As mentioned previously, the 7970 GHz edition was competing vs the 680 for 8 months - far longer than the 7970 was - and also longer than Titan currently has been vs the GHz Edition (currently 7 months).
No I do, but 680 was and still faster than 7970 it's main competitor.7970 GHz was a new sku which was beaten by 680 occasionally and then completely dethroned by Titan/780.Even 770 is faster than GHz at this moment.
Well if you don't factor in chronology, sure Nvidia never loses the performance crown. Think about your argument for a second....
As mentioned previously, the 7970 GHz edition was competing vs the 680 for 8 months - far longer than the 7970 was - and also longer than Titan currently has been vs the GHz Edition (currently 7 months).
Except I am giving both company time to retaliate.9970 can surely beat Titan but will 9970 be the performance champ of 28nm node? I don't think so.
Why not?
The 7970 was "performance champ" for 3 months, then the 680 was for 2 months, then the 7970 GHz Edition was for 8 months, currently Titan has been for 7 months.
Because if 9970 is indeed faster than Titan, a Titan ultra is surely coming out.
And AMD will come out with something that beats Titan UBER FACE MELTING GRiLLER EdITiON. That's how this business works, the 5870 had the performance crown for a really long time (for this industry) the 7970 vs. 680 was basically a wash.Because if 9970 is indeed faster than Titan, a Titan ultra is surely coming out.
And AMD will come out with something that beats Titan UBER FACE MELTING GRiLLER EdITiON. That's how this business works, the 5870 had the performance crown for a really long time (for this industry) the 7970 vs. 680 was basically a wash.
Which is exactly the same argument that could be applied to the GHz edition, yet for some reason the Nvidia fans seem to fondly remember the 680 as the performance champion for all of the 3 months it barely held that position for.
Simply because 7970 didn't set the bar that high so it was relatively easy for NV to take away the crown.
I thought the issue was just a stock clocks thing.
I doubt I'll be upgrading anytime soon as anyways.
I do like to keep up on tech tho.
Looking at the history no they won't.580 was the fastest sgpu on 40nm.
It's laughable to spin this by saying NV had the fastest 40nm GPU, small consolation when you are 8 months late and lose market share in the process. You are convincing yourself with convoluted and convenient arguments, but they really don't make much sense.
Let's get back on topic here. The thread title is amusing, no way AMD beats titan by 2X. 20% would be great, and probably realistic given what I've read about VI so far.
Looking at the history no they won't.580 was the fastest sgpu on 40nm.
The 580 came out against the 6000 series GPU's though. When the 5870 was around, the 480 was its competition, and the 5870 handily beat it. Not only was the 480 very late to production, but it had to be cut down due to heat concerns.
Although thats not to say an Uber-Titan wont be released, as I think it probably will be. The 780 is basically a Titan replacement for all intents and purposes. A Titan-II will be the next tier up.
A 9970 really needs to beat Titan by a fair margin. This would make it more competitive with the Titan-II which I am sure is already well into development. Of course this assume AMD wants to go for the super-high end market at all. Some will say "Of course they want to be the best". But there is limited profitability in that market. The only reason nVidia was able to pull it off is because the R&P was under the HPC side of things, which is extremely profitable. If AMD decides to go for a Titan killer, it means they are also going to try and step up the FirePro's performance and market it at HPC customers.
Oh, that's right. Nvidia had the 480 which was late and still flawed, but they undercut AMD with the 580 by releasing it ahead of schedule.The 580 came out against the 6000 series GPU's though. When the 5870 was around, the 480 was its competition, and the 5870 handily beat it. Not only was the 480 very late to production, but it had to be cut down due to heat concerns.
Although thats not to say an Uber-Titan wont be released, as I think it probably will be. The 780 is basically a Titan replacement for all intents and purposes. A Titan-II will be the next tier up.
A 9970 really needs to beat Titan by a fair margin. This would make it more competitive with the Titan-II which I am sure is already well into development. Of course this assumes AMD wants to go for the super-high end market at all. Some will say "Of course they want to be the best". But there is limited profitability in that market. The only reason nVidia was able to pull it off is because the R&D was under the HPC side of things, which is extremely profitable. If AMD decides to go for a Titan killer, it means they are also going to try and step up the FirePro's performance and market it at HPC customers.
480 was faster than 5870, are you talking about 470?
480 was faster than 5870, are you talking about 470?
Kindly point me in the direction where AMD makes promises about the 9xxx series GPUs.AMD always over promise and under deliver.
The 480 was faster, but it was largely panned on release due to heat and noise issues - while the 5870 was far more efficient. It should also be mentioned that the 5870 was released 7 months prior to the GTX 480.
The GTX 580 was everything the 480 *should* have been, it was a great product. But, I was not impressed by the 480 at release. If you go back and look at reviews, this sentiment was shared among many - it was a hot and loud piece of junk. That is why many folks stayed with the 5870 and/or preferred it over the 480 despite the slight speed deficiency (besides the fact that it was released much sooner than the 480).
I really feel like Fermi is why nvidia pushed so hard for efficiency with the Kepler. That was the smart business move by nvidia, nobody wants a dust buster for a GPU. Seems like the roles are reversed these days. AMD/ATI were always the more efficient uarch prior to Fermi. Again, nvidia made the right move by focusing on efficiency with the Kepler; I don't care about power consumption but I do care about the indirect effects it has on heat and noise.
I do think the 9970 will match or come close to Titan - but the most important thing in my mind is for AMD to ensure the software is up to snuff and creating an improved reference cooler. AMD used the same ref cooler from the 5870 to the 7970, it really is time to junk that reference cooler. It is long in the tooth to say the least. Really, those were the biggest problems with the 7970. It was a good product in it's own right but it took f-o-r-e-v-e-r for the software/drivers to catch up to it. That cannot happen again IMO.
Where did you draw that conclusion from? Unless this is a steering attempt?