[chiphell] kepler rumors suggest 15% better than 580.. price and transistors

Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
SirPauly

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
Crap Daddy said:
Isn't it? Back to your doubts I can recall only one benchmark which is Crysis 1 where the 285 scored around 30FPS and the 570 around 50FPS. If you still have doubts check benchmarks from reputable sites, if you don't find there the 285 then figure out how is the 460 compared to the 570 because the 460 is around a 285.
Click to expand...

All one has to do is investigate ComputerBase results and clearly in the 50-80 percent realm.
 
Vesku

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
That's the point, that is launch performance, where as these comments about the GTX 480 being 50-80% faster than a 285 are based on mature 480 drivers. If the 7970 was 50-80% faster than the 6970 with the launch drivers then I'd be in full panic mode if I was a NVIDIA executive, reality is that 7970 is only a respectable ~40% faster than a 6970 at launch time.

I certainly hope NVIDIA starts getting aggressive with pricing, 7970 with mature drivers would be tempting to me if they are forced to price drop it into the $400ish bracket.
 
Last edited:
railven

railven

Diamond Member
Mar 25, 2010
6,604
561
126
Vesku said:
That's the point, that is launch performance, where as these comments about the GTX 480 being 50-80% faster than a 285 are based on mature 480 drivers. If the 7970 was 50-80% faster than the 6970 with the launch drivers then I'd be in full panic mode if I was a NVIDIA executive, reality is that 7970 is only a respectable ~40% faster than a 6970 at launch time.

I certainly hope NVIDIA starts getting aggressive with pricing, 7970 with mature drivers would be tempting to me if they are forced to price drop it into the $400ish bracket.
Click to expand...

You can only lead a horse to water.

This is why the current trend continues from the last three generations. The increase at launch has been stagnant. But of course someone will post a benchmark dated 2-weeks ago after nVidia/AMD released the newest driver and the GTX 480/HD 5870 is now 20% faster than launch and call the successor a poor performer.

Didn't we have a thread here about the HD 5xxx series, something along the lines of "A year with Catalyst" that documented the performance increases that mature drivers brought with the product?
 
S

skipsneeky2

Diamond Member
May 21, 2011
5,035
1
71
GK104 will more then likely be the gtx285 of this generation to replace the gtx280.

Faster by specs $100 cheaper lower tdp and if it could hold up to a 7970 in some cases or come close it could be a winner.

At $400 its interesting:biggrin:
 
SirPauly

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
Vesku said:
That's the point, that is launch performance, where as these comments about the GTX 480 being 50-80% faster than a 285 are based on mature 480 drivers. If the 7970 was 50-80% faster than the 6970 with the launch drivers then I'd be in full panic mode if I was a NVIDIA executive, reality is that 7970 is only a respectable ~40% faster than a 6970 at launch time.

I certainly hope NVIDIA starts getting aggressive with pricing, 7970 with mature drivers would be tempting to me if they are forced to price drop it into the $400ish bracket.
Click to expand...

The computerbase data was with launch drivers.
 
J

JAG87

Diamond Member
Jan 3, 2006
3,921
3
76
Vesku said:
That's the point, that is launch performance, where as these comments about the GTX 480 being 50-80% faster than a 285 are based on mature 480 drivers. If the 7970 was 50-80% faster than the 6970 with the launch drivers then I'd be in full panic mode if I was a NVIDIA executive, reality is that 7970 is only a respectable ~40% faster than a 6970 at launch time.

I certainly hope NVIDIA starts getting aggressive with pricing, 7970 with mature drivers would be tempting to me if they are forced to price drop it into the $400ish bracket.
Click to expand...


The bigger problem with linking old benchmarks are that they were done with games that are old and don't actually stress the cards enough to show the true gap.

Also just as a general FYI, NV "mature" drivers will increase performance by 10-20% over the course of a card's life, AMD "mature" drivers are just the original drivers with 37 hotfixes on top. Check it out:

5870 review

20105.png


7970 review

43099.png



45.6 vs 39.7

There's "mature" drivers for you.
 
Vesku

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
All the older cards scores are lower, GTX 285 included. Wonder what changes in their test bench system causes that discrepancy.

Techpowerup tested more games than computerbase.de, so in terms of overall percentages theirs would be a more comprehensive picture of launch performance. Also, I will note that staying within review methodology by Techpowerup's tests the 7970 is only ~30% faster than the 6970 at launch. So the 480 was a bit bigger jump from 285 than 7970 is from the 6970.
 
Grooveriding

Grooveriding

Diamond Member
Dec 25, 2008
9,108
1,260
126
JAG87 said:
The bigger problem with linking old benchmarks are that they were done with games that are old and don't actually stress the cards enough to show the true gap.

Also just as a general FYI, NV "mature" drivers will increase performance by 10-20% over the course of a card's life, AMD "mature" drivers are just the original drivers with 37 hotfixes on top. Check it out:

5870 review

20105.png


7970 review

43099.png



45.6 vs 39.7

There's "mature" drivers for you.
Click to expand...

Um, you just compared the 5870's launch average FPS to the 5870's minimum fps in the 7970 review. o_O

Here is the average you were looking for

43096.png


Looks like AMD improved performance by 25% in this benchmark with driver revisions. :ninja:

So your comment:


JAG87 said:
Also just as a general FYI, NV "mature" drivers will increase performance by 10-20% over the course of a card's life, AMD "mature" drivers are just the original drivers with 37 hotfixes on top. Check it out:
Click to expand...

Is totally off-point, as it seems AMD improved performance by 25% through driver revisions in the scenario you just outlined. Perhaps a re-assessment of the value you place on AMD products in contrast to nvidia products is in order.
 
Last edited:
Vesku

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
Ah, totally missed the "minimum" bit. No wonder the numbers seemed so odd. Honest mistake, Jag87?
 
Skurge

Skurge

Diamond Member
Aug 17, 2009
5,195
1
71
Grooveriding said:
Um, you just compared the 5870's launch average FPS to the 5870's minimum fps in the 7970 review. o_O

Here is the average you were looking for

43096.png


Looks like AMD improved performance by 25% in this benchmark with driver revisions. :ninja:

So your comment:




Is totally off-point, as it seems AMD improved performance by 25% through driver revisions in the scenario you just outlined. Perhaps a re-assessment of the value you place on AMD products in contrast to nvidia products is in order.
Click to expand...

Facepalm moment right there. The 7870 results also include 4xAA while the original ones didn't. So it's more than 25%.

Someone's credibility has just been thrown out the window.
 
Vesku

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
4xAA is in the original 5870 test, it's just not listed in the same spot. Facepalms for everyone!
 
M

MrTeal

Diamond Member
Dec 7, 2003
3,569
1,699
136
Even outside average vs minimum, the increase over time is more mature drivers, but also

CPU: Intel Core i7-920 @ 3.33GHz
Motherboard: Intel DX58SO (Intel X58)
Chipset Drivers: Intel 9.1.1.1015 (Intel)
Hard Disk: Intel X25-M SSD (80GB)
Memory: Patriot Viper DDR3-1333 3 x 2GB (7-7-7-20)

vs

CPU: Intel Core i7-3936 @ 4.3GHz
Motherboard: EVGA X79 SLI
Chipset Drivers: Intel 9.​2.​3.​1022
Power Supply: Antec True Power Quattro 1200
Hard Disk: Samsung 470 (240GB)
Memory: G.Skill Ripjaws DDR3-1867 4 x 4GB (8-10-9-26)
 
Z

ZimZum

Golden Member
Aug 2, 2001
1,281
0
76
JAG87 said:
The bigger problem with linking old benchmarks are that they were done with games that are old and don't actually stress the cards enough to show the true gap.

Also just as a general FYI, NV "mature" drivers will increase performance by 10-20% over the course of a card's life, AMD "mature" drivers are just the original drivers with 37 hotfixes on top. Check it out:

There's "mature" drivers for you.
Click to expand...

Look at the numbers for the 285, 4870 and 8800. Every card thats listed in both charts shows anywhere from 33%-62% performance drop. Including the NV cards. Obviously the benchmark or the test system aren't exactly the same.
 
SirPauly

SirPauly

Diamond Member
Apr 28, 2009
5,187
1
0
To place things in perspective what both nVidia and AMD have done from a performance point-of-view currently when comparing to the older generations like the 8800 GT and HD 3870.
 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom