ATI X1900XTX vs ATI X1800XT [Archive] - Page 2

View Full Version : ATI X1900XTX vs ATI X1800XT

Pages : 1 [2] 3 4 5

OWA

01-30-06, 11:59 PM

Thats old info - it's been out since last week. It is driver related.

6.2 should fix it and offer a performance increase for X1900 owners :D
I don't appear to have the 3DMark06 issue with the drivers I'm using (the ones from the CD).

The only thing I see is that one is slightly darker than the other.

Riptide

01-31-06, 12:20 AM

I can't wait for the VF900 to come out. I want to see how well it cools these X1800/1900 series cards. :)

Skinner

01-31-06, 05:39 PM

There seems to be problem with Crossfire AA as well, NVIDIA SLI has much better picture (from FiringSquad.com):

http://radekhulan.cz/img/1/20060130-kvalita-ati-nvidia.jpg

http://radekhulan.cz/img/1/20060130-ati-nvidia-2.jpg

Do you have the link to this review. I cann't find it on fsquad.

eva2000

02-01-06, 02:48 AM

thanks for the comparison .. still using my HIS X1800XT right now :)

OWA

02-01-06, 03:04 AM

No problem. I'll hopefully get a chance to redo it soon. I ordered a X2 4800+ for that system on tuesday and should get it by thursday. I also want to test higher resolutions.

Henno

02-01-06, 01:33 PM

Do you have the link to this review. I cann't find it on fsquad.

http://www.firingsquad.com/hardware/ati_super_aa_vs_nvidia_sli_aa/page3.asp

If you compare Nv 8xAA vs Nv 16xAA you'll see the same happening, the doors in the 16xAA image are much sharper. Ati 8xAA vs Nv 8xAA is more or less the same sharpness.
What is happening in the 16xAA Nv image that makes the doors more sharp? Different Lod?

OWA

02-01-06, 01:37 PM

Thanks for the link. I tried finding it at Firingsquad also but couldn't find the right link.

Henno

02-01-06, 02:16 PM

This is what i'm trying to say, Nv8x/Ati8x looks both equally sharp. It's the 16xAA Nvidia shot that looks sharper. What is causing this?

http://members.chello.nl/h.appel1/NvAti8x.jpg

Nv 16xAA vs Ati 14xAA:
http://radekhulan.cz/img/1/20060130-ati-nvidia-2.jpg

Riptide

02-01-06, 03:12 PM

Sheesh 16x AA? What are people running in 800x600? Seems like a waste.

Skinner

02-01-06, 04:37 PM

http://www.firingsquad.com/hardware/ati_super_aa_vs_nvidia_sli_aa/page3.asp

If you compare Nv 8xAA vs Nv 16xAA you'll see the same happening, the doors in the 16xAA image are much sharper. Ati 8xAA vs Nv 8xAA is more or less the same sharpness.
What is happening in the 16xAA Nv image that makes the doors more sharp? Different Lod?

Thanx, yes, you would say lod is responsible for this.

OWA

02-03-06, 07:44 PM

I've got to leave shortly but I wanted to run one game benchmark with the new X2 4800+ before leaving.

Old result with AMD 4000+ at 2.58 and XTX at 689/797
FEAR 4xAA/16xAF, no soft-shadows
30/65/173 (0/28/72)

New result with X2 4800+ at 2.58 and XTX at 689/797
33/68/184 (0/19/81)

Higher resolutions and more benches will have to wait until I get back. I did post some 3DMark scores in the benchmark threads though. X2 mainly helped in 3DMark06.

Dazz

02-03-06, 08:34 PM

As you can see your still CPU limited the extra core is kicking in but only slightly increases the score intill the drivers mature for dual core. However the fps in the above 45fps is alot higher then the X1800XT. Also try it with soft shadows and see what it can do.

