WE’VE ALREADY examined DOOM 3’s performance on the newest generation of high-end graphics cards from ATI and NVIDIA, but we’re not done with the game just yet. While DOOM 3 is obviously most comfortable running on the latest high-end graphics hardware (especially NVIDIA’s GeForce 6800 series) the vast majority of gamers are equipped with older and more mainstream graphics cards. These cards may not have the features or horsepower necessary to churn out playable frame rates with the resolution and detail level cranked, but they’ll still run the game. The question is how well.
To find the answer, I’ve assembled an eclectic collection of 17 low-end, mid-range, and former high-end graphics cards and run them through a gauntlet of DOOM 3 performance and image quality tests. I scoured the Benchmarking Sweatshop for cards, and with a little digging and dusting, I was able to come up with two GeForce4 Titaniums, two Radeon 8500 series cards, five GeForce FXs, six Radeon 9000-series cards, one Dustbuster, and even a Parhelia. How do these cards compare in DOOM 3? Read on to find out.
A note on the testing
Because this comparison includes so many different cards, there’s a wide range graphics horsepower available. To segment things a little, I split the cards into two groups like so:
| Group one | Group two |
| GeForce FX 5200 64MB GeForce FX 5200 Ultra 128MB GeForce4 Ti 4200 128MB Radeon 8500LE 64MB Radeon 8500 128MB Radeon 9000 Pro 64MB Radeon 9550 128MB Parhelia-512 128MB |
GeForce FX 5600 Ultra 128MB GeForce FX 5700 Ultra GDDR3 128MB GeForce FX 5800 Ultra 128MB GeForce FX 5900 XT 128MB GeForce4 Ti 4600 128MB Radeon 9500 Pro 128MB Radeon 9600 Pro 128MB Radeon 9600XT 128MB Radeon 9700 Pro 128MB |
The first set of cards was tested at Medium Quality and resolutions up to 1024×768. On these cards, the game just isn’t playable at higher resolutions or detail levels. The second group of cards has a little more power, so they were tested at Medium Quality at resolutions up to 1280×1024. This second group was also tested at Medium Quality at resolutions up to 1024×768 with 4X antialiasing, and in High Quality mode.
To gain a better understanding of DOOM 3’s image quality modes, you’ll want to read id programmer Robert Duffy’s July 26 .plan update. Duffy explains the differences between the Medium and High Quality modes we used for testing and why things look the way they do. Perhaps the most important thing to take away from that .plan update is that DOOM 3’s High Quality mode uses 8X anisotropic filtering.
A shot from our single-player DOOM 3 demo
High Quality mode, of course
One more thing: although I’ve managed to include 17 different graphics cards in this comparison, this is by no means a complete collection of low-end, mid-range, and former high-end graphics cards. I’m limited by the stock in the Benchmarking Sweatshop, so if you don’t see a specific card tested, it’s because I don’t have one on hand. I do have a couple of DeltaChrome cards in-house, but S3’s current drivers don’t work with DOOM 3, so I couldn’t include the S8 and S4 Pro in the comparison, either.
Our testing methods
All cards were left at their driver default settings for image quality, with the exception that we turned off vertical refresh sync.
Our test system was configured like so:
| Processor | Athlon 64 3200+ 2.0GHz |
| Front-side bus | HT 16-bit/800MHz downstream HT 16-bit/800MHz upstream |
| Motherboard | Abit KV8-MAX3 |
| North bridge | VIA K8T800 |
| South bridge | VIA VT8237 |
| Chipset driver | VIA Hyperion 4.51 |
| Memory size | 1024MB (2 DIMMs) |
| Memory type | Corsair XMS3500 DDR SDRAM at 400MHz and 2-7-3-3 timings |
| Graphics | GeForce FX 5200 64MB GeForce FX 5200 Ultra 128MB GeForce FX 5600 Ultra 128MB GeForce FX 5700 Ultra GDDR3 128MB GeForce FX 5800 Ultra 128MB GeForce FX 5900 XT 128MB GeForce4 Ti 4200 128MB GeForce4 Ti 4600 128MB Radeon 8500LE 64MB Radeon 8500 128MB Radeon 9000 Pro 64MB Radeon 9550 128MB Radeon 9500 Pro 128MB Radeon 9600 Pro 128MB Radeon 9600XT 128MB Radeon 9700 Pro 128MB Parhelia-512 128MB |
| Graphics driver | ForceWare 61.77 |
| Storage |
Western Digital WD360GD 10,000RPM Serial ATA hard drive |
| Operating System | Windows XP Professional Service Pack 1 and DirectX 9.0b |
We used NVIDIA’s ForceWare 61.77 drivers for all of the GeForce cards, and we used ATI’s CATALYST 4.9 beta drivers, intended for DOOM 3, with all the Radeon cards. Matrox’s 107.00.089 drivers were used for Parhelia.
