AMD allegedly pushes Vega GPU forward to October

Don't hold your breath for any AMD new products lol....sometimes they end up being a year later than the rumer
 
This adds weight to the idea that Polaris is ~390 class speed. If Polaris was faster, AMD wouldn't be panicking like this.

Agreed. This is a response to the 1080.

I'm sure pricing on Polaris is being re-evaluated as well. They need to be competitive, and if they can't come close with Polaris to even a 1070, then they have to be a cheaper alternative.
 
Were you, by any chance, one of those people claiming Polaris will compete with GP104 despite all information pointing to it competing in another tier entirely?
That would explain why you're bitter, just not why you're taking it out on me. Get mad at Raja Koduri, or yourself for having unrealistic expectations...

I have to add though, when the conversation is about AMD products, talking about the NV products they compete with is very pertinent to pricing...

Why don't anyone find any quotes regarding performance of Polaris from me ;) . But shoving me inside the fanboy box is rather pathetic. Do yourself a favour stop attacking people and to claim :
The people beating their chests over Polaris will finally shut up and move on to beating their chests over Vega, lol.

What is this nonsense ? Who are you to tell people to shut up on this forum ?
 
This could be great, if a truly unleashed Vega with a full compliment (8? 12? 16gb?) of HBM2 is going to be made available this year, that would potentially hobble Nvidia well into next year. If Polaris 10 can provide 980Ti performance for the $200-400 price range through the summer, then Vega comes out and crushes the high end with HBM2 and all new architecture made for a high end card, that would put AMD way ahead. Nvidia is banking on another $700+ card (not to mention Founders Edition nonsense for even more money and no more performance it seems) with GDDR5X and GDDR5 1070 that is as of yet the only unknown - if Polaris can provide better price performance than the 1070, they'll lock up that entire price range. Maybe this will force Nvidia to have to roll out their Ti version of the 1080, which still may not touch HBM2, or try to speed up their HBM2 card roll out instead, but in the meantime AMD will capture a ton of sales and maybe make developers less likely to partner with Nvidia's proprietary GameWorks, PhysX and other junk in favor of using open standards and APIs that AMD is favoring.

They need to announce things soon, though...
 
Agreed. This is a response to the 1080.

I'm sure pricing on Polaris is being re-evaluated as well. They need to be competitive, and if they can't come close with Polaris to even a 1070, then they have to be a cheaper alternative.
Isnt polaris supposed to be 100nm smaller, so it would/should be cheaper by default.
 
I think to many here are jumping to conclusions or just hopeful thinking. I wonder if AMD will have their own AMD cards (since they produce the reference design anyways) like ATI use to years past. Since Nvidia appears to be going into the card selling market with a premium price added to it for a reference card, maybe AMD will to to get the series out faster. Who knows besides AMD.

I will just wait until the dust is settled, wait, with gpus it is never settled.
 
Not likely, but watching the Nvidia nerd rage from all the 1080 purchasers would be interesting.

I don't think so. I think the true brand loyalists are blind much like Apple folks. They will buy it regardless. I really hope this thing is competitive to drive prices down for the older cards or like the old days where price drops were a thing. I personally think AMD needs to just push SOMETHING out the door even if it's in extremely limited supply.
 
Isnt polaris supposed to be 100nm smaller, so it would/should be cheaper by default.

The prevalent association of die size and costs is a gross oversimplification of all the factors involved.

Even on the same process from the same fab die size alone is not even the only factor in determining effective yields. In this case we also have much more complications involved now with different fabs and processes. Add to this that yields itself are not the sole determinant of cost either.

Basically one known variable is really insufficient data to make any accurate determinations.
 
I remember waiting in January of last year for Fury X. Ended up being six ish months later. I don't bother waiting for AMD or even reading rumors.

AMD isn't that bad off performance wise, they are just always waayyyyy late to market.

If they could get out big Vega before big Pascal it'd be a discriminator. Personally I'd still buy nVidia. AMD isn't even on the radar except for relative's rigs, AMD doesn't support features I need.

Let's look at the rumors about Fury X. Date was off by six months, performance was off - even paper launched by AMD was way off and misleading, and the whole "over clockers" dream. What a joke. Also ran hot as balls.

So not sure why we bother with these threads... besides for funsies. No actual decisions should be made off of them.
 
Last edited:
I remember waiting in January of last year for Fury X. Ended up being six ish months later. I don't bother waiting for AMD or even reading rumors.

AMD isn't that bad off performance wise, they are just always waayyyyy late to market.

If they could get out big Vega before big Pascal it'd be a discriminator. Personally I'd still buy nVidia. AMD isn't even on the radar except for relative's rigs, AMD doesn't support features I need.

Let's look at the rumors about Fury X. Date was off by six months, performance was off - even paper launched by AMD was way off and misleading, and the whole "over clockers" dream. What a joke. Also ran hot as balls.

So not sure why we bother with these threads... besides for funsies. No actual decisions should be made off of them.
We bother with these threads because the alternative is to stand around with our dicks in our hands and not talk about Jack Shit. That's the internet brah...
 
Back
Top