• 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 "Greenland" Vega10 Silicon Features 4096 Stream Processors?

Status
Not open for further replies.
Man of Honour
Joined
21 May 2012
Posts
31,980
Location
Dalek flagship
ZBmIZ7K.jpg


http://www.techpowerup.com/221269/amd-greenland-vega10-silicon-features-4096-stream-processors.html
 
So Polaris isn't going to be what their "big" card will be on, or am I reading it wrong? Sounds like Vega is timed to release with big Pascal from that.
Right. Vega will be their big, HBM2-equipped card for next year. Polaris 10 the mid/upper mid range card and Polaris 11 the smaller, budget/power-minded card.
 
Polaris will more than likely just support GDDR5 from the commits in the open driver. it only talks about 32bit per memory channel. with 8 in total.

GDDR5X uses 64bit memory channels.
 
The real question is why are AMD abandoning HBM on this new card as it will be faster than a Fury X ?

http://www.anandtech.com/show/8526/nvidia-geforce-gtx-980-review

Why did the 680 have a smaller memory bus than the 580, why did the 980 have a smaller memory bus than the 780ti yet both faster than the predecessor? Who says the new midrange card will be faster than Fury X?

Where is AMD abandoning HBM, it's going to be on their and Nvidia's high end cards. It's expensive memory still, midrange cards have less need to push for 4k performance and midrange cards are cheaper and smaller meaning they have less space for a memory controller than a card over double the size, they have to fit into a cheaper price bracket. On top of all that, every generation brings efficiency improvements. Just because X shaders need Y bandwidth in one generation doesn't mean they do in the next generation.

Historically almost every single midrange card has had similar/higher performance to the previous gen midrange card yet less bandwidth and a cheaper narrower memory bus.

But something that has been completely standard for the past 15 years of new generations of graphics cards for AMD and Nvidia you decided to phrase as AMD abandoning HBM... sure, you have no agenda, no bias.
 
Why did the 680 have a smaller memory bus than the 580 .... yet both faster than the predecessor?
That'll be because the 680 was based on a midrange chip that performed very nicely.

The 580 was a "full fat" 2nd-gen Fermi, just as the 480 before it was a "full fat" 1st-gen Fermi and the 280 before was a "full fat" Tesla.

The 680 wasn't a "full fat" Kepler, no, that honour went to the Titan (and later, in castrated form, as the 780 and 780Ti).

As that worked so well for NVidia (charge full-fat prices for a midrange chip), it's no surprise they repeated it with the 980/Titan X/980Ti.
 
Last edited:
That'll be because the 680 was based on a midrange chip that performed very nicely.

The 580 was a "full fat" 2nd-gen Fermi, just as the 480 before it was a "full fat" 1st-gen Fermi and the 280 before was a "full fat" Tesla.

The 680 wasn't a "full fat" Kepler, no, that honour went to the Titan (and later, in castrated form, as the 780 and 780Ti).

As that worked so well for NVidia (charge full-fat prices for a midrange chip), it's no surprise they repeated it with the 980/Titan X/980Ti.

That's the point, Polaris is dramatically smaller chip. Fury X is 596mm^2, it looks likely the biggest Polaris is 232mm^2... it is precisely what the 680gtx was to the 580gtx, or the 980 was to the 780ti.

Vega 10 will likely have HBM2 and is coming next year, the chips coming this year that according to Kaap have "abandoned HBM" are dramatically smaller midrange chips.
 
I bet they will.be faster than a 290x tho. Which would be nice

Ellesmere XT should exceed Hawaii/Grenada in terms of performance and match if not exceed fiji.

I just hope their hype of bringing performance/cost down is true and these end up for around 250-300 price bracket. And they could quite easily do that considering the smaller PCB's, fewer parts in power phase, fewer GDDR Dies and with the node shrink.

The GDDR5 memory should be 1GB chips made on 20nm process so should be cooler and more efficient than older ones.

will be very surprised if they go with the inductor only power phase method such as with the Radeon Pro Duo, completely removes coil whine.
 
The real question is why are AMD abandoning HBM on this new card as it will be faster than a Fury X ?

They don't want to use HBM 1.0 with the 4gb limit on Polaris.
Vega will be HBM 2.0 equipped.

Samsung will apparently be first to market with HBM 2.0 with Hynix slightly behind them.
Nvidia will snap up a a rather large chunk if not all of the first Samsung IC

No GDDR5x available yet either, they pretty much had to go with GDDR5 for Polaris or delay even further.
Hobson's choice!

Nvidia in exactly the same position.
 
Last edited:
They don't want to use HBM 1.0 with the 4gb limit on Polaris.
Vega will be HBM 2.0 equipped.

Samsung will apparently be first to market with HBM 2.0 with Hynix slightly behind them.
Nvidia will snap up a a rather large chunk if not all of the first Samsung IC

No GDDR5x available yet either, they pretty much had to go with GDDR5 for Polaris or delay even further.
Hobson's choice!

Nvidia in exactly the same position.

An article I just read thinks the majority of Samsungs first HBM2 batch will be heading the way of the mobile market. Amd have a priority agreement with Hynix who begin mass production in q3 which I think puts them in a stronger position as far as supply goes.

Either way I doubt we will be seeing any HBM2 cards until the end of the year start of 2017.
 
Last edited:
So my theory of Vega being the big-chip this gen, with Polaris being the mid-range isn't complete nonsense after all... that's if this article is any credible. In which case, move over Polaris, I'm excited for Vega now XD
 
I was saying this for ages. No "big" cards this year they will be in 2017 which is Vega. Polaris which will be low and mid-high end this year. Why people was even thinking anything from nVidia and AMD will be the big daddy or big guns must have been having a laugh.

Now the high end cards from AMD or nVidia could quite possibly match or exceed Furyx and 980Ti in terms of performance but what will be most noticeable is the lower power usage.
 
I don't mind if real deal high end comes next year. Gives me more use of my Nano. No HBM 2.0 cards and no temptation to swap and can save money lol.

I fancy a move to the highest end Kaby Lake CPU when they drop, so that may be only upgrade this year :eek:
 
I was saying this for ages. No "big" cards this year they will be in 2017 which is Vega.

Nothing points to them being next year, in the slide Vega was shown hovering over 2017 in the timeline, so they could come late this year or early next year.

I fancy a move to the highest end Kaby Lake CPU when they drop, so that may be only upgrade this year :eek:

Might be worth waiting to see how the 8 core zen comes up, would be a better upgrade if the 8 core has the same price as a top end, but not extreme edition, i7.
 
They don't want to use HBM 1.0 with the 4gb limit on Polaris.
Vega will be HBM 2.0 equipped.

Samsung will apparently be first to market with HBM 2.0 with Hynix slightly behind them.
Nvidia will snap up a a rather large chunk if not all of the first Samsung IC

No GDDR5x available yet either, they pretty much had to go with GDDR5 for Polaris or delay even further.
Hobson's choice!

Nvidia in exactly the same position.

This is the whole point

AMD have realised this even if DM can not accept it.
 
Status
Not open for further replies.
Back
Top Bottom