OWA

02-06-06, 09:18 PM

I've kind of lost track of where I was in comparing the X1800XT to the X1900XTX. I know I needed to do some higher resolution tests so those are below. Things are slightly different now though so I'm only listing my new results. Most of the X1800XT results I obtained, I posted in this thread:

http://www.nvnews.net/vbulletin/showthread.php?t=59592

One difference now, that kind of makes my results even different than when I started this thread, is that I'm now using a X2 4800+ with the X1900XTX. Most of the results I made with the X1800XT were with a single core 4000+. So, with all that in mind, here are some results with the following conditions:

X2 4800+ at 2.58ghz, using drivers that came on the CD, Cat AI on, HQ AF off, adaptive AA off, XTX at 689/797.

FEAR 1600x1200, 4xAA/16xAF
26/52/125 (0%/45%/55%)

Half-Life 2: Lost Coast 1600x1200, 4xAA/16xAF
timedemo lost_coast: 55.24fps
video stress test: 76.22

Doom 3 1600x1200 4xAA/8xAF Ultra Quality
66.2fps

Counter-Strike:Source 1600x1200 4xAA/16xAF
170.71fps

Highest 3DMark05 so far: 11,567

Highest 3DMark06 so far: 5746

I plan to run all those again with the Adaptive AA on and HQ AF on.

OWA

02-06-06, 09:40 PM

Here are the results with Cat AI off, HQ AF and Adaptive AA on.

X2 4800+ at 2.58ghz, using drivers that came on the CD, XTX at 689/797.

FEAR 1600x1200, 4xAA/16xAF
24/47/119 (1%/48%/51%)

Half-Life 2: Lost Coast 1600x1200, 4xAA/16xAF
timedemo lost_coast: 43.73fps
video stress test: 58.45fps

Doom 3 1600x1200 4xAA/8xAF Ultra Quality
56.5fps

Counter-Strike:Source 1600x1200 4xAA/16xAF
136.53fps

Using the Cat 6.2 betas, I got 11,859 (http://service.futuremark.com/compare?3dm05=1783485) in 3DMark05 and got 6085 (http://service.futuremark.com/compare?3dm06=111217) in 3DMark06.

Edit:

After installing the Cat 6.2 betas, here are the results.

X2 4800+ at 2.58ghz, using the beta Cat 6.2s, Cat AI on, HQ AF off, adaptive AA off, XTX at 689/797.

FEAR 1600x1200, 4xAA/16xAF
26/53/129 (0%/43%/57%)

Half-Life 2: Lost Coast 1600x1200, 4xAA/16xAF
timedemo lost_coast: 55.84fps
video stress test: 79.50fps

Doom 3 1600x1200 4xAA/8xAF Ultra Quality
69.0fps

Counter-Strike:Source 1600x1200 4xAA/16xAF
173.07fps

fatal1ty

02-06-06, 11:45 PM

OWA

02-06-06, 11:52 PM

Here are the results with Cat AI off, HQ AF and Adaptive AA on.

X2 4800+ at 2.58ghz, using the cat 6.2 betas, XTX at 689/797.

FEAR 1600x1200, 4xAA/16xAF
24/48/123 (3%/47%/50%)

Half-Life 2: Lost Coast 1600x1200, 4xAA/16xAF
timedemo lost_coast: 44.83fps
video stress test: 61.29fps

Doom 3 1600x1200 4xAA/8xAF Ultra Quality
57.4fps

Counter-Strike:Source 1600x1200 4xAA/16xAF
141.38

vs the results when using the drivers on the CD

FEAR 1600x1200, 4xAA/16xAF
24/47/119 (1%/48%/51%)

Half-Life 2: Lost Coast 1600x1200, 4xAA/16xAF
timedemo lost_coast: 43.73fps
video stress test: 58.45fps

Doom 3 1600x1200 4xAA/8xAF Ultra Quality
56.5fps

Counter-Strike:Source 1600x1200 4xAA/16xAF
136.53fps

Sazar

02-06-06, 11:54 PM

I guess they are working their way round the marchitecture much like the 1800xt.