Because this is a graphics performance comparison, I went with an Athlon 64 system with plenty of RAM to make our testing as graphics-bound as possible.
The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 75Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.
We used a custom-recorded demo for testing. You can download it here. If you have questions about our methods, hit our forums to talk with us about them.
Medium Quality
We’ll kick things off with Medium Quality mode where all 17 of our cards can play, at least up to 1024×768. I’ve presented scores with both bar and line graphs, so I’ll hold off commenting until the bottom of the page.
After years of ridicule, the Dustbuster is finally vindicated in DOOM 3—not that the card’s obscenely loud cooler will let you enjoy the game’s audio content. The single-slot GeForce FX 5900 XT is quieter, cooler, and nearly as fast. Heck, it even beats out the venerable Radeon 9700 Pro.
After the 9700 Pro, the cards really start to bunch up, with the GeForce FX 5700 Ultra, GeForce FX 5600 Ultra, Radeon 9500 Pro, and Radeon 9600 XT sharing similar scores. The Radeon 9600 Pro and GeForce4 Ti 4600 aren’t far off the pace, though. Curiously, the two perform almost identically here.
Scores for our slower cards are considerably more varied, with the classic GeForce Ti 4200 leading the way. The GeForce FX 5200 Ultra and Radeon 9550 aren’t too far behind, although I wouldn’t want to play at resolutions above 800×600.
Moving along, the vanilla GeForce FX 5200 manages to edge out ATI’s Radeon 8500 series cards and the Radeon 9000 Pro, which in turn give Parhelia a healthy beating. I was impressed that Parhelia actually ran DOOM 3 without issue, but 22.6 frames per second at 640×480 is hardly playable.
Medium Quality + 4X antialiasing
Next up we have Medium Quality results with 4X antialiasing. The slower cards don’t have the power to handle Medium Quality with antialiasing or DOOM 3’s High Quality mode, so they’ve been shelved for the rest of our performance testing.
The GeForce FX 5800 Ultra and 5900 XT are all over the rest of the pack with Medium Quality and 4X antialiasing. The Radeon 9700 Pro doesn’t do too poorly here, but it’s not much faster than the GeForce FX 5700 Ultra.
Interestingly, the Radeon 9600 XT and 9500 Pro are essentially tied, as are the Radeon 9600 Pro and GeForce4 Ti 4600. The Ti 4600’s performance does drop off more dramatically at higher resolutions, though.
High Quality
Although Medium Quality with antialiasing nicely cuts down on jagged edges, especially at lower resolutions, I far prefer the visual goodness that is DOOM 3’s High Quality mode. High Quality mode’s 8X anisotropic filtering does wonders, and the visuals are truly stunning.
The Dustbuster takes a dive in High Quality mode, relinquishing the lead to the GeForce FX 5900 XT. Even the Radeon 9700 Pro pulls out ahead of the 5800 Ultra at higher resolutions, although it’s still a few frames per second behind the 5900 XT. The good news is that all three cards are butter-smooth up to 1024×768.
Our second pack of cards is led by the Radeon 9600 XT and GeForce FX 5700 Ultra, which essentially tie. The GeForce FX 5600 Ultra, Radeon 9500 Pro, and Radeon 9600 Pro are all pretty close, too, although the 9500 Pro’s performance improves as we turn up the resolution.
As for the GeForce4 Ti 4600, well, the results speak for themselves. The card is really only playable in High Quality mode at 640×480, and even then, the average frame rate is under 40 frames per second. Heck, the Ti 4600 isn’t even half as fast as its closest competition, the Radeon 9600 Pro.
Image quality 1 – Medium quality
We usually like to take our screenshots from a game with a minimum of image processing, but DOOM 3’s gamma and brightness settings don’t seem to affect the output of its screen dump routine. As a result, I’ve done a slight gamma correction to the images that follow, consistent across all of the images, in order to make them easier to see.
Because the DOOM 3 guy’s flashlight moves around a little, even when the player is stationary, the shots below aren’t exactly identical. We’re just checking screenshots to make sure there’s no funny business going on with any of the cards, though.
Since we’ll be covering the image quality of our faster cards with Medium quality plus 4X antialiasing and High Quality, I’ve left them out of our Medium Quality screenshot comparison.
You can click on each image to open an uncompressed PNG in a separate window.
GeForce4 Ti 4200
GeForce FX 5200 (shares GPU with GeForce FX 5200 Ultra)
Parhelia
Radeon 8500 (shares GPU with Radeon 8500LE)
Radeon 9000 Pro
Radeon 9550
Although it’s impressive that the card even runs the game, Parhelia doesn’t seem to be drawing this particular scene properly. Notice how the hand, left wall, and girders aren’t lit properly. Also, Parhelia doesn’t seem to be drawing the floor tile’s bump map properly. To be fair, the ridges in the floor tile don’t look right on the GeForce4 Ti 4200, either.
It’s also interesting to note that the Radeon 9000 Pro, 8500 series, and Ti 4200 appear to have slightly sharper texture filtering than their DirectX 9-class successors. The difference is minor at best, though.
Image quality 1 – Medium quality + 4X antialiasing
GeForce4 Ti 4600
GeForce FX 5600 Ultra
GeForce FX 5700 Ultra
GeForce FX 5800 Ultra
GeForce FX 5900 XT
Radeon 9500 Pro (shares GPU with Radeon 9700 Pro)
Radeon 9600 Pro (shares GPU with Radeon 9600 XT)
The GeForce4 Ti 4600 appears to suffer from the same bump map depth problem as the Ti 4200. Other than that, and the fact that ATI’s gamma-corrected SMOOTHVISION generally looks better than NVIDIA’s antialiasing, image quality is pretty consistent across all cards.
Image quality 1 – High quality
GeForce4 Ti 4600
GeForce FX 5600 Ultra
GeForce FX 5700 Ultra
GeForce FX 5800 Ultra
GeForce FX 5900 XT
Radeon 9500 Pro (shares GPU with Radeon 9700 Pro)
Radeon 9600 Pro (shares GPU with Radeon 9600 XT)
All the cards look pretty close in High Quality mode. Although the 5800 Ultra’s floor tiles appear to be more well-lit than the others, that may be an artifact of the extra sparks in the scene.
Image quality 2 – High Quality
This next image is a little more complex. You can see the amazing rock walls of this room. As I said, I suspect normal maps (which are just fancy bump maps) are in use on the walls.
GeForce4 Ti 4200
GeForce FX 5200 (shares GPU with GeForce FX 5200 Ultra)
Parhelia
Radeon 8500 (shares GPU with Radeon 8500LE)
Radeon 9000 Pro
Radeon 9550
Beyond Parhelia’s obvious problems with this scene, all our low-end cards seem to be rendering things properly.
Image quality 2 – Medium quality + 4X antialiasing
GeForce4 Ti 4600
GeForce FX 5600 Ultra
GeForce FX 5700 Ultra
GeForce FX 5800 Ultra
GeForce FX 5900 XT
Radeon 9500 Pro (shares GPU with Radeon 9700 Pro)
Radeon 9600 Pro (shares GPU with Radeon 9600 XT)
Everything looks good with Medium Quality and 4X antialiasing, too.
Image quality 2 – High quality
GeForce4 Ti 4600
GeForce FX 5600 Ultra
GeForce FX 5700 Ultra
GeForce FX 5800 Ultra
GeForce FX 5900 XT
Radeon 9500 Pro (shares GPU with Radeon 9700 Pro)
Radeon 9600 Pro (shares GPU with Radeon 9600 XT)
Everything looks good in High Quality, too.
Conclusions
DOOM 3 will no doubt drive a wave of hardware upgrades among hard-core gamers who want to enjoy the game at the highest resolutions and detail levels, but it’s comforting to know the game is playable on lesser graphics cards. Medium Quality doesn’t look all that bad, either, although there are definite perks to bumping up to High Quality.
