![[H]ardOCP](/images/blank.gif){kind=small}
- Date:
- Wednesday, July 10, 2002
- Author:
- Steve Lynch
- Editor:
- Sean Quinn
- Google +1
MSI GeForce4 Ti4200
In a sea of budget video cards, the Ti4200 is the big fish we are all looking to land. But with so many fish in the ocean, what makes one Ti4200 different from another?? Which is worthy of your hard earned cash?? MSI thinks they know...
Test Systems:
MSI Ti4200 System: ABIT KR7A-RAID / VIA KT266A, AMD Athlon XP 1800+, CoolerMaster HCC-001, 2x Corsair 256MB XMS3000 DDR SDRAM, 3Com 905CX-TXNM 10/100 PCI, Western Digital 30GB ATA/100 7200RPM HDD, Windows XP Professional, DirectX 8.1, 4n1 4.38.
Gainward / Reference System: EPoX 8K7A-RAID / VIA KT266A, AMD Athlon XP 1800+, Alpha PAL8045, 1x Corsair 256MB XMS3000 DDR SDRAM, 3Com 905CX-TXNM 10/100 PCI, Western Digital 40GB ATA/100 7200RPM HDD, Windows XP Professional, DirectX 8.1, 4n1 4.38.
The cards and configurations used in benchmarking are as follows:
MSI MS-8870 GeForce4 Ti4200 64MB - Clocked at default 250MHz / 500MHz (core / memory) driver version 28.32
Gainward Ti 4200 128MB - Clocked at default 250MHz / 444MHz (core / memory) driver version 28.32
NVIDIA Reference Ti4200 64MB - Clocked at default 250MHz / 500MHz (core / memory) driver version 28.32
Benchmarks:
This is where the rubber meets the road. We've amassed a collection of industry standard programs and the latest games that we feel test the product's stability and overall performance. While features and bundled software are nice, it is the raw performance you'll see here that will help you decide if this is the card for you or not. There is a certain percentage of folks out there who might say "Raw FPS does not make a great videocard alone". While we agree to a certain extent, it is sheer performance that makes a card's "features" enjoyable. After all, all the pixel shaders and programmable lighting effects in the world don't mean jack running at 14FPS. Let's get on with the tests.
3DMark2001 SE
3DMark 2001 SE is a DirectX 8 based benchmark that is very similar to the previous 3DMark 2000, except that it uses some totally different DirectX features that put much more stress on the VidCard. Default mode at 1024x768x32 was used. You will see the scores bunch up much more tightly here. While memory bandwidth is still an issue, if you were to see this benchmark alone, you would certainly be left with a different perception of our systems. 3DMark 2001 has moved very close to being a VidCard benchmark instead of a system benchmark. The 3DMark 2001 will surely stay an industry standard, as it utilizes the Vertex and Pixel Shader programming, but it has lost some of its usefulness as a total benchmark.
Looking at the 3DMark 2001SE scores, the three cards turned in almost identical results, with both of the 64MB cards edging out the 128MB card. This is the faster (500MHz) memory talking here, proving that more memory isn't always better.
Jedi Knight 2:
Star Wars Jedi Knight 2: Jedi Outcast is a newly released OpenGL game that many have been waiting for. It has much improved graphics over its predecessor. It fully supports advanced shaders, as well as very high texture resolutions and effects. There is one demo included in the multi-player section that is good for benchmarking use, JK2FFA.
Screen: On, Geometric Detail: High, Texture Detail: Very High, Texture Quality: 32bit, Texture Filter: Trilinear, Detailed Shaders: On.
Interesting enough, I had expected both the 64MB cards to finish with close scores and the 128MB card to walk away with this one because of the large textures and extremely intensive nature of Jedi Knight II, only to find the MSI card to hold on to the lead over the other boards.
Serious Sam: The Second Encounter
We are using Serious Sam: The Second Encounter for benchmarking. All testing was done using "Maximum Quality" settings. The Elephant Atrium Demo at 1024 x 768 32bit was used here for benchmarking purposes.
The MSI card easily handled the Serious Sam benchmark, even at Maximum Settings. Performance on all cards is almost identical with all cards able to push almost 90fps at 1600 x 1200. Very impressive.
Comanche 4:
Novalogic recently released the Comanche 4 Demo with built in benchmarking capabilities and DirectX 8.1 support. We will be using the Comanche4 Benchmark as a standard part of our graphics card benchmarking suite. The full demo can be downloaded here so you can see how your system compares. Default configuration at 1024 x 768 32bit color settings.
The Comanche 4 benchmark demo is a unique benchmark as it represents a real-world gaming experience. It contains the single player "Eagle's Talon" mission from the game as well as a detailed cinematic. This DirectX 8.1 benchmark demo will measure your system's performance in the standard frames per second format.
Flip a coin, the scores you see here are far too close to call. Something that might be easily overlooked here is the fantastic results put out by the Ti4200 across the board on this benchmark. Lower core speeds do not seem to be holding the Ti4200 back in the slightest.
Code Creatures:
Code Creatures is another new DX8.1 based benchmark that shows off the power of vertex and pixel shaders. It's a nature scene much like the ATI nature demo and the nature scene in 3DMark2001SE. It takes you through some rolling hills with grass that waves in the wind and water that reflects like a mirror with fish underneath. It's incredibly beautiful for a real time 3D rendered scene. I used the "Official Benchmark" option and ran the default benchmark run. The benchmark produces three FPS results in three resolutions.
Again, more of the same. No clear leader here even though the MSI card does eek out a small lead.
Quake3
We are using patch version 1.17 so that you can compare with all of the past reviews here at the [H]. These benchmarks are taken using Quake 3 Arena version 1.17, Max settings at 1024 x 768, and Demo001 using the QBench script which can be found at G256.com.
Over 100FPS at 1600 x 1200 is an amazing feat for any card, let alone a 64MB "so-called" budget video card.