OWA

02-08-06, 03:18 AM

I decided to continue the comparison by adding the 7800s to the mix (since I no longer have the X1800XT). The settings used are high quality settings with the optimizations disabled since that is what I use when gaming. For ATI, I disabled the HQ AF though since I was trying to make the settings as similiar as possible between the systems. The CPUs were both set to 2.58ghz but the memory timings on the Nvidia system are slightly tighter. Hopefully I'll be able to get both running the same mem settings soon. Also, I'll probably redo the tests again just as a double-check since the results were somewhat surprising (to me anyway) and I'll try to do optimizations on as well. I compare single vs single (and singles versus SLI just because I was curious to know the results of that). Cards at stock settings but bear in mind that the BFG 7800GTX 512 is factory overclocked.

One thing I did find is that gaming at 1920x1200 with a single card is more plausible than I thought.

http://img147.imageshack.us/img147/4626/xtxvsgtxsummary6jz.jpg (http://imageshack.us)

fatal1ty

02-08-06, 01:24 PM

I decided to continue the comparison by adding the 7800s to the mix (since I no longer have the X1800XT). The settings used are high quality settings with the optimizations disabled since that is what I use when gaming. For ATI, I disabled the HQ AF though since I was trying to make the settings as similiar as possible between the systems. The CPUs were both set to 2.58ghz but the memory timings on the Nvidia system are slightly tighter. Hopefully I'll be able to get both running the same mem settings soon. Also, I'll probably redo the tests again just as a double-check since the results were somewhat surprising (to me anyway) and I'll try to do optimizations on as well. I compare single vs single (and singles versus SLI just because I was curious to know the results of that). Cards at stock settings but bear in mind that the BFG 7800GTX 512 is factory overclocked.

One thing I did find is that gaming at 1920x1200 with a single card is more plausible than I thought.

You really need to replace the stock ATI cooler \ idle = 57.... damn! I'm using the Zalman VF-700 and idle is 36, load 63.

OWA

02-08-06, 01:40 PM

You really need to replace the stock ATI cooler \ idle = 57.... damn! I'm using the Zalman VF-700 and idle is 36, load 63.
Yeah, I hope to get one soon. I got my first Zalman recently but for my CPU. I really like it. It does a pretty good job and it's quiet even with the fan as high as it'll go. I'm not real fond of the idea of replacing the XTX HSF since I'm afraid I'll screw something up but maybe it's worth the risk. Hopefully it's easy. I had to remove the motherboard to install the CPU one but the actual mounting on the CPU was incredibly easy. If it works out well, I'll probably start using their coolers on everything.

Riptide

02-08-06, 01:40 PM

fatal1ty

02-08-06, 01:50 PM

Yeah, I hope to get one soon. I got my first Zalman recently but for my CPU. I really like it. It does a pretty good job and it's quiet even with the fan as high as it'll go. I'm not real fond of the idea of replacing the XTX HSF since I'm afraid I'll screw something up but maybe it's worth the risk. Hopefully it's easy. I had to remove the motherboard to install the CPU one but the actual mounting on the CPU was incredibly easy. If it works out well, I'll probably start using their coolers on everything.

Its like 7 screws - there is nothing to mess up and the VF-700 is like a 10 min install. It's cake!! :)

Funny- about your results, everyone else is seeing the X1900XTX clearly beating the GTX 512 with the new drivers 6.2. I have my old numbers with the GTX 512 and it's not even close the X1900XTX is spanking it!!!

Using the built in timedemo in FEAR (which I know is not truely indicative of real world performance) softshadows off, 1280x960 4xaa/16af my max with the X1900XTX is 345 fps!!! The GTX 512 was in the low 200's. My average was like 94 vs 70. I mean its not even close, the GTX 512 got owned!! My GTX 512 was running 600 / 1.78 and the XTX is 690 / 800 so both were well overclocked.