DOOM 3 will no doubt drive a wave of hardware upgrades among hard-core gamers who want to enjoy the game at the highest resolutions and detail levels, but it’s comforting to know the game is playable on lesser graphics cards. Medium Quality doesn’t look all that bad, either, although there are definite perks to bumping up to High Quality.
I don’t want to spend a lot of time summarizing the performance of each of the 17 of the cards we’ve looked at today, but there are a few things worth noting. For starters, I have to single out the GeForce4 Ti 4200. Although it’s two generations behind the curve, it’s quite capable of running DOOM 3 in Medium Quality, and it’s significantly faster than the Radeon 8500, which comes from the same era. Forget about High Quality, though. Even the GeForce4 Ti 4600 choked when we turned up the detail levels.
Among the current generation of low-end cards, the Radeon 9550 stands out as the best choice for DOOM 3, but only because it’s half the price of a GeForce FX 5200 Ultra. The Ultra is faster, but not by anything approaching a big enough margin to justify its significantly higher price tag.
Moving to our assortment of current mid-range graphics cards, the GeForce FX 5900 XT is the clear choice for DOOM 3. ATI’s Radeon 9600s just don’t have it, and not even the eight-pipe Radeon 9500 Pro can bridge the gap. The only cards that give the 5900 XT a run for its money are the former DX9 flagships, the Radeon 9700 Pro and GeForce FX 5800 Ultra. The cards are pretty close in High Quality mode, but the 9700 Pro stumbles a little with Medium Quality and 4X antialiasing.
And there you have it: 17 low-end, mid-range, and former high-end graphics cards in DOOM 3. Now that I have all these cards dug out and dusted, they’re warmed up and ready for Half-Life 2.
I am running Doom3 with my Gigabyte Radeon 9600XT with
core/memory 500/600 respectively, and I run it in High Quality 800×600.
The picture is excellent and performance – NOT A GLITCH
System:
512MB pc 3200 ddr 400(dual channel)
AMD 2600+ barton
Asus A7V600 400FSB
60gig 7200rpm PATA hard disk
Windows XP
Why I even tried running the game in Ultra High mode and performance has’nt changed very significantly for my system.
Why is my card not mentioned in this review much?
Or is it above the ones mentioned?
Just a question
Thanks for the hard work Geoff. The benches of the high-end cards were meaningless until this.
Speaking of the high end, I can’t believe how limited the availability still is. I wander if Best Buy, Comp Usa, etc don’t have the cards in stock because of availability problems, or if they figure that so few will purchase them that stocking them is a waste of time. I’m figuring availability is the issue. Is it just me or has this been one of the most paper-launched years to date? (p4 @ 3.4, 3.6; 6800 series to a while too) Anyway, since we can see that the old cards are doin decent, who cares!
BTW, its amazing how the 5x00s handle this game while sucking (relatively) at so much else. I chose the 9800pro over the 5900xt, and am still glad I did. Still, Nvidia needed the game a year ago.
Am I rambling? Oops.
Im surprised no one has picked it up but the High quality 5800U Image quality 1 – High quality screenie looks antialised look at the hand compaed to the other cards it is really smooth like the medium quality 4x AA screenies
Fixed.
On another note, I’m glad I snatched up a 9500 Pro before they were replaced with the 9600’s.
So you know, the rock walls are definitely normal mapped, along with about 99% of the textures in the game. I’ve been having some fun looking through the pak files.
Page 10 and 11 at the bottom say: q[<"Everything looks good in High Quality, too"<]q but only page 11 should be in HQg{<...<}g
Your demo is flawed and doesn’t show an accurate represenatation about how these cards will play this game. I have a 3.2ghz P4 with a Geforce FX5900 128 and I ran your demo on medium quality 640×480 and got 107 FPS. WOW that is great! However, when I run demo1 (included in Doom3), I get a mere 53 FPS. That’s HALF of what your demo is recording. Why? Because demo1 has more monsters and more dynamic lights, and it’s more representative of what the whole game is like. Plus, the player doesn’t spend half the time staring at a blank wall and switching weapons. I would strongly recommend to people that they not buy this game thinking they’re going to get a consistent 100+ FPS on any of the cards shown here.
Most “custom” demos (like CRUSHER in the old days) were designed to be MORE action packed and take more system resources than the default ones. What reason is there to use a demo of lesser quality?
What reason is there to stop in and take it out on an innocent webadmin who noone else seems to mind.
