Does anyone know ANYTHING solid about Cayman atm? :shrug:
Printable View
Does anyone know ANYTHING solid about Cayman atm? :shrug:
From what's been leaked so far (pending if it's true and translate into real performance) the 6950 may compete well with the 580. But we have to wait and see. Other then that there is no real solid information other then what was leaked in the slides (not the separately leaked fake ones).
I think beside manufacturing problem of the card (the hard to find MOS from TI), other probable main reason for the "delay" is AMD trying to squeeze as much performance out of this chip in the driver. Clock i think is already finalized, but getting optimal output out of a "new" mArch from the driver is plenty difficult IMHO, and with this chip, it seems AMD graphic really want to deliver a pretty serious blow against the green camp. The VLIW compiler in the driver has to be reworked seriously to make sure the promised added efficiency going from 5 way-VLIW to 4-way VLIW really materialized in MOST cases.
I was thinking about the new dual core (Graphics engines) and the 4-way VLIW architecture in Cayman,
Starting with the front end, the ILP (Instruction Level Parallelism) must be the same or better than Barts (6800 series) and the new dual core design seems better in TLP (Thread Level Parallelism). Doubling the Graphics engines in order to have a second Tessellator unit will increase the transistor count and that means it will take more space. So the new front end will be bigger than Barts in size.
Because the new 4-way VLIW don’t have a specialized T (Transcendental) unit, its job is done by the 3 out of 4 shaders insight the SP in the new VEC-4. I believe that when they calculate T(sin, cos etc) the new 4-way VLIW will be slower that the previous 5-way VLIW by up to 10-20%. FP performance could be almost the same as VEC-5 (5 VLIW) per SP but at some instances it could be slower by up to 10%. The new VEC-4 saves them 10% space per SP so in theory they could put 10% more SPs and keep the same size.
One more problem with size will be the Texture units. Because Tex Units are part of the SIMD, and we will have more SIMDs than Cypress (20) they will increase the size of the die. I have no idea how big the Tex Units are but if Cayman have 30 SIMDs (120 TMUs) then they will increase the die size a lot.
Another problem will be the need to connect all those SIMDs between them and that’s the job for the Crossbar. If we have 30 SIMDs and 1920 shaders, Cayman has 20% more shaders than Cypress and it will need more connecting lines and bigger crossbar. That will increase the complexity of the chip and will increase the size.
One more side effect of this complexity is that it could effect the yields because more broken lines could happen in manufacturing and more copper could effect the thermal characteristics of the chip.
From my point of view 120 TMUs are too much and if AMD wants to keep the die size close to Cypress then they will need to cut the SIMD count to 24 SIMDs and 96 TMUs. If the new 4-way VLIW have the same performance with the old VEC-5 and they will increase the ILP and TLP in the front end, then with 24 SIMDs and a dual Core architecture for 2x Tessellation performance, they will have a small chip at 360-380mm2 with 20% more performance in DX-9-10 but 2x in DX-11 Tessellation.
60 page of nothingness seriously wth amd
Honestly AMD, give us something. You delayed the 22nd which was a blunder because if you didn't know about the gtx580 how dumb are you? It is exactly what we all thought it was going to be. 512 version of 480 + minor tweaks = 15% faster. I would be happy with a paper launch on the 22nd because at least you can keep prospective buyers holding out for a few more weeks. AMD + ATi have always been 2nd place when it comes to marketing. Frankly if there was another player in the industry I'm sure AMD/ATi would be third when it comes to promoting your products.
I understand that misleading the public with fake information during the 4870 launch fooled Nvidia and served the purpose of making them price their cards much above their performance - and they looked like a bunch of dumb people in the process. So it was a good move by AMD.
But with this launch - 580 has already been launched, its price set, and is a good card. If AMD's trying to pull another 4870 (meaning they have a killer performing, killer priced card) what would serve them better THIS TIME is to leak true information about their cards. The reason being: 580 is a good card and you want to prevent people from buying it at least until your own product hits the shelves. This fact plus the delayed launch date, plus with the facts that this is the biggest chip AMD produced to date and is a (although not completely) new architecture, all being possible contributors to a certain GPU's failure; makes me think these cards will not be as good as people think.
