NVIDIA GF100 & Friends speculation

That looks a lot like cherry picking benchmarks to me Chalnoth.

Or maybe minimums are the new averages, just like fastest single gpu is the new fastest card.

When you've got fuck all, you spin. :)
 
Chalnoth said:
There's a hint of something special in the Anandtech review, though: minimum framerates. Check this out:
http://www.anandtech.com/video/showdoc.aspx?i=3783&p=9

For the average framerates, in these benchmarks the 480 GTX is ~10% faster than the HD 5870, and quite a bit slower than the 5970. Ho hum, nothing special here.

But look at the minimum framerates!
The 4x0 cards just wipe the floor with the competition! The 470 GTX bests the 5870, and the 480 GTX is trading blows with the 5970! This means, to me, that despite the average framerates, the 4x0 cards, at least for Crysis: Warhead, are producing much better gameplay.

Now, the question is: is this unique to Crysis: Warhead? Or is this a pattern that repeats with other games as well? Is it something that only happens in newer games? Sadly, Anand doesn't help us here, as he only tested the minimum framerates in this one game. Did any other reviewers look at the minimums in a wider variety of games? I know Kyle over at HardOCP did, but it's difficult to tell where the cards stand due to how he tests.
Click to expand...

Mins are better in several games across different reviews. Even older games like GTAIV etc.

I posted this before from PCGH.

http://forum.beyond3d.com/showpost.php?p=1413838&postcount=5313
 
Chalnoth said:
There's a hint of something special in the Anandtech review, though: minimum framerates. Check this out:
http://www.anandtech.com/video/showdoc.aspx?i=3783&p=9

For the average framerates, in these benchmarks the 480 GTX is ~10% faster than the HD 5870, and quite a bit slower than the 5970. Ho hum, nothing special here.

But look at the minimum framerates!
The 4x0 cards just wipe the floor with the competition! The 470 GTX bests the 5870, and the 480 GTX is trading blows with the 5970! This means, to me, that despite the average framerates, the 4x0 cards, at least for Crysis: Warhead, are producing much better gameplay.

Now, the question is: is this unique to Crysis: Warhead? Or is this a pattern that repeats with other games as well? Is it something that only happens in newer games? Sadly, Anand doesn't help us here, as he only tested the minimum framerates in this one game. Did any other reviewers look at the minimums in a wider variety of games? I know Kyle over at HardOCP did, but it's difficult to tell where the cards stand due to how he tests.
Click to expand...

We need some 2GB 5870 benchmarks with minimum FPS to see if it's a frame buffer issue or if Fermi simply kicks more ass ;)
 
SimBy said:
We need some 2GB 5870 benchmarks with minimum FPS to see if it's a frame buffer issue or if Fermi simply kicks more ass ;)
Click to expand...

Only if they exceed 1920x1080 'very high' with 4xAA otherwise it will not be VRAM limited in Crysis/Warhead/Crysis Wars.
 
It looks like the response is overall fairly positive from the several reviews I skimmed. Its interesting the contrasting conclusions between the different sites, some giving the nod to the Cypress GPUs and others giving the nod to Fermi. In all I think its pretty fairly balanced and it comes down to what you want.
 
Chalnoth said:
There's a hint of something special in the Anandtech review, though: minimum framerates. Check this out:
http://www.anandtech.com/video/showdoc.aspx?i=3783&p=9

For the average framerates, in these benchmarks the 480 GTX is ~10% faster than the HD 5870, and quite a bit slower than the 5970. Ho hum, nothing special here.

But look at the minimum framerates!
The 4x0 cards just wipe the floor with the competition! The 470 GTX bests the 5870, and the 480 GTX is trading blows with the 5970! This means, to me, that despite the average framerates, the 4x0 cards, at least for Crysis: Warhead, are producing much better gameplay.

Now, the question is: is this unique to Crysis: Warhead? Or is this a pattern that repeats with other games as well? Is it something that only happens in newer games? Sadly, Anand doesn't help us here, as he only tested the minimum framerates in this one game. Did any other reviewers look at the minimums in a wider variety of games? I know Kyle over at HardOCP did, but it's difficult to tell where the cards stand due to how he tests.
Click to expand...

That's why everyone's waiting to see the figures on the 2 gig 5870s. There's speculation that the lack of memory on the 1 gig 5870 is causing a performance choke that is dropping minimum framerates down.

Still [H] reckons there's little difference in actual gameplay on their tests. If it's only a "hint of something special", it may not be much to hang your hat on.
 
Last edited by a moderator:
Bouncing Zabaglione Bros. said:
That's why everyone's waiting to see the figures on the 2 gig 5870s. There's speculation that the lack of memory on the 1 gig 5870 is causing a performance choke that is dropping minimum framerates down.

Still [H] reckons there's little difference in actual gameplay on their tests. If it's only a "hint of something special", it may not be much to hang your hat on.
Click to expand...
What, what I would really like to see, that no reviewer has ever done, is a "playability" metric that strongly weights the slowest frames, and doesn't pay much of any attention to the faster ones. After all, in real gameplay, it's highly unlikely most players will ever notice framerates above 60 fps (not least because most flat panel displays have 60Hz refresh rates...), but it's comparatively easy to see dips below 30 fps.

So, one possible way of performance-weighting would be to, say, sum up the total amount of time spent below 30 fps. Lower scores would mean more playability. Could also add scores that represent total time spent below 45 and 60 fps to get a fuller picture of how the card plays.

Of course, any metric you choose is going to be arbitrary, but I think weighting slower frames more than faster frames would be a better way of measuring playability than the average.

That said, if the 4x0 cards actually do much better at minimum framerates, I would consider them much higher-performing parts, despite not being exceptional with their average framerates.

From looking at the pcgameshardware.de results, it looks like a good fraction of the time, there isn't any noticeable difference in minimum framerates (as compared to average framerates), but when there is, the 4x0 do significantly better than the 5xxx parts.
 
ChronoReverse said:
Why not look at the [H] charts? The absolute speeds aren't as relevant but you can still see where and how much each architecture spike.
Click to expand...

But most of [H] test arent done on equal basis. Amount of AA, settings etc differ. Also they only test 4 games....
 
You must log in or register to reply here.

Similar threads

M
Interesting old GPU architecture
Replies
17
Views
3K
Ailuros
Ailuros
T
GPU Stability testing for Nvidia cards - PowerSweep
Replies
0
Views
445
T2098
T
J
Graphic card/CPU configuration for best memory graphics in Flight Simulator 98?
2
Replies
37
Views
5K
homerdog
homerdog
R
HDR settings uniformity is much needed
2 3 4
Replies
69
Views
9K
RobertR1
R
D
Speculation and Rumors: Nvidia Blackwell ...
45 46 47
Replies
928
Views
103K
Kaotik
Kaotik
Back
Top