Just because you have an Intel heatfarm doesn’t mean we all have to suffer.
You can’t compare different levels, because there are different textures, animations, AI, monsters, lighting loaded each level. Also, reviewers must create their own levels or we see situations where the vid card manufacturer’s just optimize for a specific demo, such as, say, demo1g{<.<}g Just for reference my trdemo2 came out to 96.8 and my demo1 came out to 60.4
Well said.
21.8 for demo1 and 37.5 for trdemo2. When in game my performance most closely matches the trdemo2 scores for my average fps (in game is a little higher). It depends on the level of course but there is no way in heck I average 20fps in real game play.
So there is no definitive way to test it, I would say it all averages out.
*Also I am flat out amazed at my score, I can take a ss to back it up if anyone wants proof. TR got 37.7 (ti4600). Only .2 off. What really makes it strange is the fact I have my 2500 running at stock for the moment with 1gig pc2100 ram. A far cry from an a64 3200+…
Did you ever run crusher? that was an absolute WORST CASE scenario. As in, I want to punish this computer because I like Q2, so I will load up the crusher demo. It was not representative of gameplay, it was there so you knew what the lowest FPS would be. Basically, it set the floor for your system. You knew you would never go below the crusher score.
OTOH, the other demo by 3fingers was pretty good. I don’t remember the name of it right now, but it was pretty representative of actual gameplay. Then again, he had bots to play against. A lot easier to get a stressful demo when you can throw in 30 other players…
I was disappointed that they left out the 9800Pro. It can easily be considered a mid-range card at its current price. You can get it for about the same as a Gainward or MSI 5900XT (~$190USD). I am looking to upgrade my aging AIW 8500DV 64MB, and would have liked to see the 9800Pro pitted up against this lineup of cards.
I’m wondering, is a 8500 64MB @ 285/300 better than a 9000 Pro 128MB @ 300/300? This review has got me thinking that I should install the 9000 until I upgrade later on to a x800 or 6800.
I have almost this exact system, (shuttle on an Nforce3 instead of via,) and my FPS benches differ from TR’s by .2! amazing. I think, at least to meg{<.<}g Good work!
Ok, I’m a dumbass. I put the downloaded trdemo into the base folder, but ‘timedemo trdemo2’ doesn’t do anything.
You need to create a “demos” folder:
§[<https://techreport.com/forums/viewtopic.php?t=20100<]§
Dustbuster’s AF looks mighty ineffective in those high-quality shots. The floor gets to be very blurry as you look down the corridor.
Thank you thank you thank you.
I’ve been waiting for this kind of article.
AntecRep
Glad to see that the 9600 Pro I recently bought will be adequate. Anyone care to speculate on how well it will run on a GF3 Ti200 card (or has anyone actually tried it)?
I’m running it on a GF3 Ti200 right now, and it’s playable at 640×480 at low quality. Any higher than that and it becomes tough to play. “timedemo demo1” at those settings is somewhere around 25 fps.
I must say I have a Geforce 4 ti4400 128 ram.. And I was able to run the Ultra quality at 800×600 at 25fps… Though I did do the optimiations in the ini files that helps with performance first.
Diss,
On page 2, the link to the TR timedemo is broken.
Fixed.
Doesn’t seem to be fixed… §[<https://techreport.com/etc/2004q3/doom3-midlow/<]§\etc\2004q3\doom3\trdemo2.zip what an odd url...
works for me, try refreshing the page
nope, still doesn’t work. emptied cache and all. but after trying another browser it worked… i guess when you empty the cache it doesn’t actually empty the cache. anyway, i got it so i’m happy. thanks!
didn’t for me either
EDIT: looks like it’s working now
Doesn’t work for me. Cleared everything out, even tried it on another computer and browser.
O.K. had to get a 3rd computer into the mix and it’s working for that one.
The weird link only works with Internet Explorer. It’s not compatible with other browsers.
Dunno why we had a jacked up link there so long. Sorry about that… it’s fixed.
§[<https://techreport.com/etc/2004q3/doom3/trdemo2.zip<]§
Good my ti4200 can run it. Now all I need is a demo.
Well that just justified getting my fiancee’s copy of Doom 3 installed on her 8500LE 128MB card @ 285/310.
