If it can't beat the 5870 by a good margin... then nVidia might as well not bother... only 2-3 people are going to spend the kinda prices they will have to charge for the cards with that kinda performance. Also given the size of the core, the dev. time and the raw specs if it can't beat the 5870 by a good margin nVidia really screwed up.
-
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.
CES: NVIDIA shows Fermi
- Thread starter mathwat
- Start date
Perhaps they are banking on the GPGPU market as well as gaming?
Perhaps the card has been designed for scientist gamers to be able to play crysis on their lunch breaks?
Caporegime
- Joined
- 8 Jul 2003
- Posts
- 30,063
- Location
- In a house
Probably failed to pass like the cases
Probably Lol!
Wonder what the power consumptions like when you start overclocking that little fire breathing dragon to a 5890?
I don't think Nvidia will be bothered if the GF100 is slower than a 5870, Its not like anyone will be held responsible.
Tegra's a mobile chip for crying out loud! The first was a little underwhelming compared to other mobile chips (particularly Snapdragon), but the Tegra 2 chip is a dual ARM A9 part, it'll be an absolute beast performance wise in the mobile world. That chip's designed to compete with the likes of the ARM/PowerVR2 setup in the iPhone, not with things like the Core i7. Although to be fair, ARM A9 is supposed to scale up to 2GHz+ so it'll be interesting to see what comes of this.
Edit: Moreover:
http://www.youtube.com/watch?v=PpGtu_ZkwqA
Last edited:
Lightnix said "Edit: Moreover:
http://www.youtube.com/watch?v=PpGtu_ZkwqA "
But that's not impressive for next generation. We already have that in this generation. If I am not mistaken that's the UE3 engine and the exact same game and same level runs fine now on todays phones.
http://gadgetblips.dailyradar.com/video/unreal-engine-3-on-the-iphone-3gs-ipod-touch-3rd-gen/ (This video looks to be the same level/map to me, only with the brightness turned way up)
http://www.anandtech.com/gadgets/showdoc.aspx?i=3695 (more info and screenshots)
The other thing is Nvidia don't say or show there device. Is it there mobile device running that EU3 engine? Is it there laptop Tegra or some dev none final hardware that's larger then a smartphone and running faster than a smartphone chip? There is just not enough information from Nvidia.
From my point of view Nvidia are showing us behind the times stuff, trying to blag it as next generation while hiding the real details on their mobile GPU. This is just what they did with the Tegra 1.
Lightnix said "Tegra's a mobile chip for crying out loud! The first was a little underwhelming compared to other mobile chips (particularly Snapdragon),"
Why are you saying for crying out loud? Of course it was a mobile chip; I was talking about mobile chips. The first was more then a little underwhelming.
My comment on Quake was that Quake 3 with FSAA has been running fine on mobile chips for years perhaps even 5 years, I have to look back. It's like the 5870 coming out and ATI saying look Quake 3 runs well with FSAA. The other funny thing is Nvidia saying they are the fastest at Quake with zero evidence. When today's mobile devices are running the Unreal 3 engine smooth why bring up something as old as Quake 3 for your next generation device? Saying you are the fastest by not benchmarking other hardware is hardly fair. Could that be as Nvidia's next gen performance hardly beats this gen? Could it be as Nvidias performance is worse then this gen?
Lightnix said " but the Tegra 2 chip is a dual ARM A9 part, it'll be an absolute beast performance wise in the mobile world."
I don't believe it will be. Nvidia said Tegra 2 is around x2 to x3 faster than Tegra 1. That is not absolute beast performance. That means Tegra 2 will next generation just about catch up with this generation. Remember how good Nivida said Tegra 1 would be? It had terrible performance in real life.
Tegra 2 seems to be aimed at current generation device when everyone else is moving on to next generation.
http://www.youtube.com/watch?v=PpGtu_ZkwqA "
But that's not impressive for next generation. We already have that in this generation. If I am not mistaken that's the UE3 engine and the exact same game and same level runs fine now on todays phones.
http://gadgetblips.dailyradar.com/video/unreal-engine-3-on-the-iphone-3gs-ipod-touch-3rd-gen/ (This video looks to be the same level/map to me, only with the brightness turned way up)
http://www.anandtech.com/gadgets/showdoc.aspx?i=3695 (more info and screenshots)
The other thing is Nvidia don't say or show there device. Is it there mobile device running that EU3 engine? Is it there laptop Tegra or some dev none final hardware that's larger then a smartphone and running faster than a smartphone chip? There is just not enough information from Nvidia.
From my point of view Nvidia are showing us behind the times stuff, trying to blag it as next generation while hiding the real details on their mobile GPU. This is just what they did with the Tegra 1.
Lightnix said "Tegra's a mobile chip for crying out loud! The first was a little underwhelming compared to other mobile chips (particularly Snapdragon),"
Why are you saying for crying out loud? Of course it was a mobile chip; I was talking about mobile chips. The first was more then a little underwhelming.
