I remember talk here on AT from people who claimed they had inside information, but I don't recall what I would call a "FUD campaign" on the part of AMD.
I thought everyone would just remember.
I remember talk here on AT from people who claimed they had inside information, but I don't recall what I would call a "FUD campaign" on the part of AMD.
I can claim a lot of thing too...doesn't cut it for me. If a serious allegation like financial manipulation, it has to be supported by a something more than a few members claiming to have inside info IMO...
I'm not really sure why this is directed at me... All I'm saying is that I don't remember this supposed FUD campaign Keys is referring to, but I do remember a lot of people posting BS about the 6-series launch. Every time we get close to a launch it seems there is a new member on AT that claims to have information on the upcoming launch. Usually it's BS, and it has nothing to do with AMD or NV directly.
My point is that just because there is a lot of BS going around about something, doesn't mean it's orchestrated by either AMD or NV.
I doubt we're in for a drastic jump in performance. It would be nice if they came out with something affordable that can match a GTX 580 and has 2gb of vram. I highly doubt these new cards will perform much better than that.
That sounds good to me.Considering its a full node jump, you could see a 6970 with a 50% smaller die, so $150ish.
Considering its a full node jump, you could see a 6970 with a 50% smaller die, so $150ish.
I think it will be at least $200, more like $250. AMD loves slapping a premium on their new cards and there is no reason why they want to price it cheaper than a 6870 when it has 6970 perf.
With BD being total crap, this is the one game they must win this round.
Considering its a full node jump, you could see a 6970 with a 50% smaller die, so $150ish.
They switched from VLIW5 to VLIW4 on 69xx, I wouldn't call that a 'little tweak'...The 6000 series was originally set to go onto 32nm until the process got scrapped. So in the end it was a little tweak to the 5000 series on the same process and hence its performance improvements were relatively minor. Improvements are all about process improvements.
Do you really think the 7870 will be that fast? I was thinking between a 6950 and 6970, similar to how the 6870 wound up between the 5850 and 5870.
At any rate, I'm looking forward to the next cards from both vendors, as I want to upgrade from my 6870 and at this time, unless I find a crazy deal on a 6870 somewhere, I prefer to avoid Crossfire.
I don't think it's fair to be so hard on AMD's graphics division because of what happened with Bulldozer. That said, they are one and the same company, and it has tainted my view of them to a certain extent.
You know what, though, I've never found the leaks about AMD's forthcoming products to be accurate. Their best products have been released out of the blue.
It's nice to know that they're supposedly working on 28nm parts, but I refuse to get excited until I read some reviews.
I doubt we're in for a drastic jump in performance. It would be nice if they came out with something affordable that can match a GTX 580 and has 2gb of vram. I highly doubt these new cards will perform much better than that.
He said the same thing about the 6850 and it turned out to be slower than the 5850.
From a purely speculative point, I think there is a high chance one of the top cards released in the first round will be a single-chip-6990 (a.k.a "6990-on-a-stick"). now the question is, is this going to be the aboslute top high-end card or will there be a faster single chip? and of course 6 months later we'll get the "X2" = 7990
He said the same thing about the 6850 and it turned out to be slower than the 5850.
From a purely speculative point, I think there is a high chance one of the top cards released in the first round will be a single-chip-6990 (a.k.a "6990-on-a-stick"). now the question is, is this going to be the aboslute top high-end card or will there be a faster single chip? and of course 6 months later we'll get the "X2" = 7990
Why? 6970 is only around 15% slower than gtx 580 right now. If they just do what they did in going from 3870 to 4870 they'll get ~ 40% improvement.
AMD decided to change the naming scheme from the HD 5000 to the HD 6000 series so it would be more coherent and not as confusing (though others will say otherwise, but objectively it makes more sense now). The dual-GPU card before being another model number higher compared to the others made no sense as it's based on the same GPU. Not only that, but it's much easier now to identify the markets for each card. With the HD 5800 series you had the problem where the 5870 and 5850 were Enthusiast cards while the 5830 was Performance, while the 5970 being a whole number tier higher than the 5870 made it look like it used a different GPU when it simply had two of them. They'll be sticking with the naming scheme introduced with the HD 6000 series moving forward.
Again:
HD *400 and HD *500: Essential (low-end)
HD *600 and HD *700: Mainstream
HD *800: Performance (high-end)
HD *900: Enthusiast (high-end)
You will not see Radeon HD 6990 performance out of any single-GPU card. We're gonna see overall improvements in the 60% range, comparable to going from the Radeon HD 4890 to the Radeon HD 5870. It's unreasonable to think there's gonna be something with that high of a performance. After all, the HD 6990 is to this day the fastest graphics card on the market.
About the HD 8000 series, we don't know anything at all.
Currently the GTX 580 is ~12% faster than the HD 6970. If you mention slower the number is lower, but I forgot how to calculate it. Just sayin'
That's a strawman. 6850 wasn't a true successor to 5850, was it? It was a cut-down 2/3 successor that still managed to only fall ~ 7% short. We all know that 6850 was more than a theoretical 6750, but less than a true successor to 5850. They probably should have called it a 6830 and called the current 6870 a 6850, instead.
I'm interested to see if with 28nm we'll actually get some new true midrange cards, like a full lineup of 77x0 and 78x0.
They've never managed to make such a large jump in the past, what makes you think they'll do it this time? 3870x2>4870, 4870x2>5870, 5970>6970, and TSMC is experiencing (again) major issues in their node jump. If 7970 is within 10% of 6990 then I'll be impressed, especially if it comes out several months before Nvidia's next offering.
TSMC today [October 24th] announced that its 28nm process is in volume production and production wafers have been shipped to customers. TSMC leads the foundry segment to achieve volume production at 28nm node.
TSMCs 28nm process offering includes 28nm High Performance (28HP), 28nm High Performance Low Power (28HPL), 28nm Low Power (28LP), and 28nm High Performance Mobile Computing (28HPM). Among these technology offerings, 28HP, 28HPL and 28LP are all in volume production and 28HPM will be ready for production by the end of this year. The production-version design collateral of 28HPM has been distributed to most mobile computing customers for their product-design use.
The number of customer 28nm production tape outs has more than doubled as compared with that of 40nm. At 28nm, there are currently more than 80 customer product tape-outs. The TSMC 28nm process has surpassed the previous generations production ramps and product yield at the same point in time due to closer and earlier collaboration with customers. TSMCs 28nm design ecosystem is available through its Open Innovation Platform®, with qualified EDA design tools and third-party IP ready for customer designs.
Imho,
I agree with the new naming convention, and if one looks back there was a gaping hole between the performance of the 5770 and 5850 that the 5830 couldn't fill on its own. So, essentially, the 6850 and 6870's were to fill this hole with the newer generation and the higher end cards were simply moved up with the new numbers.
112 is 12% more than 100. 100 is 89.3% of 112% (100/112), so you could say 10.7% slower. It's sort of semantics at that point. When dealing with percentages, you really need to specify exactly what you are saying because there are so many ways express the values. Otherwise you create ambiguity.
Well, I'd still be surprised to see them outperforming a GTX 580 by more than 15%.Why? 6970 is only around 15% slower than gtx 580 right now. If they just do what they did in going from 3870 to 4870 they'll get ~ 40% improvement.