I just wanna say a big “WOOT” for my 5900XT! 🙂
Damn, I bet that’s the first WHOOT! you’ve given it since it’s purchase.
Actually I’ve been very happy with the performance of the card. After reading the horror stories of how bad the FX series was, I was expecting the worst. Turns out it’s a great overclocker and provides great FPS for UT2K4, FarCry, and D3. Sure I can’t play all games at 1600 x 1200 w/ 4xAA and 8XAF, but I wasn’t expecting to for $191. I got the MSI version with VIVO. I am a big video editor and it was a big plus to avoid having a buy a new video capture card. (I recently sold my 4 year old Pinnacle DC10+ PCI card)
q[
I think that was there so that the graphics cards weren’t limited by the CPU throughput…
True, it would have been nice to see a lower end system in addition to this one, but that’s a lot of benchies to ask for :-P,
-Mole
I don’t think it matters, at least not for the CPU. In fact, let me provide some numbers to back that up. I have a Ti4200 128MB card. I ran trdemo2 on my own system and got these numbers:
800×600 with medium detail: 44 FPS
1024×768 with medium detail: 30 FPS
Those line up nicely with the numbers in this review. I’m also running on this system:
AMD Athlon XP 2100+ TBred B CPU running at 2.17 GHz
1024 MB Corsair CAS2 PC3200 DDR RAM
of course it does. your running your athlon at 3200+ speeds.
No, it’s running at 2700+. This is a T-bred 2100+. Quite different then the A64 benched in the review.
Well I’m sure many users will ask themselves if they should get a CPU upgrade or a video card upgrade.
From the looks of the benchmakrs if you own a Geforce 4 Ti or higher and a Radeon 9500 Pro or higher then a CPU upgrade would be worth it.
If you own a Radeon 8500 or 9000 then it’s time to upgrade. Since I myself own a Radeon 8500 I’m wondering if it’s really just the drivers.
Considering an Athlon Mobile 2500 can be bought for $88 at newegg and easily overclocked beyond a Athlon XP 3200 many on these video card owners probably only really need a new CPU.
My Radeon 8500 system has an Athlon XP 2200 with 512 megs of ram. All on a Leadtek Nforce2 motherboard with SoundStorm.
After a few driver updates and bitching if ATI doesn’t improve performance for 8500/9000 owners I’m just going to get a new video card. I’m going to be sad replacing my AIW Radeon 8500 128. I really love the video features.
Alas, that’s the problem with the high-end All-In-Wonder cards. The 3D graphics part is obsolete long before the video tuner part is obsolete. If you still want a TV tuner in your gaming PC, consider a seperate tuner card.
I think this article shows that. If you have one of these cards and what you have right now for a current CPU. You can see what a AMD64 will do to boose performance. For instance, I have a AthlonXP 3200+ and a 9700Pro. Right now I get 30-49 FPS in Doom 3 at 1024×768@med quality. These benchmarks show me that if I had a AMD64 3200+ (2.0Ghz) I would be getting around 90FPS. It shows me that by upgrading I would be getting these speeds.
I was up at arms for my upgrade. I was thinking about getting either a AMD64 3200+ and a VIA KT800, or a X800PE. Now I know whta my chose is.
I can’t believe you’re getting such low numbers on your system. How can the A64 (with the same rating as your CPU) be getting double the FPS as your current setup? That just doesn’t make any sense.
Are you running the trdemo2? Or are you looking at FPS counter in the game? Are you running the same video drivers as used in the review?
Something is wrong. With that system you should be getting at least 50 fps at that res, if not moreg{<.<}g
I’d be much more complacent with the results if my Ti4600 hadn’t died last week after 27 months of service. 🙁
Farewell card. We knew ye well.
I’m curious why the 128MB 5900XT wasn’t used instead of the 256MB version. The 5900XT included in the benchmark was the only 256MB version in group 2. I gotta think that 90% of the 5900XT’s out there are 128MB. Any ideas how the 128MB version would perform against the 9700Pro?
Exactly the same. 128 vs. 256MB of memory doesn’t start affecting performance (in Doom 3) until 1600×1200 AA+AF high quality (much sooner at ultra quality, but that wasn’t tested here). Look for some 6800 non-Ultra D3 reviews if you want to see for yourself.
I’m suprised that you couldn’t scare up a Radeon 9800 Pro in the benchmarking sweatshop. This is the card that is priced closest to GeForceFX 5900XT and 5700 Ultra, not Radeon 9600XT.
Radeon 9700 Pro still looks pretty good for a card that first appeared on store shelves in August 2002.
Agreed, the Radeon 9800 Pro 128MB is still the best card for its price, it definetly should’ve been in there.
I agree too, 9800pro is good dealing. Look 9700pro better running on doom 3, well 9800Pro is best board, can running high res with high-quality, or low res with high quality.
9600XT was $100 cheaper than the 9800 pro on release and it runs the game great, includes the elusive vaporware HL2 as well, or so they sayg{<.<}g
My 5900 XT has already been WOOTed many times the fact it has clocked to 500/850 is absolutly amazing on stock cooling & infact i have nor found a game where my xp 2200 does not limit it
I’ve been fooling around with some of the settings in the doomconfig.cfg for my Radeon 8500 system and found some interesting stuff.
For my Radeon 9500 Pro I use these settings to increase image quality while in medium quality mode.
seta image_downSizeLimit “1024”
seta image_downSizeBumpLimit “1024”
seta image_downSizeBump “1”
seta image_downSize “1”
The default is 256 for both down size.
As I went from 256 to 512 to 1024 the image quality increased a lot as the performance decreased. Even when you start Doom3 the loading screen looks a lot sharper just from going to 512.
Now I did the same thing for my Radeon 8500 and had different results. While the image quality did in fact increase, the performance didn’t decrease at all.
Timedemo1 for 256 was 26 FPS. Timedemo1 for 512 was 26 FPS. Timedemo 1 for 1024 was 24 FPS. All tests were done twice.
Only 1024 actually got a peformance decrease and only by 2 FPS. I tried switching the doomconfig.cfg files between my 9500 Pro PC and my 8500 PC to make sure nothing was wrong. Of course I saw a major image quality increase with 1024 so the settings are correct. Even 512 looks amazing.
Something doesn’t seem right.
Yeah, stats for my 9500Pro. 🙂
Great job.
Yes finally graphics cards that are commonly owned.
I would love to see the 5600u and ti4600 era cards in more benches but that’s prob asking for a lot…
Id say the ti’s do decent job for being decrepit. Glad to see 2 year old tech (at least I am pretty sure the ti4600 was released in 2002..) is still viable. Definitely not for high rez and with aa on but what can you expect.
Yep, I got my Ti4400 in April 2002, which was when it came out in the UK …
How about a few nv1x cards?
I find it most interesting that in High Quality the Radeon 9700 Pro seems to do comparatively better (against it’s competition) than in Medium Quality. If I had to guess I would say it’s because of the Aniso. Must cost a pretty penny more on the Dustbuster. Given these scores though I wonder what would’ve happened if Doom3 came out in early-mid 2003? In time with the Dustbuster launch, if memory serves.
Finally a benchmark of graphics cards that people actually own. I’m sick of seeing X800 and 6800 benchmarks.
I find it strange how horrible the Radeon 8500 performs in Doom 3 compared to the Geforce 4 Ti 4200. Just shows you how long it’s been since ATI really touched OpenGL.
I own a Radeon 7500, 8500, and 9500 Pro. I know first hand how quickly ATI forgets it’s older products. I’m sure R3X0 owners would be lucky to see Doom 3 improvements in future drivers.
Wasn’t the 8500 a good match for the GeForce3 only? Comparing it to the GF4 is a little unfair in my opinion, given that the card got beaten by the gf4 since the first day.
I knew that my Ti4200 was a good buy.
Picked up a 5900 XT 8 months ago. Enjoying Doom 3 now and very glad I made a good choice. 😀
I’m a proud owner of the fastest “low end” card 🙂
I can overclock it well too… (my 4200)
*yawn*… Doom3 is like so /[
That’s true however, these benches are good indicators of what to except with future games based on the D3 engine. 😉
hmm, i’ve seen screen shots of areas towards the end of the game where the beast in the walls looks ‘wet’ on dx9 hardware, yet it doesn’t look that way on my 4200. and yes i was running the same quality as the shot in question.
it appears my card just can’t render that effect so its left out.
is there any chance that the 8500 (r200) core chip may be able to render more effects than the geforce4 range? saying that because r200 supports dx8.1 and the gf4ti doesnt