My comment on Quake was that Quake 3 with FSAA has been running fine on mobile chips for years perhaps even 5 years, I have to look back. It's like the 5870 coming out and ATI saying look Quake 3 runs well with FSAA. The other funny thing is Nvidia saying they are the fastest at Quake with zero evidence. When today's mobile devices are running the Unreal 3 engine smooth why bring up something as old as Quake 3 for your next generation device? Saying you are the fastest by not benchmarking other hardware is hardly fair. Could that be as Nvidia's next gen performance hardly beats this gen? Could it be as Nvidias performance is worse then this gen?
Lightnix said " but the Tegra 2 chip is a dual ARM A9 part, it'll be an absolute beast performance wise in the mobile world."
I don't believe it will be. Nvidia said Tegra 2 is around x2 to x3 faster than Tegra 1. That is not absolute beast performance. That means Tegra 2 will next generation just about catch up with this generation. Remember how good Nivida said Tegra 1 would be? It had terrible performance in real life.
Tegra 2 seems to be aimed at current generation device when everyone else is moving on to next generation.
Last edited:
Whats wrong with the quote function?
Anyway. nVidia tout slow, expensive products. People will buy. It is the way it's ment to be played
sweeet! 3D over three screens! finally an excuse for needing so much power to run a console port
Sort of cheating but this is a vid with Crysis on the current gen of iPhone. Would be fun to show freinds at the bar.
http://www.youtube.com/watch?v=IoRRiMDq5r0&feature=player_embedded
"Instead of running Crysis locally on your own machine, you can play Crysis remotely on one of thousands of OnLive servers."
EDIT:To give an idea what level current phones are at this is a vid of Need for Speed Undercover for iPhone. This does work on the hardware and you can buy it along with many other games. A lot of people don't seem to be aware of just how much phones have been catching up with PC's.
http://www.youtube.com/watch?v=1uZ3WkmOfs8
http://www.youtube.com/watch?v=s2B-imDnj94&feature=related
http://www.youtube.com/watch?v=IoRRiMDq5r0&feature=player_embedded
"Instead of running Crysis locally on your own machine, you can play Crysis remotely on one of thousands of OnLive servers."
EDIT:To give an idea what level current phones are at this is a vid of Need for Speed Undercover for iPhone. This does work on the hardware and you can buy it along with many other games. A lot of people don't seem to be aware of just how much phones have been catching up with PC's.
http://www.youtube.com/watch?v=1uZ3WkmOfs8
http://www.youtube.com/watch?v=s2B-imDnj94&feature=related
Last edited:
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
GPGPU sales are a bonus, they are talking up GPGPU features more than before but realistically very very little was added in terms of die space and functions to make it more a gpgpu than a plain old GPU. GPGPU is less than 1% of Nvidia's sales, even they aren't retarded enough to make a card purely based on GPGPU, or marginally, nor are they banking on GPGPU sales to break even, if they were, they'd be dead basically. Despite people thinking its a completely new architecture, its frankly a beefed up gt200b, with a few extras. Its most certainly fundamentally the same core idea's and no major changes in now anything works, just tweaks, as is the 5870 to the 4870, thats not a bad thing itself. They've pretty much spent more on marketing and software developement in the GPGPU market than they have in sales in the last year, but banking on huge market growth, however huge market growth over 5 years and it will still likely be less than 5% of their current sales numbers.
It should beat a 5870, not comftably, and not in every game, in the Nvidia games where they've always excelled I can't see them not retaining a pretty large lead in those games. Likewise in AMD's strongest games I can't really see them being spanked, but price wise I can't see them coming even close to being able to release a slightly lower clocked 5850 type part for under £300.
This is the main problem, though we see lots of 5970's, 295gtx's on OCUK forums, they are very very very low sales parts. The 5850/260gtx/4850/4870 are cards that sold hugely hugely more than the 280/295/4870x2 cards.
We'll have to see, and I think Nvidia still have to see where the new silicon comes in and then what clocks/shader numbers they can release cards at.
The less high end parts they can make and sell at a higher price, the higher the lower end price needs to be to make up for it, unfortunately its highly suspected their yields will be woeful for the top end parts. I can realistically seeing them dropping to somewhere between 256-320 shaders to get a sub £250 part realistically out there, at which point its performance will be questionable compared to a 5850.
The last and maybe most important question is, just how much DX11 stuff are they emulating rather than have hardware acceleration for, everything, or nothing, will DX11 take a huge speed hit compared to AMD. Even Nvidia are pushing now to say how big a difference tesselation is, if they have a huge speed disadvantage for it while it might perform well now, in 6 months it might be no where near as good in newer games.
Anyone remember what the next big tradeshow is and when? Hopefully they'll do a paper launch as soon as they get yield info back on the new version of the cores, then we all know where we stand(roughly) and people can decide more easily to wait or buy now.
lol
What percentage of GPU buyers really end up using a multi monitor setup anyhow? Seems like an odd pitch from both ATI and nVidia, banging on about the number of screens they can support.
I can see it being important in a professional design/financial enviromant though.