Sign in with
Sign up | Sign in
Your question

x800 pro, getting "bad" framerates

Last response: in Graphics & Displays
Share
June 24, 2004 7:53:17 AM

Hi all.

I'm new here so be gentle. I'm a rare breed, a female gamer who loves to kick arse in ut2004. After suffering horrid framerates in certain ONS maps on my old GF4, I decded to get myself a Sapphire x800pro (after reading the benchmarks from Tom, how could I resist?).

However, I'm shocked at how badly the card is actually performing, in comparison to Toms benchmarks, on my system. Has anyone else here actually obtained an x800pro yet? I'd be curious to see what sort of frame rates they get.

I was playing crossfire last night, and got avg 70fps in spec mode with no players, near the core. Once the action started, it was averaging at 33 fps - not good enough. Same on any map. Worse on big open maps like dawn - get an average 28 there.

I've tried various levels of AA and filtering but nothing seems to actually improve the frames per second. I couldn't find an option to turn AA off altogether in the graphics card opengl/direect3d settings (it was either 2x or "application controlled").

I'm hoping that the bad performance is due to some old Nvidea driver being left around, so will do a format / reinstall windows tonight. I'll probably reinstall UT too (and I don't have EAX selected in ut because it reputedly has a performance hit)

Nerdy info:

Intel p4 2.8 clocked to about 3.1 (10%)
asus p4p800 motherboard
Sapphire radeon x800pro
Creative audigy player
bog standard ddr333 pc2700 ram (1gig)
OS: WinXP with no service pack.

I know all too well my system isn't as good as the one Tom used in his benchmarks. However it's not insuperior to the point where it should knock 60fps off the benchmarked fps!

Other notes about UT:

I run at 1280 res, 32 bit colour, have tried various detail levels but framerate remains at a constant (low) average.
There's no hardrive thrashing, it just goes quite slow in places.
Avg fps without firefights is 50
avg with firefights is 28-33
June 24, 2004 8:00:06 AM

I know of at least one member here, Speeduk, who HAD one (he voltmodded, aggressively OCed and unlocked all 16 pipes and sold it for a large profit). He had higher frame rates in FarCry...so something is wrong.

When you switched from GF4 to X800P...did you reinstall windows or use some driver deletion programs to FULLY remove all of nVidia's drivers? Also, what Catalyst version do you have?

Maxtor disgraces the six letters that make Matrox.
June 24, 2004 8:10:34 AM

"When you switched from GF4 to X800P...did you reinstall windows or use some driver deletion programs to FULLY remove all of nVidia's drivers? Also, what Catalyst version do you have?"

I plan on formatting tonight. There could well be some drivers left kicking around, though I did uninstall the software and uninstall the card from device manager.

Using the latest cats, 4.5 I think.
Related resources
June 24, 2004 8:14:08 AM

Yeah, remnants of nVidia drivers can really screw up ATi's performance...a clean format should do you wonders (unless you have a near-dead card in addition). Also, CAT 4.6s are out...so are 4.7betas (which have excellent performance).

Maxtor disgraces the six letters that make Matrox.
June 24, 2004 8:18:39 AM

Thanks for the info. I'll try find those beta drivers and install them after I reformat.

I'll let you know if it works.

What's a "near dead" card?
June 24, 2004 8:22:33 AM

A bad card that will still work (albeit barely).

Most cards work perfectly, some are DOA, others are bad (or near-dead as I said). Usually the bad ones are secondhand, however (first owner OCs the crap outta the card). Bad cards also artifact a lot...so I highly doubt you have a bad card.

Maxtor disgraces the six letters that make Matrox.
June 24, 2004 2:29:14 PM

You Cpu 2.8B or 2.8C? 533Mhz FSB or 800Mhz FSB.
June 24, 2004 2:38:23 PM

"You Cpu 2.8B or 2.8C? 533Mhz FSB or 800Mhz FSB."

800mhz.

Just did a format and clean install, latest drivers, clean install of UT.

28 fps min at 1024, 32 bi tcolour.
33 average.

Totally slow in other words..
June 24, 2004 2:41:38 PM

How the heck is your puter running on PC2700?
June 24, 2004 2:43:02 PM

What do you mean?
June 24, 2004 2:47:22 PM

All of the reviews said you needed at least PC3200. Thats if you OCed which I think you said you did.
June 24, 2004 2:47:52 PM

Mem divider probably. Shouldn't make that HUGE of a difference...my 9800Pro on a 3.06B gets higher fps, so something else is wrong.

Maxtor disgraces the six letters that make Matrox.
June 24, 2004 2:49:51 PM

WTF--run 3DMark03 if you can and post your score. You can download it <A HREF="http://www.futuremark.com" target="_new">Here</A>.

