NVIDIA GeForce 6600GT and 6600 (NV43): Part 1

Test results: performance comparison

We used the following test applications:

Return to Castle Wolfenstein (MultiPlayer) (id Software/Activision) – OpenGL, multitexturing, ixbt0703-demo, test settings – all to maximum, S3TC OFF, you can download configurations from here
Serious Sam: The Second Encounter v.1.05 (Croteam/GodGames) – OpenGL, multitexturing, ixbt0703-demo, test settings: quality, S3TC OFF
Quake3 Arena v.1.17 (id Software/Activision) – OpenGL, multitexturing, ixbt0703-demo, all test settings to maximum: detailing level – High, texture details – #4, S3TC OFF, curved surfaces are strongly smoothed using variables r_subdivisions «1» and r_lodCurveError «30000» (note that by default r_lodCurveError «250» !), the configurations can be downloaded here
Unreal Tournament 2003 v.2225 (Digital Extreme/Epic Games) – Direct3D, Vertex Shaders, Hardware T&L, Dot3, cube texturing, default quality
Code Creatures Benchmark Pro (CodeCult) – gaming test demonstrating how the video card works with DirectX 8.1, Shaders, HW T&L.
Unreal II: The Awakening (Legend Ent./Epic Games) – Direct3D, Vertex Shaders, Hardware T&L, Dot3, cube texturing, default quality
RightMark 3D v.0.4 (one of game scenes) – DirectX 8.1, Dot3, cube texturing, shadow buffers, vertex and pixel shaders (1.1, 1.4).
Tomb Raider: Angel of Darkness v.49 (Core Design/Eldos Software) – DirectX 9.0, Paris5_4 demo. The tests were conducted with the quality set to maximum, only Depth of Fields PS20 was disabled.
HALO: Combat Evolved (Microsoft) – Direct3D, Vertex/Pixel Shaders 1.1/2.0, Hardware T&L, maximum quality
Half-Life2 (Valve/Sierra) – DirectX 9.0, demo (ixbt07. The tests were carried out with enabled anisotropic filtering as well as in heavy mode with AA and anisotropy.
Tom Clancy's Splinter Cell v.1.2b (UbiSoft) – Direct3D, Vertex/Pixel Shaders 1.1/2.0, Hardware T&L, Very High settings; demo 1_1_2_Tbilisi
Call of Duty (MultiPlayer) (Infinity Ward/Activision) – OpenGL, multitexturing, ixbt0104demo, test settings – maximum, S3TC ON
FarCry 1.1 (Crytek/UbiSoft), DirectX 9.0, multitexturing, demo01 (research) (the game was started with the -DEVMODE option), test settings – Very High.
DOOM III (id Software/Activision), OpenGL, multitexturing, ixbt1-demo (33MB) test settings – High Quality. Configuration files were created with caching in order to optimize the game and to reduce the jerks. You can download them as well as a batch file to launch the benchmark here (35K)
3DMark03 v.340 (FutureMark/Remedy), DirectX 8.1/9.0, multitexturing; Game1/2/3/4, MARKS.

If you want to get the demo-benchmarks, which we use, contact me at my e-mail.

Quake3 Arena

GeForce 6600GT vs. OTHER CARDS	GeForce 6600GT vs. GeForce 6800/6800GT	GeForce 6600

The easiest modes without AA and anisotropy: 6600GT easily outscored its cheaper competitors (X600XT and PCX5900) but it lost the battle with GeForce 6800/6800GT at even frequencies (later on, I'll call these cards "emulators"). This is really interesting. Ok, in case of 6800 (AGP) we can write it off to platform differences, but why does it lose to 6800GT? Larger memory capacity in the latter? - For Quake3 it does not matter, as is well known. Now let's recall the synthetic tests, where we demonstrated that because of the memory bandwidth limitation the processor does not write all 8 processed pixels at once – at first it writes 4, and then, when the next textures are processed, it writes the remaining 4. Sort of caching and buffering to use memory more rationally with a limited bus. There is hardly any influence of core differences on this test. 6600 behaved like a regular trump, though it lost a little to its competitors X600XT/PCX5900 (but they are much more expensive! And they are positioned differently). This card is a successor to PCX5750, and it deals successfully with the task.

With enabled AA: 128-bit bus in 6600GT has its effect! Catastrophic loss both to R9800PRO and 6800LE! These cards possess the 256-bit bus and more memory controllers. We can also see a serious blow to this product and its complete fiasco dealt by its competitors X600XT/PCX5900, though X600XT is considered a cheaper video card. It's not clear why this card with greater memory bandwidth (6600GT has 128bit*1000 MHz) is defeated by its competitor with a lesser memory bandwidth (X600XT has 128bit*760 MHz). The number of pipelines is also smaller. In general, it's a strange result, perhaps it can be written off to buggy drivers (Quake3 is infamous by such things). Approximately the same picture is with both 6600 and 6600GT in comparison with "emulators". On the whole, the AA speed in this game is horribly slow in NV43.

With enabled anisotropy: Here the picture is mostly the opposite. And we see 6600GT succeed almost in all the fronts, except for the battle with "emulators". Here is the effect of 8 pipelines and the frequency of 500 MHz. But why the defeat to 6800/6800GT... Perhaps these cards have some caches or units, which were cut down in 6600 (see the synthetic tests section). This may be the effect of a larger number of memory controllers, despite the reduction of memory frequencies (we must admit that a 256-bit bus has an advantage over a 128-bit bus even with equal memory bandwidth). Though the memory bandwidth is less important for anisotropy than the core capacity. So, 6600 acts perfectly! It outscored even more expensive competitors: X600XT and PCX5900!

The final heaviest mode with AA and anisotropy: It's clear that the general defeat of 6600GT/6600 almost to all competitors is due to the loss in AA.

Thus, on the whole:

GeForce 6600GT versus RADEON 9800 PRO (ATHLON64 3400+) – defeat (due to AA)
GeForce 6600GT versus GeForce PCX5900 – the same picture
GeForce 6600GT versus RADEON X600XT – the same picture
GeForce 6600GT versus GeForce 6800LE (ATHLON64 3400+) – also a defeat (because of the bus)
GeForce 6600GT 325/700 MHz versus GeForce 6800 (8/3) 325/350 MHz (ATHLON64 3400+) – the same picture
GeForce 6600GT 350/1000 MHz versus GeForce 6800GT PCX (8/3) 350/500 MHz – the same picture
GeForce 6600 300/600 MHz versus GeForce PCX5900 – I think that we must take into account the victories in all modes except for AA here, especially when the competitor has a higher price
GeForce 6600 300/600 MHz versus GeForce PCX5750 – excellent success
GeForce 6600 300/600 MHz versus RADEON X600XT – the only victory is with anisotropy, but considering the other defeats and the fact that these cards will hardly be manufactured with 600 MHz memory - general defeat

Pay attention to the performance gain from 6600GT overclocking – almost directly proportional to frequency increase (18%).

[ The previous part (3) ]

[ The next part (5) ]

Andrey Vorobiev (anvakams@ixbt.com)
Alexander Medvedev (unclesam@ixbt.com)

07 September, 2004

Write a comment below. No registration needed!