Nvidia screen shot competition

[Charles Wirth]

Most people don’t know the reduction of image quality Nvidia is capable of to compete at high end benchmarking. That is the premise of this contest. I would like to see Nvidia with higher standards as well as benchmarks that reported various settings that produce these results. For those who see these high scores, this is a sample of what some of them see on screen.

Rules:
Any Nvidia video card
Must be submitted to the ORB
Any 3Dmark product
One entry per person
All entries must be submitted by Monday 12 noon PST
LODBias settings must be included with the post

Scores will be based on:
Image quality (lack of)
Tearing (text must be readable/visible lower right)
Image with no textures or flat shades are a plus
Antialias round tires into square boxes
Objects missing from scene are a plus
Post a compare of the machine below you (ORB)
Abnormally high frame rate

Cooling, speed, and card type will not be scored.

Comments and criticism welcomed.

The winner shall receive an X800XL video card and a one lifetime membership to ati.com

[fatfreepork]

3d01 - does it have to be run at default settings?

How do i show Level of Detail Bias settings?

And how do i make my card do that?


Image quality (lack of)
Tearing (text must be readable/visible lower right)
Image with no textures or flat shades are a plus
Antialias round tires into square boxes
Objects missing from scene are a plus
Post a compare of the machine below you (ORB)
Abnormally high frame rate

Sorry for questions. :slap:

[Disposibleteen]

what about us ati guys????????

[Charles Wirth]

Just post what you used for LOD

You can run 2K1, 2K3, 2K5 at any settings, sceen shot any frame and post it in this thread.

[Charles Wirth]

what about us ati guys????????

Its not too hard to locate a Nvidia card capable of competing in this competition.

[faruquehabib]

lol..just sold my 6600gt for a 9600xt vivo :D

[fatfreepork]

I just gave a quick tries and 3dMark01 is the only one that i can select settings so its a little unfair for people without pro licenses. And for some reason, when i take a screenshot in 3dMark, I get nothing when i paste into microsoft paint.

[Charles Wirth]

I asked Kunaak to post an example.

you can run one test and submit it, no need to run all test. just submit it to show machine specs.

[Kunaak]

use FRAPS, and hit F10.

http://www.fraps.com/

[Kunaak]

heres something for the hell of it, to show how insane LOD can get... by the way, do I have some good timing or what?? 1 frame off from getting the exact same frame for both pics, damn

I just did this a second ago, to show what to expect.

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31374&stc=1

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31373&stc=1

I did nothing more then enable the power user range to 15 for mipmaps in rivatuner. other then that, it's just one was at zero, one was at 15+ LOD.

however, to be somewhat fair to Nvidia... LOD this low is a huge hinderance on the score.
I lose about 1000 frames from the first 3 tests of 3dmark2003 and about 500 from the last test, with anything higher then LOD + 2.

[fatfreepork]

use FRAPS, and hit F10.

http://www.fraps.com/

Oh thanks! I totally forgot about Fraps' capability to take screens. :doh:

[Disposibleteen]

Its not too hard to locate a Nvidia card capable of competing in this competition.
my only nvidia is an fx 5200, i wouldnt call that capable ;)

[Charles Wirth]

my only nvidia is an fx 5200, i wouldnt call that capable ;)

That would run 2K1 and DX8

[Disposibleteen]

whos got a download link for DX8, i want to try this

[fatfreepork]

nVidia card: 6800LE

Benchmark: 3dMark01se

Comparison link: http://service.futuremark.com/compare?2k1=8591408

These are with LodBias at 15
http://www.ocrig.com/xs/1.JPG
http://www.ocrig.com/xs/2.JPG
http://www.ocrig.com/xs/3.JPG
http://www.ocrig.com/xs/4.JPG
http://www.ocrig.com/xs/5.JPG
http://www.ocrig.com/xs/6.JPG
http://www.ocrig.com/xs/7.JPG
http://www.ocrig.com/xs/8.JPG
http://www.ocrig.com/xs/9.JPG
http://www.ocrig.com/xs/10.JPG
http://www.ocrig.com/xs/11.JPG
http://www.ocrig.com/xs/12.JPG
http://www.ocrig.com/xs/13.JPG
http://www.ocrig.com/xs/14.JPG
http://www.ocrig.com/xs/15.JPG
http://www.ocrig.com/xs/16.JPG
http://www.ocrig.com/xs/17.JPG
http://www.ocrig.com/xs/18.JPG
http://www.ocrig.com/xs/19.JPG
http://www.ocrig.com/xs/20.JPG

If the images are too big, then just tell me and i will remove the image tags. Sorry :eek:

[calcal]

wut is LOD????????

how do i change the quality to bad?

[BrokenWall]

i got an eye opener coming here

[zabomb4163]

ive done worse when benching. lol. where are the memory overclocks then we'll see some real tearing.

[calcal]

can anyone tell me how to make the quality like fatfreeport?

[BrokenWall]

Its my 6800 Ultra
3DMark01se
LOD +15
Some other tweaks.

http://service.futuremark.com/compare?2k1=8592608

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31433&stc=1

Hope this one counts, cause its the vertext shader test of 3DMark01se

[Sentential]

First we have the non-LOD picture and Score:

http://img.photobucket.com/albums/v180/Sentential/3DMark2001SE2005-05-2814-06-09-84.jpg

BENCHMARK SCORE: http://service.futuremark.com/compare?2k1=8467631

Next we see my LOD +3 (aka Force Mips To Lowest Levels):

http://img.photobucket.com/albums/v180/Sentential/LOD36600.jpg

BENCHMARK SCORE: http://service.futuremark.com/compare?2k1=8490269


In addition we have LOD-3:

http://img.photobucket.com/albums/v180/Sentential/LOD-36600.jpg

Stark difference if I do say so myself. Now I will show the steady image loss created by LOD:

First we have LOD bias set at 0:

http://img.photobucket.com/albums/v180/Sentential/0.jpg

Then 1:

http://img.photobucket.com/albums/v180/Sentential/1.jpg

Then 2:

http://img.photobucket.com/albums/v180/Sentential/2.jpg

Then 3:

http://img.photobucket.com/albums/v180/Sentential/3.jpg

Now some 3d05 Pictures:

First LOD at normal levels:

http://img.photobucket.com/albums/v180/Sentential/01.jpg

Then +1:

http://img.photobucket.com/albums/v180/Sentential/11.jpg

Then +2:

http://img.photobucket.com/albums/v180/Sentential/20.jpg

Finally +3:

http://img.photobucket.com/albums/v180/Sentential/30.jpg

[calcal]

what is LOD and how i do i turn it on POEPLE COME ON!!!!!!

[fatfreepork]

Level of Detail

[forey]

i dont really understand what to do, what do you to in rivatuner to make it look like that?

edit i found it, doesnt make it look as bad as the guys up there

[CTKP]

use FRAPS, and hit F10.

http://www.fraps.com/

I can't get fraps to take a screen, any alternatives? I could always use my camera, any objections fugger?
Edit- fixed, color quality needs to be 16bit+

[Charles Wirth]

An easy way to adjust LOD is to download rivatuner from our friends at guru3d.com

Nice submissions so far.

[Kunaak]

Rules:

"One entry per person"




posting 100 shots doesn't mean your gonna win :eek:

[JasonDTM]

Any Idea of what I can use.... I tried to set it LOD +15 in rivatuner (When I was benching) and it never really worked because it didnt look like utter :banana::banana::banana::banana:e :)

[fatfreepork]

posting 100 shots doesn't mean your gonna win :eek:

I only had one entry. Nowhere does it say only 1 image allowed. Fugger listed some things he wanted and it takes more than one scene to get that.

[drcrawn]

Ok, I know these are probably invalid because the text is not "visible", but these are too good not to show!

LOD+ 3.5
"other tricks"
http://service.futuremark.com/compare?2k1=8592475

http://img290.echo.cx/img290/8310/1st2ye.jpg (http://www.imageshack.us)
***
http://img286.echo.cx/img286/3742/2nd9zz.jpg (http://www.imageshack.us)
***
http://img286.echo.cx/img286/3254/3rd8uc.jpg (http://www.imageshack.us)

[Charles Wirth]

Drcrawn, those will be accepted. If there was tearing and I could not read text then i would be rejected. Nice entries btw

[Pinnacle]

lol..just sold my 6600gt for a 9600xt vivo :D

Why on earth would you wanna do that?

I had both and the GT smacks that XT around

[drcrawn]

Drcrawn, those will be accepted. If there was tearing and I could not read text then i would be rejected. Nice entries btw

:toast: Great, thanks! Btw is the 800xl an agp or pcie card? Great idea also for this comp whoever came up with it. :clap:

[kryptobs2000]

While I do not have my old 6800s I do have 2 comparison benches. The image quality looks identical to what was already posted.

I would also like to note that the LOD tweek does not affect 03 or 05. This is a specific 01SE hack.

First we have the non-LOD score and picture:

http://img.photobucket.com/albums/v180/Sentential/3DMark2001SE2005-05-2814-06-09-84.jpg

BENCHMARK SCORE: http://service.futuremark.com/compare?2k1=8467631

Next we see my LOD +3 (aka Force Mips To Lowest Levels):

http://img.photobucket.com/albums/v180/Sentential/LOD36600.jpg

BENCHMARK SCORE: http://service.futuremark.com/compare?2k1=8490269


In addition we have LOD-3:

http://img.photobucket.com/albums/v180/Sentential/LOD-36600.jpg

Stark difference if I do say so myself.

Next Ill show the PS hacks...


Why does a negative LOD increase fps?

[Nazaar]

http://service.futuremark.com/compare?2k3=4004032

i only have screens from the first test cause i'm on a geforce 2MX 200 and it can only run one test in 3dmark2k3 :fact:

lod +15 and some other fun stuff :)
http://www.xtremesystems.org/forums/attachment.php?attachmentid=31455

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31454

the first non-game test (fill rate i think) is just completely gray :p:

[forey]

my LOD is -3 > 0 > +3
no higher than 3?

[WeStSiDePLaYa]

hmmmmm guess its time for me to tinker with my vmodded 5950U and nature 03 :p:

[drcrawn]

my LOD is -3 > 0 > +3
no higher than 3?

Here dude, that's how you get -15 to +15.

http://img132.echo.cx/img132/4712/riva8ql.jpg (http://www.imageshack.us)

[BrokenWall]

think everyone is got about the same :banana::banana::banana::banana:ty quality pics. LOL

[WeStSiDePLaYa]

yup, good idea, i can run benchies that are unbelievably crappy. Well I think I might have killed my 5950U trying this. BUT..... I CAN make it through a whole run with this happening. My card didn't like it and think its toast. :( I hope I win because I'm getting some decent pixel explosions even on desktop.

I included other pics of image quality to show the HUGE difference there is between it but still able to run a full bench and get valid score.


LOD 3+

http://service.futuremark.com/compare?3dm05=24138

http://xtremesystems.org/forums/attachment.php?attachmentid=31465&stc=1



http://xtremesystems.org/forums/attachment.php?attachmentid=31466&stc=1

[crmaris]

LOD +15
gforce mx 4000 :banana:

http://service.futuremark.com/compare?2k1=8593430

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31468&stc=1

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31469&stc=1



http://www.xtremesystems.org/forums/attachment.php?attachmentid=31471&stc=1

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31472&stc=1
http://www.xtremesystems.org/forums/attachment.php?attachmentid=31530&stc=1

[wittekakker]

Lol, I had the same results yesterday, but fraps wouldt take screenshots for gods sake.

Anyway, here is some 3d mark 05 stuff, looks very good if you first played around with 3d mark 01 :)

LOD +15
Nvidia 6800LE

http://service.futuremark.com/compare?3dm05=889657

http://home.scarlet.be/~gebrvd00/3d01.jpg
http://home.scarlet.be/~gebrvd00/3d02.jpg
http://home.scarlet.be/~gebrvd00/3d03.jpg
http://home.scarlet.be/~gebrvd00/3d04.jpg
http://home.scarlet.be/~gebrvd00/3d05.jpg

[dubya01]

Geforce mx440se, Duron 1.6 @ 1.92.

LOD is at 15, and some other various tweaks.
ORB: http://service.futuremark.com/compare?2k1=8593827

http://img281.echo.cx/img281/1039/07sb.jpg
http://img281.echo.cx/img281/6950/56567sk.jpg
http://img281.echo.cx/img281/1263/18zs1.jpg
http://img281.echo.cx/img281/6015/71yk.jpg
http://img281.echo.cx/img281/2124/58xb.jpg
http://img281.echo.cx/img281/1194/986wi.jpg

[P_1]

EDIT: Got valid ones now but not as impressive as the others lod 15
here is 01 compare link: http://service.futuremark.com/compare?2k1=8594714
03: http://service.futuremark.com/compare?2k3=4006906
http://www.xtremesystems.org/forums/attachment.php?attachmentid=31566&stc=1 http://www.xtremesystems.org/forums/attachment.php?attachmentid=31567&stc=1 http://www.xtremesystems.org/forums/attachment.php?attachmentid=31563&stc=1 http://www.xtremesystems.org/forums/attachment.php?attachmentid=31564&stc=1 http://www.xtremesystems.org/forums/attachment.php?attachmentid=31565&stc=1
aww look at that cute little butterfly^
http://img254.echo.cx/img254/2816/37de.jpg
cute turtle ^^

[dubya01]

Wow lol

[WeStSiDePLaYa]

Here is a video of a 2001se car low run. Thought this would be a little treat to show how low nvidia cards can go.

It only has a transfer of 10mb's a day. so would someone be generous enough to host it somewhere for me? or direct me somewhere i could host it better. if its down and you want to see it send me an email.

http://home.ripway.com/2005-5/316764/Video.zip

[WeStSiDePLaYa]

[QUOTE=P_1]These were taken on a 6600gt but 3dmark01 wont let me publish.These probably arent valid but these are also too good to pass up :) LOD @ 0


you probably cant publish because 3dmark catches those tweaks. so i guess its not valid for the contest? :confused:

[Highland3r]

Here is a video of a 2001se car low run. Thought this would be a little treat to show how low nvidia cards can go.

It only has a transfer of 10mb's a day. so would someone be generous enough to host it somewhere for me? or direct me somewhere i could host it better. if its down and you want to see it send me an email.



rehosted :D (http://www.drunken-student.co.uk/xs/Video.zip)

~ 24gb bw still left this month, so shouldnt be a problem :)

Will stick a 4200 in later and give it a whirl. The run the screengrabs were taken from needs to be uploaded to ORB right?

[Highland3r]

http://service.futuremark.com/compare?2k1=8594115
LOD @ -24 (didnt seem to make much difference beyond about 0 tho... o_O) Quincux, 8*AF, texture sharpening. Poor little 4200 struggled with most of the tests, as can be seen by score and FPS in screengrab :S

http://www.drunken-student.co.uk/xs/3DMark2001SE%202005-05-30%2008-48-12-53.jpg

http://www.drunken-student.co.uk/xs/3DMark2001SE%202005-05-30%2016-36-36-28.jpg

http://www.drunken-student.co.uk/xs/3DMark2001SE%202005-05-30%2016-39-03-65.jpg

<edit> 2 more added. Specs as per initial link. Bottom 2 running lod of 13.5, bottom pic in 640*480*16bit

If I need to choose one from this lot, let me know :D </edit>

[Charles Wirth]

Highlander, you have a few more hours to add to your entry like everyone else.

[Highland3r]

Highlander, you have a few more hours to add to your entry like everyone else.

