Do a comparison for interest - no problem at all, but the 480 must be judged against the 5870 as it stands today.
-
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.
1st Fermi review
- Thread starter Azza
- Start date
Do a comparison for interest - no problem at all, but the 480 must be judged against the 5870 as it stands today.
Haven't they had working versions of these cards for awhile now? Sure not in any massive quanitities but surely enough to test drivers internally. Being late to the party, won't they have wanted their driver team to squeeze as much out of these cards as possbile so as not to loose face?
I'm sure we will see driver improvements, but who is to say they haven't already seen significant gains via drivers under internal testing? Who is to say the ATI cards wont see a few more frames per second squeezed out with new drivers?
We don't know whats going to happen.
Maybe you will compare it with today's drivers so you can claim a small win (or perhaps minimize a loss) but I, and I'm sure others, will be interested to compare the cards with like-for-like drivers to get a more balanced and accurate view of performance.
How much faster? Is the difference like in between 5850 and 5870 or more? Does it justify £100-150 cost difference as well?
Soldato
nVidia have has working hardware for around 6 months, they have been optimising the drivers.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
Those graphs don't mention what voltage they used to overclock that far, which basically make the chart useless, add .3v onto every chip and they'd go up dramatically also. Thing is the i7 really doesn't appear to need that much voltage to do 4Ghz, if they simply used loads extra or had one very bad chip it makes the results pointless.
As for being a great gpgpu, I really wish people would stop saying that, AMD aren't actively marketing it to gaming companies as a great GPGPU, but the 5870 certainly is.
The benchmark the other day, of a company that uses Nvidia hardware and CUDA to run their data, they did a quick port, got the same software working on AMD hardware in OpenCL(and didn't take them at all long apparently) and a 5870 was, was it 4 times faster than a 285GTX doing the same "GPGPU" type work Nvidia don't shut the heck up about.
Nvidia may, or may not be the GPGPU kings, but to say it because Nvidia say it, with no comparisons of software being run on both, its just an accepted thing people now say with zero proof. AMD sell GPU's first and foremost so talk about GPU's, they sell professional versions and they aren't slow, they just don't go on, and on, and on about it.
The problem being also that Nvidia seems to want to persaude every gamer he needs a massively powerful GPGPU and massive DP performance to run the latest games. Yet another CUDA, TWIMTBP and fancy Nvidia water effects using game in Just Cause 2........... uses Havok, doesn't require a GPGPU style device and doesn't require a high level of DP power(even though the 5870 has it anyway).
The 480gtx loses on the only scales that matter, its slower per £, its slower per Watt, and its slower per transistor = big fail. Even with all that it would be ok if they could drop prices, but they can't, they are going to make a loss at inflated prices, and can barely produce them.
The real sorry thing is for all the fanboys that think somehow it being marginally faster is fantastic news, the majority won't ever be able to find one to buy.
As for being a great gpgpu, I really wish people would stop saying that, AMD aren't actively marketing it to gaming companies as a great GPGPU, but the 5870 certainly is.
The benchmark the other day, of a company that uses Nvidia hardware and CUDA to run their data, they did a quick port, got the same software working on AMD hardware in OpenCL(and didn't take them at all long apparently) and a 5870 was, was it 4 times faster than a 285GTX doing the same "GPGPU" type work Nvidia don't shut the heck up about.
Nvidia may, or may not be the GPGPU kings, but to say it because Nvidia say it, with no comparisons of software being run on both, its just an accepted thing people now say with zero proof. AMD sell GPU's first and foremost so talk about GPU's, they sell professional versions and they aren't slow, they just don't go on, and on, and on about it.
The problem being also that Nvidia seems to want to persaude every gamer he needs a massively powerful GPGPU and massive DP performance to run the latest games. Yet another CUDA, TWIMTBP and fancy Nvidia water effects using game in Just Cause 2........... uses Havok, doesn't require a GPGPU style device and doesn't require a high level of DP power(even though the 5870 has it anyway).
The 480gtx loses on the only scales that matter, its slower per £, its slower per Watt, and its slower per transistor = big fail. Even with all that it would be ok if they could drop prices, but they can't, they are going to make a loss at inflated prices, and can barely produce them.
The real sorry thing is for all the fanboys that think somehow it being marginally faster is fantastic news, the majority won't ever be able to find one to buy.
A small amount faster and bloody hell no it doesn't justify the inevitable high price compared to the 5870.
Got to say though Dirt2 is an embarrassment for ATI, how did that happen!
pretty sure they would have had working 8800gt hardware for a long time before launch also but they had huge performance gains in drivers for those cards for a long time after launch, same with ati and their 58** series now
LOL,
That wont really so anything IMO. You cant draw a fair performance comparison as Nvidia have had six months extra to work on drivers no matter what spin is put on the results.
Soldato
I recon they have been struggling with the hardware so much the drivers won't be anywhere close to optimied!
Soldato
Those Dirt 2 results (apparently are DX9), if you look at the DX11 results both cards perform pretty much the same - the 480 is a couple fps faster.
Associate
What happens when they are clocked?
This would be my question.
This would be my question.
Soldato
Boom.
And I'm sure ATI had 6 months to work on drivers before their release.
ATI wasn't moving the day of release for months like nVidia did.
The GTX 480 runs at 92 degrees on stock clocks. Good luck overclocking that.
Its kind of like the 4870, too hot and inefficient.
Soldato
I really hope that is the case!
Associate
- Joined
- 3 Mar 2009
- Posts
- 591
Hey! be reasonable now, the 4870 wasn't this bad!