Amd Volcanic Islands details

[Ace123]

We know that Hawaii is the flagship chip of the AMD Volcanic Islands series but we have new updates. There are three new names revealed by HWINFO which fall under the AMD Volcanic Islands series ? Tonga, Iceland and Maui. According to 3DCenter, the New Zealand chip which was supposed to make a debut with the HD 7000 series southern islands series but it got Malta instead. The AMD Volcanic Islands series would feature the New Zealand as the flagship dual chip card followed by Hawaii (High-End), Maui (Performance), Iceland (Mainstream) and Tonga (Entry Level).

Read more: http://wccftech.com/codenames-amd-vo...#ixzz2bIpMQfjN

Found a little more information here, also a VERY promising Firestrike benchmark screen shot from a member at Chiphell Forums
http://videocardz.com/45324/possible...nds-hawaii-gpu

According to this fudzilla article, Amd will price their newest top hawaii card below 600, undercutting Nvidia, and if the leaked benchmark is any indication, possibly outperforming them as well
http://fudzilla.com/home/item/32199-amd-hawaii-will-be-priced-below-us-$600

Update: 09/13/13

Here is some more info

"GPU-Z 0.7.3:

Added preliminary support for AMD Radeon R7 240, R7 250, R7 260X, R9 270, R9 270X, R9 290, R9 290X"

Also appears Vesuvius may be their dual GPU card

http://videocardz.com/45569/amd-rade...aming-unveiled

"It seems that someone from HIS made a huge mistake today by posting more information about the future AMD products.

AMD Radeon R9 280X is a successor to Radeon HD 7970"

http://videocardz.com/45608/amd-rade...b-gddr5-memory

[AMDforME]

These cards should be a significant performance bump.

[Lanek]

Im a bit doubtfull ( for stay nice ), about this New Zealand chip..

[Greg83]

I'm suspecting a huge performance jump. Why else this $699, $709 in canada for the 7990, making it below the cost of 2 7970's now. 8 free games. and $70 MIR, so $639 after MIR.

Must be something close to 75-80% the performance average of the 7990 in a single card

[Russian]

I'm suspecting a huge performance jump. Why else this $699, $709 in canada for the 7990, making it below the cost of 2 7970's now. 8 free games. and $70 MIR, so $639 after MIR.

Must be something close to 75-80% the performance average of the 7990 in a single card

I am not sure which card you're talking about, but I think Hawaii will have to be damn close to a 7990 in performance to compete with NV.

[Greg83]

noticing that 780GTX is priced the same after MIR as 7990 now. and 8 free games. i see what ati is doing. moved 7990 vs titan in price to 780gtx in price. or even beating it by $20. even if close to 7990. beating 780GTX at $400 for hawii. or even $500 at start. still a winner.

[Andrew LB]

2 Days ago TSMC said 20nm won't be entering production till Q1 2014. http://www.xbitlabs.com/news/other/d...arly_2014.html

Morris Chang, chief executive officer and chairman of Taiwan Semiconductor Manufacturing Co., said that his company will initiate volume production of chips using 20nm process technology at its Fab 14/phase 5 manufacturing facility in February, 2014. The company will also be producing chips using 20nm fabrication process at its Fab 15/phases 3 and 4 starting from May, 2014.

“On 20nm [process technology], we see little competition. The risk production has started in the first quarter [2013] and the volume production will start in early 2014 next year […] the equipment [is] already being installed, the equipment [is] streaming in and [is] being installed. […] The volume production will start in early 2014,” said Morris Chang, the head of TSMC, during quarterly conference call with financial analysts.

Now that being said by the CEO of TSMC either shows AMD is going to be doing a VERY long paper launch, or these are just rumors with no factual backing.

[Ace123]

added some information and a chiphell screen shot

[DilTech]

I'm suspecting a huge performance jump. Why else this $699, $709 in canada for the 7990, making it below the cost of 2 7970's now. 8 free games. and $70 MIR, so $639 after MIR.

Must be something close to 75-80% the performance average of the 7990 in a single card

I wouldn't expect TOO terribly much as this chip is on the same node as the 7970. There's definitely room for more transistors, but only so much more can be done unless they go massive chip like the titan.