So we can enter more than 1? :D

[Charles Wirth]

Just in that post

[Highland3r]

Ah ok :D Cheers so its 1 post per person

[Bennah]

So let me get this right its a screenie using an nVidia card with LOD tweaks etc so it messes up basically... ?

[simplyadvanced]

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31469&stc=1

my fav!

1970s animation of the mars rover!

[perkam]

i got an eye opener coming here

I USED TO THINK LOD JUST BLURRED IT A TAD BIT....OMG WAS I WRONG.

Though now I can respect Charlie's decision not to post his WR 05 score cos it had LOD :)

perkam

[WeStSiDePLaYa]

So any word on when winner will be announced?

and thanks a ton for rehosting my video! you rock man! :banana: :clap:

[Entity_Razer]

dam i didn't have time to do it and as gonna do it next weekend ....
DAM,DAM,DAM,DAM

that old geforce 256 would've done wonders :D

[P_1]

what happens if the post has the wrong time on it? (this post is in b4 noon pst)

[perry_78]

Hehe, nice idea :D

Shame I don't have the paste to fire the old 6600GT up.

Maybe I'll try my GF3Ti200 :)

[Kazoo]

Okay, this contest is officially closed...winner announced shortly -- stay tuned!

[fatfreepork]

ooh! i was gonna post worse screens! sucks i posted first cus all the people after me could see what i did. oh well. Good luck everybody!

[WeStSiDePLaYa]

Okay, this contest is officially closed...winner announced shortly -- stay tuned!
yaaaaaaaay! I could use a new card now that mines unusable in anything besides 2D basic windows stuff :( :bsod:

[Entity_Razer]

begging won't help ya know, Xtremesystems mods and admins are made up of pure silicone and Thermalpaste as blood.

:)

[WeStSiDePLaYa]

begging won't help ya know, Xtremesystems mods and admins are made up of pure silicone and Thermalpaste as blood.

:)


LOL, with a quad core FX-78 for brains i bet too. who needs to buy TIM when you just cut yourself and it flows out.

[wittekakker]

If it was my chose, I would take the GF4 screenshot with the green water texture.
And I have a strong feeling that he is going to win.
What wonders me is that there are not much people joining this competition.
It's my first one on XS, I expected alot more entrie's.

[fatfreepork]

look at the rest of that image though. its still reasonably good quality.

[Charles Wirth]

Winner will be announced shortly.

[WeStSiDePLaYa]

Winner will be announced shortly.


OMG, ill be watching this thread like a hawk.

[dubya01]

Already am :p

[CD 1986]

I should look at the competition section more often! Wish i'd have seen this one. Oh, and i like the way the winner of this "nvidia-based" competition gets an ATI video card!!! :p:

Good luck guys!!! :toast:

[WeStSiDePLaYa]

I should look at the competition section more often! Wish i'd have seen this one. Oh, and i like the way the winner of this "nvidia-based" competition gets an ATI video card!!! :p:

Good luck guys!!! :toast:


because it is ment to show how nvidia can up their standards with IQ and make benchs more equal and less optimizations that have no real world use at all. except make their cards "look" faster. ;)

[MaxxxRacer]

I think west side has this one won by a mile with his wings of fury 03 image... i can barelly even tell what the heck it is... lol

[P_1]

i had those kind of pics too but i didnt wanna post them because i didnt think they were interesting. I had pics where it was completely dark with a spec or 2 of lights. I also had a pic where there was a white box around the proxy con soliders with the ground grating all messed up. Ill post them after the competition is over because i dont know if the mods will like it if i put up new pics now.(i dunno what the limit was so i just picked the most interesting ones and posted them)

[dubya01]

So any word yet?

[WeStSiDePLaYa]

So any word yet?
patience, we will know soon enough.

[Charles Wirth]

Congratulations to Nazaar for this winning entry. :toast:

http://www.xtremesystems.org/forums/attachment.php?attachmentid=31455

PM me with your mailing info and slot type needed.

Thanks everyone for the awesome entries. This was a tough decision, sorry for the delay.

[fatfreepork]

congrats nazaar!

