PDA

View Full Version : Single and SLI results on 7600GT@850/2016 (G73-B1 1.1ns)



S_A_V
04-09-2007, 04:41 AM
Testing configuration:
Single XFX 7600GT XXX Edition @ 850(796)/2016 (Vgpu=1.69V, Vmem=2.22V, delta=-50)
Core2Duo E6600 (L630A) @ 9x440=3966MHz 1.53V for 3DMark01-03-05 and AM3
Core2Quad Q6600 (L640A) @ 9x345=3106MHz 1.40V for 3DMark06
Asus P5N32-E SLI Plus (PCI-E@133MHz)
2x512Mb Micron D9GKX DDR2-1000 @ 880MHz 3-3-3-4-1T 2.60V
PSU Tt Toughpower 750W

Cooling: +4°c water at CPU and both GPU's
http://img470.imageshack.us/img470/974/coldwater4cyl9.th.jpg (http://img470.imageshack.us/my.php?image=coldwater4cyl9.jpg)

ForceWare v93.81 used because older drivers (including FM-approved for GeForce7 series) not works in SLI mode with nforce 650i SLI chipset :(

Single card results:

AM3: 150447 (http://www.hwbot.org/result.do?resultId=594094)
http://img292.imageshack.us/img292/4950/full150447benchmark2007ah9.th.jpg (http://img292.imageshack.us/my.php?image=full150447benchmark2007ah9.jpg)

01: 51969 (http://service.futuremark.com/compare?2k1=9175811)
http://img69.imageshack.us/img69/4035/51969singlexfx7600gtxxxjj1.th.png (http://img69.imageshack.us/my.php?image=51969singlexfx7600gtxxxjj1.png)

03: 22591 (http://service.futuremark.com/compare?2k3=5287229)
http://img262.imageshack.us/img262/3315/22591singlexfx7600gtxxxaa2.th.png (http://img262.imageshack.us/my.php?image=22591singlexfx7600gtxxxaa2.png)

05: 9670 (http://service.futuremark.com/compare?3dm05=2860832)
http://img262.imageshack.us/img262/5554/9670singlexfx7600gtxxx8iv0.th.png (http://img262.imageshack.us/my.php?image=9670singlexfx7600gtxxx8iv0.png)

06: 5425 (http://service.futuremark.com/compare?3dm06=1588984)
http://img237.imageshack.us/img237/9294/5425singlexfx7600gtxxx8pv1.th.png (http://img237.imageshack.us/my.php?image=5425singlexfx7600gtxxx8pv1.png)

Results in SLI mode:

AM3: 196609 (http://www.hwbot.org/result.do?resultId=595095)
http://img261.imageshack.us/img261/7088/info196609benchmark2007wy8.th.png (http://img261.imageshack.us/my.php?image=info196609benchmark2007wy8.png)

01: 62784 (http://service.futuremark.com/compare?2k1=9175765)
http://img262.imageshack.us/img262/8271/62784slixfx7600gtxxx850oy1.th.png (http://img262.imageshack.us/my.php?image=62784slixfx7600gtxxx850oy1.png)

03: 37818 (http://service.futuremark.com/compare?2k3=5287069)
http://img176.imageshack.us/img176/78/37818slixfx7600gtxxx850es2.th.png (http://img176.imageshack.us/my.php?image=37818slixfx7600gtxxx850es2.png)

05: 17256 (http://service.futuremark.com/compare?3dm05=2860444)
http://img261.imageshack.us/img261/4666/17256slixfx7600gtxxx850iz2.th.png (http://img261.imageshack.us/my.php?image=17256slixfx7600gtxxx850iz2.png)

06: 9633 (http://service.futuremark.com/compare?3dm06=1589012)
http://img265.imageshack.us/img265/1853/9633slixfx7600gtxxx8507fj1.th.png (http://img265.imageshack.us/my.php?image=9633slixfx7600gtxxx8507fj1.png)

And few pics of the rig:
http://img214.imageshack.us/img214/3593/017600gtsliwatercooledlx4.th.jpg (http://img214.imageshack.us/my.php?image=017600gtsliwatercooledlx4.jpg) http://img294.imageshack.us/img294/7493/027600gtsliwatercooledrh5.th.jpg (http://img294.imageshack.us/my.php?image=027600gtsliwatercooledrh5.jpg)
http://img294.imageshack.us/img294/6374/037600gtsliwatercoolednw0.th.jpg (http://img294.imageshack.us/my.php?image=037600gtsliwatercoolednw0.jpg) http://img294.imageshack.us/img294/179/047600gtsliwatercooledkl7.th.jpg (http://img294.imageshack.us/my.php?image=047600gtsliwatercooledkl7.jpg)
http://img214.imageshack.us/img214/1852/057600gtsliwatercooledow4.th.jpg (http://img214.imageshack.us/my.php?image=057600gtsliwatercooledow4.jpg)

Only one run in each benchmark, and no way to improve scores - both cards and Q6600 returned back to pc parts shop where I got it only for 2 days for quick benching.

