NVIDIA GF100 & Friends speculation

[DavidGraham]

No, he has to learn to fake first
Putting some idea of performance gain into a spreadsheet without jittering your results is not a good idea. Relative performance of gtx480 from his table (same row/columns):

1,29    1,61    1,25    1,28    0,75
1,22    1,17    1,16    1,15    1,21
1,50    1,50    1,50    1,50    1,50
1,67    2,12    1,88    1,67    1,17
1,95    2,25    2,11    1,59    1,14
2,33    2,08    2,16    1,88    1,41

So it's always exactly 50% faster than gtx295

Seriously -No troll intended- , but I don't see the part where GTX 480 is always 50% faster than GTX 295 :

 

[ShaidarHaran]

Seriously -No troll intended- , but I don't see the part where GTX 480 is always 50% faster than GTX 295

I'm not sure if you're serious, but every number in the GTX 480 column is precisely 50% faster than that in the GTX 295 column...

 

[Razor1]

yep they are, probably fake.

 

[Silent_Buddha]

Seriously -No troll intended- , but I don't see the part where GTX 480 is always 50% faster than GTX 295 :

You really don't see it? Those values are all exactly 50% higher.

[edit] bah this is what happens when you have to answer the door in the middle of a post.

Regards,
SB

 

[DavidGraham]

You really don't see it? Those values are all exactly 50% higher.

[edit] bah this is what happens when you have to answer the door in the middle of a post.

Regards,
SB

Sorry must be the way I calculated the percentage :

I take GTX 295 score , divide it on GTX 480 score and then multiply by 100 :
(87 / 130) X 100 = 66%

So GTX 295 scores 66% of GTX 480 Performance .


OR I subtract GTX 295 score from GTX 480 score , and then divide the result with GTX 295 and then multiply by 100 :

(130 - 87 ) / 87 X 100 = 50%

That way GTX 480 scores are 50% faster than GTX 295's .

I am really confused , which method reflects the real situation ?!

 

[Alexko]

Sorry must be the way I calculated the percentage :

I take GTX 295 score , divide it on GTX 480 score and then multiply by 100 :
(87 / 130) X 100 = 66%

So GTX 295 scores 66% of GTX 480 Performance .


OR I subtract GTX 295 score from GTX 480 score , and then divide the result with GTX 295 and then multiply by 100 :

(130 - 87 ) / 87 X 100 = 50%

That way GTX 480 scores are 50% faster than GTX 295's .

I am really confused , which method reflects the real situation ?!

130.5/87 = 1.50, in other words it's 50% more.

 

[NathansFortune]

Ironically, this can also be a great way to lose the price war.

Winning the battle to lose the war...

Nvidia are very cash rich and could go for a long time loss-leading on top-end products to maintain market share, but I think they are divesting (Tesla, Tegra) specifically to avoid expensive price-wars in the GPU arena.

 

[Kynes]

They say in the article that they are making an extrapolation based on nVidia marketing figures of "40-60% higher performance than 295", so those aren't real numbers.

 

[DavidGraham]

130.5/87 = 1.50, in other words it's 50% more.

Wow , Simple yet effective .. thanks !

 

[digitalwanderer]

They say in the article that they are making an extrapolation based on nVidia marketing figures of "40-60% higher performance than 295", so those aren't real numbers.

Yeah, looks like not a lot of other people went to the time of translating the page and trying to read it.

The data used are based on a review conducted by the ATI techreport.com 5970, and has been a 50% improvement in FPS on the GTX 295, NVIDIA reported as improvements in the environment of 40-60% compared this graph. We could just say it, but it seems better to present the possible scenario that will move the graphics from NVIDIA if fulfills the promise of "40-60%" of improvement on their last dual graphics. That's why this is a preview, not a review, and data need not correspond to the actual card once these go on sale.

 

[KamraK]

New scores from Japan or China i dunno

3Dmark Vantage Extreme Preset

HD5850 = X6200
GTX470 = X7100
HD5870 = X8000 13% faster then GTX470

 

[KamraK]

New scores from Japan or China i dunno

3Dmark Vantage Extreme Preset

HD5850 = X6200
GTX470 = X7100
HD5870 = X8000 13% faster then GTX470

Link:
http://translate.googleusercontent....gle.pl&usg=ALkJrhh4bEHrFmiDuGZ3il29eQYMn2jtcQ

 

[Kaotik]

New scores from Japan or China i dunno

3Dmark Vantage Extreme Preset

HD5850 = X6200
GTX470 = X7100
HD5870 = X8000 13% faster then GTX470

http://forum.beyond3d.com/showthread.php?p=1397987#post1397987

 

[KamraK]

http://forum.beyond3d.com/showthread.php?p=1397987#post1397987

That is fresh new score from today... From vga.it168.com i forgot write a link

 

[nutball]

Have we now got to the point that we need websites to help us out with simple arithmetic?

 

[Jawed]

Right so far?

Yep.

Okay, my brain fires up, better wait for confirmation (GPUs == in-order?).

Processing of ALU instructions is strictly in-order in ATI.

There's some debate about whether NVidia's GPUs actually re-order ALU instructions - I think they do.

Jawed

 

[Recall]

Seems Hilbert has had his cage rattled by Charlie.

http://www.guru3d.com/article/nvidia-geforce-470-480/

the GeForce 470 seems to be clocked at roughly 650 MHz, that's lower than I expected

 

[nutball]

Seems Hilbert has had his cage rattled by Charlie.

That article makes no sense. It a) regurgitates a whole pile 'o' stuff that's been out on t'Internet for months now, BS rumour or otherwise, and b) says "we know the specs" / "we're not going to tell you" / "the clock speeds are roughly 650MHz". So... he did tell us, even though he said he wasn't going to?

 

[ZerazaX]

New scores from Japan or China i dunno

3Dmark Vantage Extreme Preset

HD5850 = X6200
GTX470 = X7100
HD5870 = X8000 13% faster then GTX470

Someone said that Cat 1.3 beta gives 1000 more points to the 5870 in Vantage for P preset... will be interesting to see where this goes

 

[Silent_Buddha]

Starting to remind me of X1800 XT versus 7800 GTX with 7900 launching soon after X1800 XT. Would have been competitive if it had launched when it was supposed to (Win7 launch), but compared to a refresh product, may or may not look good. Over 6 months (Sept. 10 -> Mar. 26) after the competition, even X1800 XT wasn't that late.

That Guru3D editorial was interesting.

Regards,
SB