Soldato
- Joined
- 26 Sep 2013
- Posts
- 10,756
- Location
- West End, Southampton
Will small Vega and full fat Vega be launched together? I'm guessing one will be H1 '17 and the other 2H '17.
-
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 VEGA confirmed for 2017 H1
- Thread starter Doom112
- Start date
- Status
im guessing the same thinking they may have to concentrate on just the one to get something out the door sooner rather than later
Permabanned
- Joined
- 12 Sep 2013
- Posts
- 9,221
- Location
- Knowhere
We'll probably get a similar release to Fiji where we get the main chip closely followed by a cut down chip made from the failures. I just wish they'd do it already.
That's semi custom silicon - take nothing from it. The arch may be the same and it is x86 but apart from that MS could have had AMD tape out anything within reason. Or it could just be totally incorrect and use HBM.
Last edited:
I had that card too. We must remember that standards are different now also so the goalposts have not necessarily moved. In 2004, 1600x1200 was the benchmark set. If you wanted to play at that res now a £100 card could do it.
But we already know it is 2H 2017 so not like they are late to their own launch projection. Late vs the competition yes, but I am hoping Vega will be looking to compete with Volta, not Pascal which will have been out for a year by the time Vega comes out.
This is why I am almost certain that unless they ****** up the architecture we will easily get a card that will be able to be as good if not better than titan px. I will be here to gloat when this happens as I was one of the first here to predict this while people laughed it off saying they expect 1070/80 performance; or it will be a sad day for competition at the top end of the GPU market as Nvidia will have it all to themselves.
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,358
- Location
- Planet Earth
I actually want smaller Vega to use GDDR5/GDDR5X - it should enable them to fight Nvidia on price easier. Having to use HBM2 might mean a cheaper PCB,but the problem is HBM2 probably is not cheap as is all the effort needed to use an interposer,etc as that I believe that part of the integration needs to be done by another company.
No we don't.
Oh really? It must have been one of those rumours from wccftech you guys love quoting all the time, must have read it here so many times I thought it was an official thing!
My bad
Permabanned
- Joined
- 12 Sep 2013
- Posts
- 9,221
- Location
- Knowhere
Weren't WCCF also claiming this year not long ago? You need to ignore clickbait sites like that, What we know from a safe source, namely AMD is that it'll be sometime in the first half of 2017, But that's still a 6 month window sadly.Oh really? It must have been one of those rumours from wccftech you guys love quoting all the time, must have read it here so many times I thought it was an official thing!
My bad
You'd think they'd want a full year of revenue from the 480. OTOH, if Vega 11 isn't that big an improvement over the 480, then delaying it makes it look even worse when it eventually comes.
I mean, if Vega 11 is ~980 perf, and they release it closer to 2018 than 2016, then nV have a good chance to blow it away with their 1160.
That's assuming the early info is correct and the Vega 11 is only ~16% better than a 480.
I do not go there. Not been there since all the lies about the RX 480. WCCF claim all sorts of things. lol.
I read about 2H 2017 on this forum and it was said so many times I thought it was what AMD said
Unless they were going to release it at 470 prices, that'd be a totally failed product.
I doubt it's only 16% better than a 480.
I read about 2H 2017 on this forum and it was said so many times I thought it was what AMD said
That's one of the main problems with some of the bs thats spouted, a rumour suddenly becomes "amd said" and everyone believes it. Then the usual people that have an axe to grind beat it to death.
Caporegime
- Joined
- 8 Jul 2003
- Posts
- 30,063
- Location
- In a house
AMD RX 490 High-End 4K VR Graphics Card Inbound – Announcement and Release Expected in December
http://wccftech.com/radeon-rx-490-vr-4k-december/
Last edited:
His post looks confused to me.
He's talking about Vega being 16% faster than the 480 and still slower than the 980, where did he get this from?
The 480 is on par with the 980 in DX11, in DX12 and Vulkan it blows it out of the water.
If Vega can't beat the 980 it can't beat the 480
Last edited:
That's based on the known TF values. 480 being 5.8 and Vega 11 being 7. Actually that works out to 20% not 16%.
Given that the 480 is behind the 980 in some titles, that 20% will put it narrowly ahead of a 980 or a 1060. It really isn't going to be a big jump.
So again, if Vega 11 is closer to 2018, it's hardly worth shouting about.
"~" means "approximately"; "thereabouts" - not "worse than". So I don't need to refute your first point any further.
As for your 2nd point... nobody cares about extreme fringe cases. Vulkan games are nowhere. There are a couple DX12 titles at best.
The vast majority are DX11 console ports (and will continue to be so in future), and when we look at the vast majority of titles the 480 does not "blow the 980 out of the water".
I keep expecting posters here to rise about the "Green is better/Red is better" nonsense, and I keep being proved wrong :/
Btw my current card is a 480. The 980 is better in most DX11 games; I can handle that without turning red and exploding. It's time we all grew up here.
Ignoring Vulkan and DX12 just shown how little you understand, its all performance and there is a reason for the huge performance of the RX480 with those API's,
The compute and render throughput of the 480 is actually near that on the GTX 1070 (5.8 vs 6.5TFlops), thats why its much closer to the 1070 with those API's, it is simply bottlenecked in DX11.
Its the same reason why often the Fury-X is little or no faster than the RX 480 or 390X at 1080P, a massively more powerful GPU constrained to the same levels of performance as those much weaker GPU's.
If AMD improve their own DX11 overheads to the levels of Nvidia and add about 100Mhz to the RX 480 they already have a GTX 1070 level card.
With twice the number of shaders Vega would blow the GTX 1080 out of the water, possibly also beat Pacal TX.
Your ignorance of this is causing you to approach it from completely the wrong angle, AMD could make a GPU 5x as powerful as the 1080 and yet if they don't sort their own DX11 bottlenecks it will be little to no faster than the RX 480/.
Last edited:
It's my understanding that comparing TF to TF across vendors is not a good comparison by which to measure performance.
Polaris's 5.8 TF to Vega 11's 7 TF is more meaningful.
And no, Vulkan means nothing until there are Vulkan games widely available. Do we need to remember Mantle? It doesn't matter how awesome an API is (or could be) if it never hits critical mass.
They probably couldn't, even if they wanted to. nV can afford to throw money at R&D on a scale that AMD can't match. Same vs Intel on the CPU side.
Polaris's 5.8 TF to Vega 11's 7 TF is more meaningful.
And no, Vulkan means nothing until there are Vulkan games widely available. Do we need to remember Mantle? It doesn't matter how awesome an API is (or could be) if it never hits critical mass.
humbug said:
They probably couldn't, even if they wanted to. nV can afford to throw money at R&D on a scale that AMD can't match. Same vs Intel on the CPU side.
Last edited:
If AMD has a faster card than the XP then NV can only match or beat AMD if NV already have a faster GPU ready to go. R&D takes a lot of time, having the money is only half of the equation.
- Status