Fr3ak
04-09-2007, 05:23 AM
Nice scores, but I have the feeling thats not the max of that setup :) There seems to be some more potential.

Nice to see some pics of your setup. Are those flexible copper pies and why did you choose them over pvc or whatever tubing?

What kind of chiller do you use? And what pump? I was tempted to get one myself, but I have been told its troublesome.

Almost 200k in AQ3 :) Good look reading that mark with 7600GTs.
What CPU and GPU temp do you have with a wartertemp of 4C? I am running my E6600 with 1.6V set in bios for benching on non chilled water, which is why I am wondering why you didnt use a higher vcore, especially with chilled water.

Anyway. Great work S_A_V! As usual :D
:toast:

JT from Russia
04-09-2007, 07:06 AM
Nice 7600GT clock! ;)
Keep on!
:toast:

S_A_V
04-09-2007, 07:47 AM
Fr3ak:

I have the feeling thats not the max of that setup There seems to be some more potential.
975X chipset would perform better, especially in 3DMark01, but I only have nforce650i based mobo. And only 345 MHz FSB - max stable for Q6600 on non-vgtlref modded mobo. Also I not tried higher voltage on GPUs, may be something like 1.85V can help to reach 900+ MHz.

Are those flexible copper pies and why did you choose them over pvc or whatever tubing?
It's not copper, this is standard (stock) tubes of Aucma Cool River watercooling system (http://itc.ua/img/dpk/2006/01/012718.jpg), it just looks very old and dirty after 2 years of intensive usage :)

What kind of chiller do you use? And what pump? I was tempted to get one myself, but I have been told its troublesome.
No chiller, no compressor and no pump. See this pic:
http://img267.imageshack.us/img267/7139/watersourceoe8.th.jpg (http://img267.imageshack.us/my.php?image=watersourceoe8.jpg)
I just connect my watercooling system to it and that's all :)
Water not circled and always have stable temp at ~+4c late night.

What CPU and GPU temp do you have with a wartertemp of 4C?
P5N32-E SLI Plus BIOS show cpu temp as +10c:
http://img402.imageshack.us/img402/6804/10ccputempinbiosidlepa7.jpg
In Windows CPU temp (CoreTemp) is +17...+20c idle and GPU temp (RivaTuner) is +21c...+24c idle. I don't measured exactly loaded temps, but while running 3dmarks it's about +35c...+40c on CPU and GPUs.

I am running my E6600 with 1.6V set in bios for benching on non chilled water, which is why I am wondering why you didnt use a higher vcore, especially with chilled water.
I tried 1.6V Vcore and got more temp on CPU but the same clocks :(

Anyway. Great work S_A_V! As usual
Thanks :)

JT from Russia:
I got this pair of 7600GT specially for Titan Cup 4 competition (http://www.titancup.ru/) and hope to see your scores too ;)

SF3D
04-09-2007, 09:25 AM
Nice cooling method :) Good scores!

I have offtopic question for you: What lod setting you are using in 3DM03 with 8800GTX. I have similar setup now, but can't score as much as you are.

Fr3ak
04-09-2007, 09:44 AM
Ahh thats how it works! :D
Great! Too bad the cold water is much warmer here. Otherwise I would give it a try :)

NickS
04-09-2007, 09:53 AM
Nice result for a pair of 76's :)

S_A_V
04-10-2007, 04:05 AM
SF3D:

What lod setting you are using in 3DM03 with 8800GTX. I have similar setup now, but can't score as much as you are.
lod +15.
I used higher shader clock flashed to bios for 3dmark03 run. Compare my score with standard shader clock 720(1674) MHz 46107 (http://img201.imageshack.us/img201/4813/46107axle8800gtx768mb72ml5.png) and with 720(1890) MHz 46945 (http://img180.imageshack.us/img180/1658/46945axle8800gtx768mb72qk5.png).

SF3D
04-10-2007, 02:38 PM
Thanks S_A_V. That LOD setting doesn't work for me :)

Our scores are identical with default shader speed.