• 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

Status
Not open for further replies.
Rroff said:
I played through a good bit of Doom on a 970 - 1440p everything turned up was most of the time comfortably well over 60fps.

Was a bit of a mixed story though - some older games where performance came down a lot to shovelling lots of primitive geometry, etc. my 780 could be quite a bit faster.

They are all old cards now though despite the justifications of some - if it wasn't for the fact nVidia have trickled out performance increments (and AMD were nowhere to be seen) they'd have been utterly obsolete months ago as developers targetted a new generation of hardware.
Click to expand...
Games that are already designed to run at 60fps on consoles will inevitably run better on PC, too. And Doom is especially well-optimized. This is not the norm for AAA releases, though. Even though there's been a fair few quite recently with the multiplayer FPS trifecta of TF2, BF1 and IW.

On the whole, people who think a 970 or 290/390 are good for 1440p are going to end up disappointed in plenty of titles, as they might already be. And no, I'm not talking about just max settings. Even with a 970/290, you're still gonna have to turn down settings in plenty of games just to play them at *1080p* and 60fps.

humbug said:
Your complaint there is unwarranted.

Plenty of credit is given to the 970, i have said repeatedly they are both great cards and as far as i can see no one is denying that.
Click to expand...
No one? Well you certainly went out of your way to make it sound exactly the opposite, as you were complaining about your 970, and Nvidia cards in general, not being good cards in hindsight. And then CAT rode that wave and was absolutely *insisting* that the 970 was not a great purchase and argued against it. And then you tried to continue that argument on without every clarifying you ever meant anything different. So please dont act like I'm just making all this up out of nowhere.
 
Last edited:
TNA said:
I will be disappointed if that is their best card after this long wait. If that is the case I will definitely be waiting for Volta. Hopefully that is not their best card and a third tier card or something.
Click to expand...

This is most probably the smaller version, as Vega was said to have 16GB memory from the beginning
 
Seanspeed said:
No one? Well you certainly went out of your way to make it sound exactly the opposite, as you were complaining about your 970, and Nvidia cards in general, not being good cards in hindsight. And then CAT rode that wave and was absolutely *insisting* that the 970 was not a great purchase and argued against it. So please dont act like I'm just making all this up out of nowhere.
Click to expand...

Not at all, i think it has some flaws, i also think the 290 has some flaws, i have been vocal on both, thats it, IMO you're raging completely hyperbolic, it seems to me anything negative said about Nvidia, or even just not overly positive is like a red rag to some people, i think that is where you are getting this impression.

I don't hate my GPU, i like it.
So lets just move on eh? :)
 
bru said:
Well there is a wealth of info guesses after the NDA on that event lifted, so lets see what we've got.

8GB ram 'confirmed'
512GB/s memory bandwidth 'either 4 stacks at 500MHz or 2 at 1000MHz'
12.5 TFlop/s FP32
< 300w TDP in server style form (fanless)


Can you get 2G HBM2 stacks? I thought it was either 4G or 8G, which would mean 2 stacks at 1000MHz tying in with above.

NCU ? Is this the GCN replacement ? 'next compute unit'


The opposition for reference

Tesla P40 16GB GDDR5x ram (GP102)
345GB/s
12TFlops FP32
250w TDP in server style form (fanless)


So the first info on Vega is certainly looking good, AMD could have a real winner on their hands.

The only downside I can see is the power usage, which is 20% more than the P40, a little bit higher than I expected considering it is using HBM2 as well.

Overall good first look and info, hopefully we will get a bit more tomorrow with the ZEN event. :)
Click to expand...

There's a whole lot more involved than just comparing a p40 inference accelerator vs amd's vega training accelerator.
 
About the Doom Demo on Vega.

"A Vega 10 has 8 GiByte HBM2 is unoptimized at GTX-1080 level"

A GeForce GTX 1080 overclocked at over 1,900 MHz is there about 10 per cent slower. Note, however, that first, no Vega-optimized driver was used, but simply a Fiji driver with a little additional debugging work.
Click to expand...

https://translate.google.co.uk/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.pcgameshardware.de%2FAMD-Radeon-Grafikkarte-255597%2FSpecials%2FVega-10-HBM2-GTX-1080-1215734%2F&edit-text=

:p
 
FoxEye said:
Ah, "AMD Rebrandeon" again.