Here is hexus' results with the new betas...
http://img.hexus.net/v2/graphics_cards/ati/r580xt_xtx/images/bench-fear.png

In fact across the board the XTX is much faster than the GTX 512. In many apps the X1800XT is even faster than a single GTX 512 now.

http://www.hexus.net/content/item.php?item=4477&page=11

Just wanted to point that out. I don't know if you are using full reformats cause the Catalyst drivers usually suck if you install them after say Nvidias have been installed first. I had to clean install everything to really get the most from the drivers.

fatal1ty

02-08-06, 01:53 PM

I'd wait for the VF900. The VF700 lowers idle substantially on the X1800XT (and I'd wager 1900 also) but there is practically zero improvement in load temps. Need something a bit more robust and w/the 900 right around the corner why not wait it out? :)

I have a VF700 cooling mine and I like how it's quieter and looks better. But load temps are hitting almost 80 with it and that's not unusual based on what I've seen over @ [H].

Hell my X1900XTX was loading about 90, my VF-700 with some decent case airflow is max in the 60's! I've got the Lian Li PC V-2000 with the front bay expansion and 120mm fan blowing over the GPU area with the backplates off around the GPU. So the 120 is rushing the hot air around the GPU right out of the back of the Case, anything the VF-700 cannot handle gets exhausted. Its nothing revolutionary just maintaining Case airflow. If you are hitting 80s then the air isn't getting exhausted and you may want to address your cooling within the Case itself. Afterall the X1900XTX stock cooler exhausts the hot air, there is no reason that the VF700 shouldnt at least be given some cool air to accomplish it's task.

The guy from 3DXtreme came up with a great idea I saw in the XS forums...
the Sumbeam anywhere bracket...

http://www.3dxtreme.net/other/ontap/Lian Li PC-V2000/IMG_2789.jpg

http://www.3dxtreme.net/other/ontap/Lian Li PC-V2000/IMG_2793.jpg

That willl probably be my next mod once I get some money together...

Also - I copied his idea with the front fan zip tied to the expansion cage...

http://www.3dxtreme.net/other/ontap/Lian Li PC-V2000/IMG_2794.jpg

Good artists copy, great ones steal!!!

Riptide

02-08-06, 02:09 PM

Thanks for the idea but the last thing I want to do is add another fan. I have a serious hate for fans.

At some point I want to go with a Reserator and a Phantom PS. Then all I will have is my two enermax 120mm case fans. Those guys have been totally trouble-free for me unlike most Zalman fans I've had.

I had a 7000cu that developed squeak after about a year or so. I had a fatal1ty GPU cooler that had some chatter coming from it and this VF700 Cu which I just installed has some bearing noise it's producing.

Basically I'm just sick and tired of fans period and I don't want to go there any more. :)

OWA

02-08-06, 02:34 PM

That system has only had ATI cards in it.

If you look at my FEAR results, the XTX was almost double the single GTX 512 (edit: max fps that is, plus the average is a good deal higher) and the single XTX was close to the SLI results. I'm not sure what you're complaining about there.

I'm also running with all the optimizations disabled and most sites compare ATI Cat AI on versus Nvidia Quality. I'm comparing Cat AI off versus High Quality since that's what I use when gaming. With the Cat AI disabled, it is probably hurting the Doom 3 results for ATI since ATI made up most of the ground with an optimization that is probably being disabled. Don't worry, I plan to do optimizations on tonight.

I was basically trying to compare things using the settings I use when gaming. I wanted to see how it would be if I used those settings with just a single card and still used the Dell LCD's native resolution of 1920x1200.

I can try reinstalling the Cat 6.2 betas but my 3DMark05 and 3DMark06 scores fall right in line with others so I don't think there is anything wrong. One thing that I did that I need to check on was that I forced 16xAF from the CCC and Nvidia CP (except for Doom 3). I usually leave both systems that way since some games don't provide the AF settings or don't provide 16x as an option. AA is almost always available so I set those using in-game settings. I'll play around with that some also to see if it makes a difference.