the professional space and a lot of the design went into compute capability. This meant a lot of the capability wasn't really geared to gaming from the start which meant it wasn't as efficient per TFlop as a lot were expecting.

Considering the shoestring budget AMD have worked on the last 5 years or so it's impressive they made one as competitive as they did in my opinion.

Vega can actually be decently fast even in today's titles considering its age and problematic start. The issue however has been work that needed to be done, aka optimization due to the architecture. Now I don't find it bad anymore due to more modern APIs(Vulkan, DX12) being used and better dx11 support with newer driver releases. Now of course it's not perfect and a rare game here and there will show some lackluster performance but I find those to be rare nowadays.

Also, it's not really a GPU for the nonenthusiast as to really get the GPU shining you have to tinker with it.
 
Why did Vega have such a high TFlop capability but performed like nothing of the sort?

There was a bunch of features intended for it that never got completed/switched on. (As I said at the time would happen to much derision from the usual suspects).

Major hype train derailment that one was

Polaris should have been a fair bit better to be fair - early samples indicated much better voltage/frequency capabilities - I think a lot of the reason for the flop (relatively speaking) was down to GlobalFoundries. If they'd manufactured them at Samsung you'd have probably have been looking at a 20-25% faster GPU at each tier.
 
Polaris should have been a fair bit better to be fair - early samples indicated much better voltage/frequency capabilities - I think a lot of the reason for the flop (relatively speaking) was down to GlobalFoundries. If they'd manufactured them at Samsung you'd have probably have been looking at a 20-25% faster GPU at each tier.
Yea, maybe. Shame as if they were it would have done much better, still not met the expectations of the hype, but not would not have got half the negativity it did in the end. Did AMD still have contractual obligations with GloFo at the time I wonder? Happy that they are using TSMC these days.
 
Did AMD still have contractual obligations with GloFo at the time I wonder?

I might have my timelines mixed up but IIRC yeah they had to pay GF a certain amount for every wafer they had produced elsewhere for much of their product range (as well as other agreements).
 
"Big Navi is a halo product” and “enthusiasts love to buy the best, and we are certainly working on giving them the best."

This makes it sound like its going to cost more than people are expecting or just the usual marketing talk, I really hope this isn't going to be another dud release from AMD.
 
Last edited:
This makes it sound like its going to cost more than people are expecting or just the usual marketing talk, I really hope this isn't going to be another dud release from AMD.


Depends what you consider a "dud", given past history nvidia are probably cooking up something to launch a few days previous to the new amd cards, as has been the case for a lot of the amd launches, fury x had 980ti launching ahead of it, vega had 1080ti a few months ahead of it, and 5700xt had the "super" cards a few days before their launch.

if amd want to retake the performance crown they have to produce something that violates the 2080ti so bad it'll think it just dropped the soap in prison, and they also need to it be competitive with ampere so it's not a short term win, so they need something about 40% or so faster than the 2080ti. Not an easy ask.
 
As far as I know it the architecture was designed to cover both gaming and the professional space and a lot of the design went into compute capability. This meant a lot of the capability wasn't really geared to gaming from the start which meant it wasn't as efficient per TFlop as a lot were expecting.
I did a lot of OpenCL work with mine in Capture 1 and it was an absolute beast for that. It also paid for itself when I mined with if in the crypto craziness (it would have paid for itself and the rest of the machine if I'd sold it at the time though!)
 
I did a lot of OpenCL work with mine in Capture 1 and it was an absolute beast for that. It also paid for itself when I mined with if in the crypto craziness (it would have paid for itself and the rest of the machine if I'd sold it at the time though!)

That's what I mean. It was a hell of an all rounder, where Nvidia had tesla cards at the top end and their main architectures for the gaming cards. It meant Nvidia could specialise a bit more.

Obviously the internet loved claiming Vega was a failure but I am pretty sure AMD actually did pretty well financially with it.
 
Status
Not open for further replies.
Back
Top Bottom