But I don't expect this will be true. Small Vega is said to be only 1TF or so faster than the 480. Therefore I personally expect Vega 11 to replace the 480, not coexist with it.

And if that's the case, then £270 for the cheapest small Vega is nonsense.
Click to expand...

Davedree said:
There's a whole lot more involved than just comparing a p40 inference accelerator vs amd's vega training accelerator.
Click to expand...

I'm sorry, what else should we be comparing the AMD Radeon Instinct MI25 with NCU too?


There is a mistake in the previous post as well, the 8GB 'confirmed' was from the card used in the Doom demo not a MI25.
 
humbug said:
Not at all, i think it has some flaws, i also think the 290 has some flaws, i have been vocal on both, thats it
Click to expand...
You're backtracking hard man. Your whole point was that you were disappointed by your 970 purchase. Even after further arguments, you never once stepped in to say, "Oh they're both pretty similar actually, I think it's a great card". Nope. You specifically argued for exactly what I interpreted you to say(which was hardly ambiguous), even.

humbug said:
it seems to me anything negative said about Nvidia, or even just not overly positive is like a red rag to some people, i think that is where you are getting this impression.
Click to expand...
Haha. Seriously, this is where it just gets sad. Everybody on either side always seems to think 'the other side' is the bad one. Can you really not see how there's a *whole lot* of AMD fans here who do the same things or worse? Well, maybe you cant see that, cuz you're one of them.

I dont mind criticisms about Nvidia. But your comment was *totally* making it sound like you thought your 970 was not a good card nowadays and thus wasn't a good purchase back on launch. Actually no, that's not what it 'sounded like', that's what you were saying. You're just trying to be deceptive and act like that wasn't exactly what you said now.
 
bru said:
Well there is a wealth of info guesses after the NDA on that event lifted, so lets see what we've got.

8GB ram 'confirmed'
512GB/s memory bandwidth 'either 4 stacks at 500MHz or 2 at 1000MHz'
12.5 TFlop/s FP32
< 300w TDP in server style form (fanless)


Can you get 2G HBM2 stacks? I thought it was either 4G or 8G, which would mean 2 stacks at 1000MHz tying in with above.

NCU ? Is this the GCN replacement ? 'next compute unit'


The opposition for reference

Tesla P40 16GB GDDR5x ram (GP102)
345GB/s
12TFlops FP32
250w TDP in server style form (fanless)


So the first info on Vega is certainly looking good, AMD could have a real winner on their hands.

The only downside I can see is the power usage, which is 20% more than the P40, a little bit higher than I expected considering it is using HBM2 as well.

Overall good first look and info, hopefully we will get a bit more tomorrow with the ZEN event. :)
Click to expand...

According to pcgameshardware linked I posted earlier the Vega chip in the INSTINCT MI25 and the gaming card previewed is meant to have a 2048 bit memory controller:

https://translate.google.co.uk/tran...als/Vega-10-HBM2-GTX-1080-1215734/&edit-text=

That would indicate two stacks of 1000MHZ HBM2,which would give the same amount of bandwidth as a Fiji based card,but also with two generations more of bandwidth compression.

Also as it is using a tweaked Fiji driver that hints at a 4096 shader configuration too,ie, the smaller of the two Vega chips.
 
Last edited:
LoadsaMoney said:
About the Doom Demo on Vega.



https://translate.google.co.uk/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.pcgameshardware.de%2FAMD-Radeon-Grafikkarte-255597%2FSpecials%2FVega-10-HBM2-GTX-1080-1215734%2F&edit-text=

:p
Click to expand...
That would be awesome, even considering that Doom + Vulkan is an extraordinary case and the average would probably be somewhat less than a GTX1080. Still, anything better than a GTX1070 for *small* Vega would be great.

If this is big Vega, it's not nearly as impressive.
 
CAT-THE-FIFTH said:
From what I read elsewhere,apparently the card might have been downclocking a bit as the exhaust was partially covered by tape.
Click to expand...

It could have been, I suppose that all depends on what sort of cooler it has on it to begin with.

At this stage it would probably be some sort of prototype cooler, which could be anything. Of course the demo was in a small case (looks like a corsair carbide 240) so it cannot of been some 10*140mm fan monstrosity. :)

It also looks like they taped up any place that people could get a look at it through, which makes sense for AMD I suppose, doesn't help us though. :D
 
Status
Not open for further replies.
Back
Top Bottom