As for the price drop on the 7990, it's because those cards have been selling horridly, and they want to clear stock as fast as humanly possible before the new parts come out.

[paulbagz]

Supposedly beats a Titan with 7816 on 3DMark Ice Extreme.

-PB

[Mumak]

Almost every popular website shared the information about the leaked codenames of Volcanic Islands GPUs. The leak appeared on HWINFO change-log website. All these sites forgot that the same website posted the information about TITAN ULTRA, which turned out to be a complete fake.

wtf? I never said/posted anything about a "TITAN ULTRA"

[informal]

If this thing has 37% more SP units than Tahiti then it will have no problems beating Titan in anything.

[Ace123]

According to this fudzilla article, Amd will price their newest top hawaii card below 600, undercutting Nvidia, and if the leaked benchmark is any indication, possibly outperforming them as well
http://fudzilla.com/home/item/32199-amd-hawaii-will-be-priced-below-us-$600

[tajoh111]

If this thing has 37% more SP units than Tahiti then it will have no problems beating Titan in anything.

37% won't be enough.

Just think of it this way the 7970 and gtx 680 are competitive with each other. One has 1536 cores, one has 2048 cores. Basically they are competitive with each other with AMD having 33% more cores.

Now lets takes your 37% case and extrapolate, Titan has 2688 cores or 75% more cores. If we add 37% more cores on Tahiti xt = 2800 cores. There is only 4% more cores vs the keplar architecture. Considering Nvidia was competitive with 33% less cores earlier, how can you expect AMD to easily beat titan with only 4% more cores?

AMD needs to change the architecture and make each GCN core more efficient and larger. This could turn it into a big chip. Unless AMD makes their GPU over 450mm2, I see them having a hard time beating Titan. I would love for AMD to make big GPU's again, but it seems very uncharacteristic of them as of late.

According to this fudzilla article, Amd will price their newest top hawaii card below 600, undercutting Nvidia, and if the leaked benchmark is any indication, possibly outperforming them as well
http://fudzilla.com/home/item/32199-amd-hawaii-will-be-priced-below-us-$600

If AMD prices this thing at 600 dollars, I expect this thing to be more or less competitive with a gtx 780.

AMD has been pricing their new products high lately. Initial 7970, 7950, 7870 and the 7990 of course. The 7990 was the ballsiest move of all but showed AMD was unrelenting on its pricing, as it should have been priced lower considering the price of their single GPU cards at the time, the lateness of the card and the lack of superiority. None the less it was priced high. Its CPU division has been mad like this too charging 800+ dollars for a fx 9590.

Everything is rumors at this point, but I think if AMD had a titan beater, they would price the card higher than a gtx 780 which is fine in my mind. E.g 749. Titans pricing is ridiculous.

If this card = 549-599.99, I expect this card to perform somewhere along the lines of 10-30 faster depending on the game compared to a 7970 ghz or roughly 20 percent faster overall.

I think the leak(4800 extreme score) is outright false for performance and in the rare possibility it isn't, i.e AMD was able to pull an architecture miracle and make the card 45% faster than a ghz edition, AMD is going to charge a whole lot for this type of jump.

If AMD charged under 600 dollars for that type of jump, it would just be a linear price increase from its current cards($400 for a ghz edition) and AMD or Nvidia haven't ever been this generous for their top card. EVER.

[flopper]

According to this fudzilla article, Amd will price their newest top hawaii card below 600, undercutting Nvidia, and if the leaked benchmark is any indication, possibly outperforming them as well
http://fudzilla.com/home/item/32199-amd-hawaii-will-be-priced-below-us-$600

Has to beat Titan, 2 years for a new serie and wont beat the high end of Nvidia is a fail.
Unless it OC like a mad dog on steroids ofc

[MaddMuppet]

If it comes within 5% of Titan and costs US$600, put me down for 2. Now that Crossfire drivers have been improved substantially with stuttering, it's a viable option. And the best part is the drivers will only get better.

[informal]

@ tajoh111
You are wrong.
Here is how much Titan is faster than 7970GHZ Ed.
2560x1600
GTX Titan 6GB 182%
7970Ghz Ed. 3GB 139%

Difference is 1.31x or 31% (182/139). So if the Hawaii has 37% more SP with removed front end (and other) bottlenecks that were limiting Tahiti, AMD will no problem performing at least 31% better than 7970Ghz Ed.

