|
XFX GeForce 6800 GT 256 MB
Test Setup
Due to the release of Doom 3 and the Half-Life 2 Source Engine benchmarking tool, we have revised the list of games that we benchmark with. To test the performance of the different cards, we will be using six popular games: Far Cry, Need for Speed Underground, Splinter Cell, Call of Duty, and as I previously mentioned, Doom 3 and the Half-Life 2 benchmark. These games vary greatly in their use of shader effects and range from almost pure geometry (COD), to games where every surface has some sort of advanced shader effect (Doom 3), and everything in between. The system we are using for benchmarking consists of an ABIT IC7-MAX3 motherboard which utilizes the Intel 875P chipset. An Intel Pentium 4 2.4c with 800 MHz FSB and Hyper-Threading technology is used along with 1 GB of Corsair PC3200 XL Pro memory (2-2-2-5 timings). Every component of the system is at stock or default speeds set by the manufacturer. While this definitely is not a top-notch system, it is sufficient for our purposes and if anything will result in benchmarking scores that the majority of our readers can relate to. Windows XP was installed onto a freshly formatted Seagate Barracuda 160 GB SATA hard drive connected via on-chip SATA. All the latest Windows Updates were downloaded and installed except for Service Pack 2. The latest video drivers at the time of testing were installed - Forceware 61.77 drivers for the NVIDIA cards, Catalyst 4.8 drivers for the ATI cards. In the video card control panel V-SYNC was disabled and the cards were set to "quality" over "performance." For NVIDIA cards, the "Image Settings" slider was set to "High Quality." For the 6800 GT, this automatically disables the Trilinear optimizations. However, the 5950 Ultra does not have the ability to disable the optimization. For the 9800 XT and X800 Pro that means the "Texture Preference" and "Mipmap Detail Level" were both set to "High Quality" while Anti Aliasing and Anisotropic Filtering were custom for each test. Everything else was left to the default settings. Through our testing, we want the cards to display their best image quality as fast as they possibly can. This usually requires a few basic tweaks to the card's settings through the graphics control panel. However, we do not want to go too in depth and start tweaking variables that the average user wouldn't even touch. Our benchmarking results display how well each card performs when displaying the best image quality possible. Because some cards may support optimizations even at the highest quality settings, we compare the quality of the images each card produces on the "Image Quality" page. To ensure the best possible results, we benchmark each game three times for each test and average the results. In most cases, we play a recorded time-demo which prints the average FPS (frames per second). However, some games differ in the way they record and play demos. Additionally, not every game supports demo or time-demo playback. More information and details regarding how each game benchmark is carried out can be found on the respective benchmarking pages. Test Computer: ABIT IC7-MAX3 Intel Pentium 4 2.4c 1 GB Corsair TwinX PC3200 XL Pro (Dual Channels; 400 MHz DDR @ 2-2-2-5) Seagate Barracuda 160 GB SATA hard drive Windows XP Professional SP1 w/ DirectX 9.0c and latest updates When analyzing the performance graphs on the following pages, I will sometimes refer to settings that offer the best gaming experience. To me, a good gaming experience offers the right mixture of speed and image quality. Playing a game at 50 FPS with smooth edges and crisp textures would offer a better overall gaming experience than playing at 100 FPS with "jaggies' and noticeable texture filtering. Personally, I think 40-50 FPS allows for a smooth experience and my comments will stem from that belief. Your opinions may differ.
Page 1: Introduction
Page 2: The Card Page 3: The Bundle Page 4: Test Setup Page 5: Benchmarks: Doom 3 Page 6: Benchmarks: Half-Life 2 Page 7: Benchmarks: Far Cry Page 8: Benchmarks: Need for Speed Underground Page 9: Benchmarks: Splinter Cell Page 10: Benchmarks: Call of Duty Page 11: Image Quality Page 12: Overclocking Page 13: Conclusion 
Related Reviews Video Cards September 23, 2009
Video Cards June 25, 2009
Video Cards June 8, 2009
Video Cards April 6, 2009
Add Comment
8 User Comments 1 - Posted by
Guest
on September 15, 2004 - 1:48 pm It'd be nice if you could include the temperature readings after the overclocking.
2 - Posted by
Guest
on September 16, 2004 - 4:00 pm how about Con: poor AA & AF performance? Or takes up 2 slots, suckin up the nearest PCI slot cuz of the mondo huge fan. Gee, the reason why the Ultra is getting creamed is the need for 2 distinct molex power connectors. Can't wait for PCIexpress. 3 - Posted by
Guest
on October 25, 2004 - 5:17 pm Just bought this card, it is awesome. Runs doom 3 on ultra-high res and had 113.00 fps on the half life 2 stress test. This isn't even overclocked yet!! I highly recommend this card
4 - Posted by
Guest
on December 7, 2004 - 9:38 pm hi, just thought i wud leave a comment saying how wicked this card is, i got it on Nov. 25th and i havent had one day of bad gaming.... realistic, quality gfx and you cant go wrong if you have 300 pounds... multi-display too?! mmmm
5 - Posted by
Guest
on December 25, 2004 - 9:24 am I'm gonna buy him directly (I need a new PC)
6 - Posted by
Guest
on August 31, 2005 - 5:44 pm I find this card in the Dual DVI version (still AGP) in one of my favorite online stores, thinking of buying it. A tad dissapointed that it didn't overclock stable to full 6800 Ultra specs. So I'm thinking of Gainward GeForce 6800Ultra or XFX GeForce 6800Ultra instead. They cost more but not much more then more expensive 6800GT by other brands. On the other hand, watercooling might get better OC and I want to get watercooling to keep noise levels down anyways. Also, is it really worth 100-150$ extra to buy a Ultra instead of a GT to get full Ultra specs (or beyond) when a GT, especially clocked near Ultra speeds, performs almoust as good. A decent graphics card and soundcard, and ofcourse the low noise watercooling, is all I miss in my comp. Currently using integrated sound and GeForce 4 MX440. Thats on a P4 Northwood "C" 3.2GHz with 2048MB primary memory. /GW 7 - Posted by
Guest
on September 18, 2005 - 5:12 pm This review was helpful. I am buying one now off of newegg.com for $290. I like that you can put it in SLI and its PCI-E. I do have a question if the DVI hookups would support 1920x1200 for my Dell 24" widescreen. My analog VGA hookup on my GeForce Fx5600 supports it, but not the DVI on it, so i was wondering if the DVI on this supports it. I dont know if 2048x1536@85Hz pertains to the DVI hookups.
8 - Posted by
Guest
on September 19, 2005 - 6:20 pm lol nice try, i think ill be keeping this monitor:)But thanks i just didnt know if it was a DVI thing or just my current graphics card.
Add Comment To add a comment without being a member, you may omit the password field, but you must enter your name (or nickname) along with your comment. * Denotes required fields. |
Recent Articles Editorials March 15, 2010
Editorials August 27, 2009
Editorials August 18, 2009
Editorials March 17, 2009
Recent News Jason Bradbury Mar. 19, 2010 - 10:52 pm
Rock Paper Shotgun Mar. 19, 2010 - 10:44 pm
joystiq Mar. 19, 2010 - 6:58 pm
Tom's Hardware Mar. 19, 2010 - 4:42 pm
Rock Paper Shotgun Mar. 18, 2010 - 11:58 pm
Fudzilla Mar. 18, 2010 - 11:57 pm
CVG Mar. 18, 2010 - 11:53 pm
Tech Radar Mar. 16, 2010 - 11:58 pm
|
|||||||||||||||
