• 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.

AMD challenges Nvidia

bhavv said:
Well .....



So Nvidia are a little bit faster. But the problem with this card is that is is held too far back by its lousy 600 Mhz clock speed, and you cant even get them up to 700 Mhz without them blowing up :x

The G110 seems to work at its best around 850 Mhz, the GTX 590 is simply bottlenecked too much by its low frequency. The GPU is actually a lot more advanced that ATI's current ones are, but if they cant get a decent clock speed on them, then it doesnt matter, they are too bottlenecked at 600 Mhz.
Click to expand...

I don't think you should be considering lousy resolutions for a £600 card to determine which is better.

People are obviously only going to be buying these cards for high resolution/multi monitor setups. ie 2560x1440 or higher or multiple 1080p screens.

That graph you shows considers resolutions like 1280x720 which these cards ARE NOT going to be used for. The only reason nVidia reaches 103% on that graph is because it's faster at lower resolutions.
 
I think its too hard to swallow for some people that AMD might have the fastest currently available graphics card on the market.

However, If you look at this objectively it's clear the 6990 is a faster GPU, sadly that wont wash for some, hence why they will compromise on it being draw.
 
I would say it is a tie with a very slight edge in favour of 6990.

I am glad it is a tie overall otherwise it would have upset lots of people on both sides if one card was say 25% faster than the other and vice-versa lol
 
I think it will come more down to minimums, averages and other points like smoothness and performance hit when 4xAA or more applied, rather than one card attaining 150fps vs the other falling 'short' at 135. The maximum frame rate means very little in most situations.
 
Last edited:
NZXT30 said:
did i miss something
Click to expand...
Common knowledge by now

Nvidia statements about the fault:
'rest assured that the GTX 590 operates reliably at default voltages.'

‘Nvidia has worked with several water-cooling companies to develop waterblocks for GTX 590,'

'and these solutions will help provide additional margin for overclocking, but even in this case we recommend enthusiasts stay within 12.5-25mV of the default voltage in order to minimise risk.’

They have also came out and said that overclocking and overvolting invalidates the warranty since the fault was found.

WingZero30 said:
I would say it is a tie with a very slight edge in favour of 6990.
Click to expand...
How can it possibly be a tie with a very slight edge, your contradicting yourself there.

Gold, Silver, and Bronze are awarded in competitions, everybody doesn't get Gold medals because it's a draw with a slight edge.
 
tommybhoy said:
How can it possibly be a tie with a very slight edge, your contradicting yourself there.

Gold, Silver, and Bronze are awarded in competitions, everybody doesn't get Gold medals because it's a draw with a slight edge.
Click to expand...

What I meant was that generally 6990 is better in some games and 590 is better in the others.

However specifically in 3d benchmarks 6990 gives slightly higher FPS than 590 in most of those programs.
 
I see the blogger at AMD didn't link to Toms Hardware or Bi-Tech lol but he does have a point as it would be nice to see how Nvidia qualifies there claim to the fastest video card.
 
I, would say its a win for amd

From what I read they are similar overall but where amdwin is important

Games like metro and crysis1 and at the 2560 resolution is where it pulls ahead ...who needs more fps when both cards are pushing 100 anyway
 
Cythx said:
I don't think you should be considering lousy resolutions for a £600 card to determine which is better.

People are obviously only going to be buying these cards for high resolution/multi monitor setups. ie 2560x1440 or higher or multiple 1080p screens.

That graph you shows considers resolutions like 1280x720 which these cards ARE NOT going to be used for. The only reason nVidia reaches 103% on that graph is because it's faster at lower resolutions.
Click to expand...

It was the 6990 @ 103% not the GTX590... also some people will use them for lower resolutions i.e. I would (if I hadn't already got GTX470 SLI) because I use a 1680x1050 120Hz panel and like to keep as close to 120+fps as possible.

arc@css said:
I think it will come more down to minimums, averages and other points like smoothness and performance hit when 4xAA or more applied, rather than one card attaining 150fps vs the other falling 'short' at 135. The maximum frame rate means very little in most situations.
Click to expand...

Agreed, unfortunatly hard to get some decent statistics on this as no one does proper decent frametime graphs - indications are that the 590 has a bit more consistantly framerates but even from the few sites that do have mins and avgs its hard to draw a decent conclusion without knowing i.e. how long each spent below 30fps and so on, couple of sites like ***** do have graphs but their way of producing them and their overall testing methodology makes it impossible to draw anything useful from them (you need graphs something like this for multi GPU comparisions: http://aten-hosted.com/images/fpsexample.jpg ).

TerryH said:
I think its too hard to swallow for some people that AMD might have the fastest currently available graphics card on the market.

However, If you look at this objectively it's clear the 6990 is a faster GPU, sadly that wont wash for some, hence why they will compromise on it being draw.
Click to expand...

If you look at it objectively neither is clearly the faster GPU... a range of benchmarks would tend to indicate the 6990 has a small overall lead but so small that for every day useage its negligible unless you happen to play one of the games where one or other of the cards is much faster than the other.
 