[SKYMTL]

I think a lot of this comes down to what NVIDIA decides to do with Maxwell. AMD may find themselves competing against a 20nm Maxwell-based GeForce card quite early in the life-cycle of VA.

This could shape up to be an interesting year!

[boxleitnerb]

Nvidia being first on 20 nm? That would be unusual.

[Lanek]

Nvidia being first on 20 nm? That would be unusual.

Maybe, but this have 2 advantages for AMD specially if 20nm dont come as early some imagine.

- They can release their product early of Nvidia. ( ~ 6 months, maybe less )
- They can die shrink to 20nm the architecture current 2014 then.... With enough possible boost on specifications for compet or try be competitive with whatever will be maxwell gpu's out at this time.

[SKYMTL]

Maybe, but this have 2 advantages for AMD specially if 20nm dont come as early some imagine.

- They can release their product early of Nvidia. ( ~ 6 months, maybe less )
- They can die shrink to 20nm the architecture current 2014 then.... With enough possible boost on specifications for compet or try be competitive with whatever will be maxwell gpu's out at this time.

Exactly. One way or another, with product cycles the way they are these days, Volcanic Islands and its eventual 20nm shrink will be AMD's answer to Maxwell, not Kepler.

With that in mind, their "fully enabled" high end single chip card doesn't just need to compete with Titan. It needs to demolish it. Whether or not AMD decides to go all-in at first remains to be seen.

[Alex-Ro]

since so much time has passed after HD7970 i expect nothing less but 30% performance gains in the worst case.

[boxleitnerb]

Maybe, but this have 2 advantages for AMD specially if 20nm dont come as early some imagine.

- They can release their product early of Nvidia. ( ~ 6 months, maybe less )
- They can die shrink to 20nm the architecture current 2014 then.... With enough possible boost on specifications for compet or try be competitive with whatever will be maxwell gpu's out at this time.

That's a matter of perspective. You could also say they are late. Titan was released in February, GTX 780 in May. With Volcanic Islands@28 nm, this round is concluded. Next round is on 20 nm.

Exactly. One way or another, with product cycles the way they are these days, Volcanic Islands and its eventual 20nm shrink will be AMD's answer to Maxwell, not Kepler.

Hm, no. Hawaii is their answer to GK110 (Kepler). The 20 nm shrink will compete with Maxwell...unless Nvidia really releases Maxwell@28 nm first.

[Iconyu]

I think a lot of this comes down to what NVIDIA decides to do with Maxwell. AMD may find themselves competing against a 20nm Maxwell-based GeForce card quite early in the life-cycle of VA.

This could shape up to be an interesting year!

I heard that Maxwell 'large core' won't turn up in H1, but they may repeat what they did with Kepler and release a '104'.

[SKYMTL]

Hm, no. Hawaii is their answer to GK110 (Kepler). The 20 nm shrink will compete with Maxwell...unless Nvidia really releases Maxwell@28 nm first.

Kepler was launched in March 2012. You're telling me VA is meant to compete against an architecture that's already a year and a half old? If that's the case, AMD will be in serious trouble.

At this point, after skipping over their planned GCN refresh, they need to prepare for what's coming rather than what's already been available for a long time.

Think of it this way. If AMD launches a 28nm VA in October, they will likely have a 6-9 month buffer zone before NVIDIA starts talking Maxwell in earnest. In order to compete with Maxwell, they'd need a chip that's (on paper at least) TWICE as powerful as TITAN if NVIDIA's latest slides are any indication of potential performance. Since refreshes typically garner at most 20% better performance than their predecessors (see previous reviews), AMD would have to come out of the gate extremely strong or at least keep something waiting in the wings to respond to Maxwell once VA gets its refresh.

I really have to wonder which is better for AMD though; wait a little while for 20nm (6 months) and compete against Maxwell on slightly more level footing or launch something now that still uses the 28nm node and hope NVIDIA doesn't pull in their launch timeframe due to strong TSMC yields.

The real question is whether VA is a GCN refresh or a brand new architecture. I'm hoping for "refresh" so they can quickly transition to a new 20nm architecture in time to compete against Kepler.