Well the 580 is hardly in stock in most countries, so I don't think AMD is too worried about sales right now
No, I don't see it as anything since like you, I have no real info regarding this card. I'm just saying that everyone's expectations are too high, but to me the circumstances indicate otherwise. I'd be very pleasantly surprised if 6950 competed with 580 and 6970 blew everything out of the water (I've sold my 5970 and I have been waiting for a fast single GPU card).
you're forgetting Nvidia still hasn't launched the gtx 570 yet and when it comes to volume the 6950 will sell more than the 6970. So Nvidia can still adjust their clocks on the 570 to make it more competitive than the 6950 and that is a risk for AMD. Untill all of Nvidias products are out, including the fabled 595, AMD does have a lot to lose by leaking benches and facts.
Why does it make you think that?Quote:
But with this launch - 580 has already been launched, its price set, and is a good card. If AMD's trying to pull another 4870 (meaning they have a killer performing, killer priced card) what would serve them better THIS TIME is to leak true information about their cards. The reason being: 580 is a good card and you want to prevent people from buying it at least until your own product hits the shelves. This fact plus the delayed launch date, plus with the facts that this is the biggest chip AMD produced to date and is a (although not completely) new architecture, all being possible contributors to a certain GPU's failure; makes me think these cards will not be as good as people think.
Today 10:35 AM
A 2+ week delay to make sure your product is as competitive as possible is a bad thing? If AMD launched a card that they could have gotten more out of would have an effect on sales. Folks like you would complain that "AMD shouldn't have released it; a short delay wouldn't have hurt."
Before you say that nobody would say that, we've all seen it plenty of times before. Lets not forget how long nVidia delayed. Now you're acting like they're right on top of the market. But obviously they were doing the right thing.
I see this as no different. AMD wants to make sure their offering is tip top before release. That is good business. Short selling your customers just to get it out the door is bad business.
+1
if its really good, then amds pr = fail...
if anything pr says COULD be a hint to nvidia and MIGHT hurt amd, then why does amd have pr at all? seriously amd, you guys might as well close the pr department all together... cause nvidia can ALWAYS lower prices, launch a tweaked card, launch a dual gpu card, tweak the driver...
your argument makes no sense, if nvidia could indeed easily do all those things, it would actually benefit AMD much more to keep them in the dark about upcoming products. Nvidia cant do those things nearly as easy as you claim but if indeed they could, how is it helpful to leak performance numbers of an unreleased product?
The more nvidia can react, the more AMD needs to keep them in the dark.
You know the lanch date and you have to wait if you want comparisons anyway, in two weeks they can still improve drivers and xfire support and what not, and its not because there is a tech comunity crying out for benchmarks that they need to comply.
In the end, the GTX580 is not really doing as much damage to AMDs sales as people seem to believe, and it is in short supply. I see no reason at all why AMD needs to telegraph their competitor what they will be releasing in 2 weeks, especially since there is still a 570 to come.
Well when you consider how how many on the red team were talking early on about how the green team was simply going to paper launch 580 to ruin the 59xx series launch it seems the same basic principle would still apply, to take the attention away from the competition.
Competing brands have to sling promotional pr at the public to keep relevant in the consumers mind and win their business.
AMD should be pitching some marketing pr with the new product since 580 is Nvidia's mid life kicker until the next node, it would be very odd for the green team to pull out anything else until then so the cards are on the table already.
Even if though the launch has been delayed I don't see any reason why they would not want to start promoting their new flagship product by priming the market.
Is it me or are some post missing? Lets all get along.:D
Heh.
AMD is not loosing anything by keeping this as tightly guarded as they have till now.
Even if its spanking chip with performance, there is little to gain from leaking the info month or so before launch.
Sure 580 is out, but its not available in any good quantity anywhere to worry AMD.
Second, they are in a position to hurt the Nvidia right before holiday season or / and right after it. Nvidia will have to go to troubles of lowering the price, compensate the AIBs, who will need to compensate the etailers and retailers. Its a lot of work. So why should they leak any info and give their competition time to prepare their strategy and action right now.
There was nothing bad to hide on last 2 major product launches by ATI/AMD. It was all good news, still they managed to keep everything silent right until the launch day to drop a nice surprise for customers and their competition.
How can this be fail?
we are all talking nonsense though, just making up numbers and hoping then complaining then making up new numbers, etc.
secrets are not always good, especially when there is another card out on the market already. it is the season for buying and people want to know what their options are.