Last edited:
Rroff said:
Agreed, unfortunatly hard to get some decent statistics on this as no one does proper decent frametime graphs - indications are that the 590 has a bit more consistantly framerates but even from the few sites that do have mins and avgs its hard to draw a decent conclusion without knowing i.e. how long each spent below 30fps and so on, couple of sites like ***** do have graphs but their way of producing them and their overall testing methodology makes it impossible to draw anything useful from them (you need graphs something like this for multi GPU comparisions: http://aten-hosted.com/images/fpsexample.jpg
Click to expand...

Yeah, I see what you mean. I don't have the technical experience but I can see the logic in how time is relevant to the experience generated with regard to how smooth feedback is. :)
 
Yeah if I'm looking at buying a multi GPU setup and comparing different options what I want to know is, how smooth (microstutter, etc.) each one is, how much time they spend below 30fps and how much time they spend at or above 60fps - unfortunatly very few sources tell you this.

OK if the average framerate is above 60 and the min is decently high its a reasonable indication, but doesn't really tell you the whole story.
 
Overall this round is a win to AMD but only due to the 590 melting issues. The 6990 is far too loud.

But for any side to try and claim the performance crown is just a joke as clearly it's too close to call. Claim a win in a specific title fine, both sides can do that.

And can we please cut out the garbage 'fanboy' insults, it's this kind of trash that has brought this forum to it's knees recently.
 
this is just daft, AMD say they are faster just because they quoted an industry standard benchmark. lol OK so nvidia just have to show one like this then they can say they are faster in an industry standard benchmark.

bottom line the two cards are very close and each win in some and lose in other situations.


haven.jpg
[/URL][/IMG]



tommybhoy said:
They have also came out and said that overclocking and overvolting invalidates the warranty since the fault was found.
Click to expand...

very interesting, do you have a source for this, as the last i read was that was the cards were overclocking fine and it was covered by the warranty
 
Last edited:
Tesselation is useless just now though. :\

Unless you like staring at a benchmark...

And AMD managed to get the same performance on a smaller die here. NV need to stop using huge dies, keplar/better bridge will be interesting.
 
tommybhoy said:
Common knowledge by now

Nvidia statements about the fault:
'rest assured that the GTX 590 operates reliably at default voltages.'

‘Nvidia has worked with several water-cooling companies to develop waterblocks for GTX 590,'

'and these solutions will help provide additional margin for overclocking, but even in this case we recommend enthusiasts stay within 12.5-25mV of the default voltage in order to minimise risk.’

They have also came out and said that overclocking and overvolting invalidates the warranty since the fault was found.


How can it possibly be a tie with a very slight edge, your contradicting yourself there.

Gold, Silver, and Bronze are awarded in competitions, everybody doesn't get Gold medals because it's a draw with a slight edge.
Click to expand...

how would they know that you overclocked your card? if the card burns you rma it and there is no way any company would ever deny you a replacement.
 
Last edited:
Nope, the "new" Nvidia' solution is driver based, the OVP/power throttling, some is in the bios, like throttling at 100C or the like, the things for Furmark and extreme usage are based in the driver and are mostly there to protect the VRM's. I have zero problem with this, I've been telling people they're idiots for testing Furmark/kombuster for ANY reason, there is none, stable in Furmark doesn't indicate your stable overclock in any other application or game. People are dead set on having their card survive Furmark so will disregard a completely stable 100Mhz faster overclock, much lower temps, quieter operation just because it can only not pass Furmark. Its a literally worthless piece of software. Its the equivilent of downloading a virus ON PURPOSE for no other reason than to see if your AV can handle it, and the same virus over and over, with zero benefit to you if the virus gets caught and deleted, or trashes your data.

The biggest problem is, if the VRM's go just like that with a not even close to extreme voltage, how long will they last at the already very high temp default clocks, apparently some of the review ones that died WERE at stock.

Theres a reason the VRM's don't die on 6990's and overclocked they still don't die, Nvidia seems to have cut the margins insanely close on max wattage the card can use and the power componentry used to supply that. Worse still is its not even close to the max the card can do, no ones pushed Furmark full whack, 6990's can handle furmark at OC settings and then overvolted and overclocked further, benchmarks(almost all) and basically every game gets no where near that kind of load on the VRM's. So it can handle the extreme use and normal use is no where near.

Nvidia are having trouble under normal use, I wouldn't be surprised if they would fry easily under Furmark, and if they are dangerously close to the edge just how long will they last.


Simple fact is the 6990 IS faster, and techpowerup is still HORRIBLE for comparing AMD/Nvidia, they've removed the best games, or gimped xfire on them, Call of Juarez, if they managed to get it working right, (as they had over a year ago, but not since) that would be another game showing a massive lead, and extending the gap. If you remove the CPU limited games, the lead would extend, if you just ignore the low res situations which you absolutely should do, the lead would extend, if you remove RTS's(because there really isn't one out there that isn't based on CPU performance) the lead extends.


The problem is, AMD didn't release the 6970 and claim it to be the worlds fastest graphics card. AMD did it for the 5870, accurate, the 5970 accurate, the 6990, accurate.

Nvidia did it for the 480gtx, inaccurate, fastest gpu, fastest card, not close(5970 was long since out). 580gtx, inaccurate, still the 5970, by a still large gap, 590gtx, inaccurate, not the fastest gpu OR card.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom