Does this increase seem reasonable with Detonator FX?!

[Ante P]

Shadermark 1.7
43.45: avarage of all PS2.0 tests: 60 fps
44.03: avarage of all PS2.0 tests: 190 fps

TB you know more about your proggy than anyone else, do you think this looks reasonable?

 

[Doomtrooper]

ON which card ??

What does this demo show between the two driver sets ??

http://www.daionet.gr.jp/~masa/rthdribl/index.html

 

[Luminescent]

I believe Ante is referring to the FX 5800 (the Abit, no).

 

[AJ]

Shadermark 1.7
43.45: avarage of all PS2.0 tests: 60 fps
44.03: avarage of all PS2.0 tests: 190 fps

TB you know more about your proggy than anyone else, do you think this looks reasonable?

Brent, you're pretty familiar with ShaderMark. Any thoughts?

Cheers,

AJ

 

[Arun]

I remember seeing a site saying that the DetFX made the NV35 look 100% normal, but the other NV3x didn't look good at all, even worse than before sometimes. You might really want to do IQ tests there...


Uttar

 

[Bambers]

So the 5800 has gone up in that bench as well?

When I saw hardocps review of the 5900 I assumed the boost was down to the 5900s touted 2xfp power, if the 5800 has gone up too then I'd be very suspisious. Nvidia know who are getting the cards first and hardocp has used that bench a couple of times before (mainly showing how the 5800 got beaten by the 9700 (even my 8500 beat it on the 2 spot light test amongst a couple of others on a slower system)). Considering digitlife/ixbts rightmark ps2.0 results where both the 5800 and 5900 are down at under 1/2 the r350s speed...

:|

 

[boobs]

Shadermark 1.7
43.45: avarage of all PS2.0 tests: 60 fps
44.03: avarage of all PS2.0 tests: 190 fps

TB you know more about your proggy than anyone else, do you think this looks reasonable?

You might want to indicate that it's Shadermakr 1.7 or TB might not even read the thread. :?

 

[DegustatoR]

So the 5800 has gone up in that bench as well?

http://www.digit-life.com/articles2/gffx/5900u.html

There are no visual difference in quality. So it looks like just an optimization.

 

[antlers]

So the 5800 has gone up in that bench as well?

When I saw hardocps review of the 5900 I assumed the boost was down to the 5900s touted 2xfp power, if the 5800 has gone up too then I'd be very suspisious. Nvidia know who are getting the cards first and hardocp has used that bench a couple of times before (mainly showing how the 5800 got beaten by the 9700 (even my 8500 beat it on the 2 spot light test amongst a couple of others on a slower system)). Considering digitlife/ixbts rightmark ps2.0 results where both the 5800 and 5900 are down at under 1/2 the r350s speed...

:|

I might be suspicious even if the 5800 did not go up, as well. If a driver is cheating, it's easy enough to only cheat for the NV35 and not the NV30.

 

[Brent]

i haven't tested this scenerio yet

i will get to it though, i have more then a few things on my plate right now to work on 8)

 

[Ante P]

So the 5800 has gone up in that bench as well?

When I saw hardocps review of the 5900 I assumed the boost was down to the 5900s touted 2xfp power,

exactly, that's why I was so interested in the results

btw the card is a FX 5800 Ultra from Gainward

I spoke to nVidia a month ago about their low shadermark 1.7 performance, at the time they were unaware that the benchmark even existed and said they'd check it out

as for quality it's indentical with both drivers which in turn is indentical to ATis

 

[Ante P]

ON which card ??

What does this demo show between the two driver sets ??

http://www.daionet.gr.jp/~masa/rthdribl/index.html

I'll check it out, is there any benchmark mode or do I need to just look at the fps or use fraps?

 

[Brent]

ON which card ??

What does this demo show between the two driver sets ??

http://www.daionet.gr.jp/~masa/rthdribl/index.html

I'll check it out, is there any benchmark mode or do I need to just look at the fps or use fraps?

There is no benchmark mode, unfortunately

just have to look at the fps

 

[Doomtrooper]

Fraps can log the average frame rate though, I've found fraps very accurate.

 

[Tagrineth]

I spoke to nVidia a month ago about their low shadermark 1.7 performance, at the time they were unaware that the benchmark even existed and said they'd check it out

Dammit! Now you've gone and asked them to optimise.

 

[jjayb]

Interesting numbers in that digitlife review. These in particular:

http://www.digit-life.com/articles2/gffx/5900u.html#p9

They show the nv30/nv35 both still run at about 1/2 the performance of the r350 using pixel shader 2. Even more interesting is that they show the 5800 actually beating the 5900 in some of those tests. Nvidia hasn't optimized to boost the scores on this yet?

 

[AJ]

Dammit! Now you've gone and asked them to optimise.

Hmm... Interestingly enough we did see quite a drop in frame rates with PS 2.0 shader test (not the game test, but the feature test) once the detection mechanisms in the drivers were turned off.

Cheers,

AJ

 

[Miksu]

Shadermark 1.7

Any new info about this?

 

[Doomtrooper]

This Demo would show any PS 2.0 improvements pretty quickly too, any update on this from anyone ??

http://www.daionet.gr.jp/~masa/rthdribl/index.html

 

[Doomtrooper]

Bump:

http://216.239.37.104/translate_c?h...icles/468/page4.html&prev=/language_tools