[Nazaar]

wow thanks a lot :toast:

[Sentential]

Congrats

[Highland3r]

Nice one mate :D A great, crap quality screenshot LOL!

[wittekakker]

Concratz nazaar.
I'll keep an eye on TN's 3d mark bench thread :)

[drcrawn]

Congrats Nazaar. This was a fun little contest. Enjoy your ATI x800xl!

[Kunaak]

that shot there was one of the most suprising to me, cause it's a totally neuterd shot, of a extremely recognizable image from 3dmark2003. the sky being reduced to just 3 blocks, and the plane itself reduced to just 2 colors, is almost shocking how much LOD and other nvidia based tweaks really can go.

Congrats to Nazaar for a hell of a shot, and Enjoy the Card :D

[P_1]

i had those kind of pics too but i didnt wanna post them because i didnt think they were interesting. I had pics where it was completely dark with a spec or 2 of lights. I also had a pic where there was a white box around the proxy con soliders with the ground grating all messed up. Ill post them after the competition is over because i dont know if the mods will like it if i put up new pics now.(i dunno what the limit was so i just picked the most interesting ones and posted them)
Here are the pics... just wished i posted these instead of the other ones.

http://img84.echo.cx/img84/3294/ihadonetoo5dv.jpg

http://img84.echo.cx/img84/6039/asd1oy.jpg

[P_1]

err for some reason the other 2 pics didnt come out on the post above
http://img27.echo.cx/img27/4182/asd23tz.jpg

http://img27.echo.cx/img27/5743/blocksoldiers5zt.jpg

[Kunaak]

holy cows, I just realized, hes gonna go from a Geforce 2 MX200 to a X800XL :D

[WeStSiDePLaYa]

nice job man! enjoy your new card. thats a damn big improvement over what you have now. good contest! :toast:

hopefully a new contest will come around again that i have a chance of winning and can get e gfx card to replace my now crap heap. lol

[Charles Wirth]

Next contest is in July

[texuspete00]

I just don't get the ATi fan*&( aire in some posts. I mean both nV and ATi look like crap when you turn down the driver quality and as far as I know..... these images that are so bad they are funny are only achievable with rivatuner. Am I missing something?

[Charles Wirth]

You cannot do this with ATI

This is what we see when we bench for the top spots with Nvidia, and it is considered legal. It gives up 10+ FPS in each test = hundred/thousands of fair points...

[P_1]

You cannot do this with ATI

This is what we see when we bench for the top spots with Nvidia, and it is considered legal. It gives up 10+ FPS in each test = hundred/thousands of fair points...
do we get a prize if we pull it off?

[Charles Wirth]

I will give you a custom title if you can get close. Use the winner as reference.

[P_1]

I will give you a custom title if you can get close. Use the winner as reference.
k ill try to have pics as soon as i finish with all this other work :D

[P_1]

heh i did a bunch of crazy things now my wings of fury is at 253fps while b4 it was at 199.5 i set lod bias to 15 and flip queue size to 5 and in enabled every optimization i could find. This resulted in a black screen that fraps could not even load on. http://service.futuremark.com/compare?2k3=4009858 will do a full test on this to see if i get increases all around

edit: HAHAHAHAHHAHAHA I HAVE DONE IT!!!!!!!!!
http://service.futuremark.com/compare?2k3=4009879

33,944 3dmarks 03 no screens tho cause it was so bad that the screen was all black :toast: btw this is stock 9800xt clocks (this is a 9800pro flashed to xt with r360 core with samsung chips)

[P_1]

bump

[Charles Wirth]

Hmm, let us know what you did. If it is legal for Nvidia it should be legal for ATI :ROTF:

[P_1]

Hmm, let us know what you did. If it is legal for Nvidia it should be legal for ATI :ROTF:
First heres a pic that is just too good: http://img191.echo.cx/img191/8417/teh0wnage0fe.png

*removed tweaks for safety and cleanliness of the orb and also due to many users to remove*

[Stuperman]

no, that is a 9600XT @ stock, ya bunch of cheaters:)

[kryptobs2000]

Can you tell me where you found the lod tweak for the ati cards in the registry? I searched for a key of lod and didn't come up with anything.

[P_1]

oops i forgot to delete this one

[Nazaar]

holy cows, I just realized, hes gonna go from a Geforce 2 MX200 to a X800XL :D

haha lol, no that 2MX 200 is the only nvidia card i have atm. I also have a 9600pro and a x800 xt-pe :p:

[MaxxxRacer]

thats some sweeet tweakiing P 1.. lol

[P_1]

thats some sweeet tweakiing P 1.. lol
thx :D :woot:

[texuspete00]

You cannot do this with ATI

This is what we see when we bench for the top spots with Nvidia, and it is considered legal. It gives up 10+ FPS in each test = hundred/thousands of fair points...

Seems to me then the gripe is with futuremark users. Seems silly to hold Nvidia accountable for things that can only be done exclusively with rivatuner. It's an extremely robust tool set that can can almost be described as hacking the drivers. Sorry ATI Tool doesn't have the options of sending IQ into the dirt. I just see your argument falling apart over "considered legal." That's a interpretation of people's views that might be more split then you think. I guess I'm reading the tones right... there is a lot of pent up anger at NV over rivatuner in this thread.

I know it's not what you want but you'll probably be closer to achieving parity when some bloke finds a trigger in an ATi driver that can do this. Having a bit of programming background, I'm damn near sure what goes in is they key on a value that was never meant to be taken out to such extremes in range, aka LOD. It could be less obvious in ATi code, or perhaps wizzard doesnt want to be in this mess and include it (more power to him) and there are really no other robust OC Ati apps. Rivatuner is made with NV in mind first, the LOD hack not working for ATi in it doesn't mean much.

I've got a 6800GT and an X800XL. I don't use hacks though so my NV sucks at 3dmark. I guess the nV guys are desparate to keep up... blame them. I've got my own questions on the integrity of 3dmark scores.... namely why my bad clocking ATI can slaughter my GT in 3dmark, but then probably lose every single game. It's no mystery to me that my well clocking GT beats my bad clocking XL in games, that's easy, just why the "gamer's benchmarks" are so bad at reflecting this. If I could crank them both, no tricks, to a 3d03 of 6k exactly, the NV will win like ALL games. Just seems so ATi friendly. That's for another day though.... rivatuner <> Nvidia thats all.

[P_1]

ARGH seems like my sound storm chip on my nforce 2 has died from this... this is what the burning smell was....

[Kunaak]

seems like this thread just killed all benchmarking for me...

now people are gonna be doing LOD stuff with ATI cards, and already are from the looks of the post you made in the 3d area...

are we really so desperate for a high number, that we won't benchmark without making everything look like it was made for an atari??

[P_1]

seems like this thread just killed all benchmarking for me...

now people are gonna be doing LOD stuff with ATI cards, and already are from the looks of the post you made in the 3d area...

are we really so desperate for a high number, that we won't benchmark without making everything look like it was made for an atari??
fugger said it couldnt be done on ati so i tried to prove him wrong. and the tweaks i did didnt even show anything on the monitor, fraps did not even show at all with my tweaks on. Infact these damn tweaks killed my damn audio(even restoring the backup of my registry wouldnt bring it back) GG i have been owned. :(
edit:: heres a pic of what happens when i try to reinstall the nforce drivers
http://img19.echo.cx/img19/7976/bsod6nb.jpg

[P_1]

anyway i still havent solved the audio problem but i unpublished the scores from orb and deleted the tweaks due to some requests.

[Remij]

seems like this thread just killed all benchmarking for me...

now people are gonna be doing LOD stuff with ATI cards, and already are from the looks of the post you made in the 3d area...

are we really so desperate for a high number, that we won't benchmark without making everything look like it was made for an atari??

Exactly...whats the point of benchmarking when the stuff on your screen looks like ****?

I dont care how fast **** flys on my screen...but rather how pretty it looks while flying.

