| Thread Tools |
10th September 2008, 22:51 | #1 |
Madshrimp Join Date: May 2002 Location: 7090/Belgium
Posts: 79,021
| Tropics 3D Benchmarks - A real GPU Load Test Tropics Benchmark Features: * Support of DirectX 9, DirectX 10 and OpenGL * Dynamic sky with light scattering * Live water with a surf zone and caustics * Special materials for vegetation * HDR rendering * Parallel split shadow map * Depth of field * Real-time ambient occlusion * Benchmark mode Software requirements: * DirectX (August 2008 or newer) * .NET framework 2.0 (Windows version only, optional) * OpenAL (included into Windows installer) * Latest stable video drivers Hardware requirements: * ATI Radeon X700 and higher or NVIDIA GeForce 6XXX and higher (recommended: NVIDIA 8800 / AMD 3800) * 256 Mb of video memory http://unigine.com/download/#Tropics make sure to download DirectX distro ( http://www.microsoft.com/downloads/d...DisplayLang=en ) - One Global Score: xxxx easy to rank - Looks impressive - result screen shows GPU and CPU make/model - result screen shows settings in detail - 60mb download (+80mb DX August) - Works with DX9 We need to do some testing, but this can be a very nice addition to HWbot, Suggested setting: 1280x1024 DX9 Fullscreen --- on C2D @ 3.5Ghz only 14-15% CPU usage at 1024x768 DX9 windowed mode (default setting thus) 9800 GTX 512Mb XP 32bit C2D@3GHzParallax, Translucence and DOF remain disabled under DX10 too, no options to enable them either? 9600 GT 1024Mb 700Mhz GPUThis would be the first 3D benchmark which is not CPU bound it would be a shame not have it as a global benchmark for all to enjoy. Really sweet that OCing CPU 1ghz makes no difference; and reacts nicely to GPU OC! ------------ Anybody up for some testing at 1280x1024 DX9 0XAA Fullscreen, looking for more CPU scaling info for duallies and quads, as well as Crossfire and SLI scaling and GPU OC performance. With enough bench data we can then move forward and make this a refreshing addition to HWbot
__________________ Last edited by jmke : 10th September 2008 at 23:06. |
11th September 2008, 00:04 | #2 |
Posts: n/a
| It will be a nice idea to put this benchmark on hwbot ! |
11th September 2008, 12:01 | #3 |
[M] Reviewer Join Date: Nov 2004 Location: Waregem
Posts: 6,466
| Yeah, I like the idea of adding a new DX10 benchmark. DX9 is old-skool, time to push the DX10 cards |
11th September 2008, 12:06 | #4 |
Madshrimp Join Date: May 2002 Location: 7090/Belgium
Posts: 79,021
| ge gaat ook ni beginnen als dien hopeloze fin op hwbot hopelijk? anyway; if somebody with multi GPU could test scaling in DX9/DX10 that would be swell; under DX9 (and Vista) one or two GPU don't make a difference; is under DX10 it does make a difference there is a reason to go DX10. if it doesn't scale under DX10 either, we can throw this benchie away
__________________ |
11th September 2008, 13:36 | #5 |
[M] Reviewer Join Date: Nov 2004 Location: Waregem
Posts: 6,466
| I think I was the first one to really vote for DX10? |
11th September 2008, 14:08 | #6 |
Madshrimp Join Date: May 2002 Location: 7090/Belgium
Posts: 79,021
| doesn't matter much who voted what as long as we don't have SLI/CF scaling numbers from DX9 vs DX10
__________________ |
12th September 2008, 13:44 | #7 |
Posts: n/a
| This benchmark will appear on the bot ?? |
12th September 2008, 14:07 | #8 |
Madshrimp Join Date: May 2002 Location: 7090/Belgium
Posts: 79,021
| maybe, need more testing; currently looking for somebody who can test SLI scaling in Vista using DX9/DX10 render path, with original name as well as rename .exe to 3DMark03.exe
__________________ |
12th September 2008, 21:21 | #9 |
[M] Reviewer Join Date: Nov 2004
Posts: 4,127
| 317 oink? HD3870, E6600 3GHz... because of 16xAF? |
12th September 2008, 21:25 | #10 |
Madshrimp Join Date: May 2002 Location: 7090/Belgium
Posts: 79,021
| what did you test and how? what OS what setting DX9 or DX10 noAA/AF please!
__________________ |
Thread Tools | |
| |