We need some sort of benchmarking software that analyzes just the card...and outputs a beautiful scene on the screen...the faster the card...the more that is output and the better the scene looks. The program would automatically lock all aspects of the driver so everyone would be benching the same settings.

Scores would be based on how many verticies generated among other effects that are being used...

It would be nice to see people that really overclock their PC's, get the recognition they deserve by a benchmark that would make the GPU clocks and processor the only variables in the equation.

[Crankster]

Yeah well, now the Nvidia bashing might stop for good.
Good work uncovering the truth P_1.

[Fr3ak]

Yes nice work P_1.
But..
the thing with benchmarks is that you want to compare system a with system b in terms of speed. But to do so the 'object' (3DMark) which is used for testing has to be equal on both systems to determine which system is faster and its not the same with those tweaks :/

I hope you got what I was going to say. Its like comparing a quarter mile race between a car driving on the street and the other one on the gras..

[Crankster]

Now both are on the street.

[P_1]

Yes nice work P_1.
But..
the thing with benchmarks is that you want to compare system a with system b in terms of speed. But to do so the 'object' (3DMark) which is used for testing has to be equal on both systems to determine which system is faster and its not the same with those tweaks :/

I hope you got what I was going to say. Its like comparing a quarter mile race between a car driving on the street and the other one on the gras..

yes i understand but i was merely trying to show that ati can use tweaks like nvidia and in some cases the tweaks are even worse than what nvidia is capable of. I was being sarcastic when i said i broke the world record for 9800s and that it is legit.

[Fr3ak]

Yes, I already got that before, but I somehow had to say what I was thinking :)

[Charles Wirth]

The average joe who looks at the top scores and compares the two products does not see a real representation of two equil cards when checking the ORB.

I think the benchmarks need to be fixed to some sort of image quality standard.

[Nasgul]

I'm certainly very content with the performance of my PIC-E GeForce 6600 GT, even though I used to own an AGP GeForce 6800 GT but not big difference as far as I can see.

Here's my 3DMark 05 score: http://service.futuremark.com/servlet/Index?pageid=/orb/projectdetails&projectType=12&projectId=849357

[saaya]

i think that the 3dmark series benches already have a decent feature, the screenshot image quality test... the thing is nobody uses it...

imo the benchmarks should capture a certain frame on each benchmark run, always the same frame. capture 2 frames for each game test and then show those frames on the orb when you submit the benchmark run.

then everybody can compare the image quality of different drivers and cards easily!
its all possible, its just a matter of interest, and ati and nvidia unfortunately dont seem to push this hard enough, and futuremark wont move until it really sees a demand for it i guess :/

[MaxxxRacer]

saaya i really like that idea. it would serioulsy help out with the benchmark cheating that has been going on over the years.

[WildKard]

imo the benchmarks should capture a certain frame on each benchmark run, always the same frame. capture 2 frames for each game test and then show those frames on the orb when you submit the benchmark run.




Best idea in ages...you should win a "great idea" trophy! It'll definately push both card makers to deliver top notch image quality at the speeds they "say" their cards can do :)

[brandinb]

lol they will "optomize" the drivers for the two frames and the rest of the frames for speed lol

haha j/k tho that is actually a very good idea and would make benchmark racing much more fun and would make it better to compete because artifact free runs require more oc skill then black screens that regurgitate a score or garbage scattered on the screen which im sure every top score on the orb consists of
________
Oregon Dispensary (http://oregon.dispensaries.org/)

[_Eduard_]

nope because then the new cheats will include prerendered frames to use as comparison, or maybe even bmp's or something :)

[wittekakker]

How can u use prerenderd frames if 3d mark is making the screenshot in-bench?

[Evil_Spork]

How can u use prerenderd frames if 3d mark is making the screenshot in-bench?
the cheat will throw in a perfect looking picture of the frame when the time comes. to trick 3dmark into making it look good.

i have no idea how this would work but i think thats what he means.

[Trice]

Who would give FM the Terrabytes theyd need to host all those pics? lol

Wish they would do something though.