Maxtor disgraces the six letters that make Matrox.
June 24, 2004 3:00:04 PM

Latest intel chipset drivers from www.asus.com for my p4p800
Latest audigy drivers
Latest (non beta) cat drivers

dx9b installed as part of ut2004

Will run 3dmark soon
June 24, 2004 3:08:11 PM

I know.. I don't understand it :)  My gf4 was faster.

The format's made no difference at all. I've knocked it down to 1024, and without aa or anything on it's still dropping to 28 when theres action on screen (can be up to 70 with no action in closed off areas).


Just downloading 3dmark now, will be an hour or so.
a b U Graphics card
June 24, 2004 4:31:52 PM

If you are basing it on AVG fps, first thing first, remove Vsync, it's better for gaming mos times, but benchies are usually run without Vsync to allow highest max fps.

Also consider that the game may have defaulted to higher details, viewdistance, etc. based on it's recognition of a more powerful card.

Also on games that don't require the latest feature DX7/8 based games, often you will see little performance difference because they are both running the game without a problem (the GF4ti is a very VERY good DX7/8 card), and then you become limited by your system's abilities to run the game (thus being CPU/memory/etc limited). If you look at most X800/GF6800 benchies, they aren't impressive at 800x600/1024x768, and even sometimes 1280x1024, but it's when he resolution goes higher or the AA/AF are turned up that they show the greatest improvements.

I think getting faster memory would help, but unless the annoyance is equal to the price of some PC3200CL2 stuff, then it's a minor issue, and running asynchronously with the FSB will work, just not as well as DDR400.

Once you post 3Dmk03 (and I would also recommend posting 3Dmk01 results) then we'll get a better idea of how well your system perfroms, and where there may be bottle necks. And yes using the Catalyst 3.6 would be a good idea, and the beta 4.7s 'may' also help. But if it is a system issue, you may be getting all you will ever get regardless of benchies.


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil: 
June 24, 2004 5:21:56 PM

Your 3DMark Score is 5255 3DMarks!

Grapeape-

It's ut2004 being tested. It uses dx9b.
If it is a system issue then why;

1, Are my results so differenty to Toms, when I only have a slightly slower cpu, and slightly slower ram?
2, was my gf 4 faster?
3, Is the lowest 3d mark score on the search 4k higher than mine?

Basically my score is half what it should be and my fps is half what it should be.

Can anyone think what could possibly cause this? A bios setting? A hardware issue?
June 24, 2004 5:28:17 PM

anyone got that link for a fill rate tester? might want to rule out its not a card or driver glitch first.
June 24, 2004 5:58:12 PM

Heh.. ignore my marks up til now.. I'm clearly an idiot.

I'd left 4xaa and 8xaf on. With the card at default settings I get just over 9000 points. Which is better, but still a couple thousand lower than people with similar systems.

Could that be due to my ram?

Also, the reason I had 4xaa and 8xaf on first place, was to compare my frames to the reputed benchmark Tom did of 92 fps (1280, 32bit, 4xaa, 8xaf). I get 30 though, not 92.. really is puzzling me.
June 24, 2004 6:02:30 PM

hehe, yes that could be due to your ram :) 
June 24, 2004 6:49:02 PM

Why are you running 166Mhz memory on a 200Mhz CPU? I would definitely make that your next change to see if you get better results. I would think you'd be crippling the system with that memory. I don't know how huge the difference would be but you seem to have tried everything else.

<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN</font color=blue>
<font color=yellow>SWEDEN SE SWEDEN SWEDEN</font color=yellow>
<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN
June 24, 2004 6:56:26 PM

heh :) Cheers for help so far guys.

I just have one burning question - do you really think that the ram could cause ut to underperform by a whopping 50 or so fps?

I don't want to fork out cash only to have it not remedy the problem. UT is my most played game, ut is what I want to optimize. A constant 70 fps would be lovely, but 30 is a joke with a new "top of the range" card.

What I don't understand is why reducing settings just doesn't improve matters at all. I can reduce the resolution to 1024, I can reduce the in game settings, I can turn aa off. Still, I get fps dropping to 28 in firefights.

Can anyone shed any light?
June 24, 2004 7:28:34 PM

What's your refresh rate set to? If it's 60Hz, that could be your limiting factor. I don't know if UT2004 is OpenGL or DirectX... but if it is DX, you can override your refresh rate settings so that they'll run higher than the default 60Hz.

<font color=red> If you design software that is fool-proof, only a fool will want to use it. </font color=red>
June 24, 2004 7:33:03 PM

60? My brain would melt out of my ears if I tried to play at that refresh! :) 

I play at 1280 which is 100 refresh on my monitor. And turning vysnc off has no noticable effect or improvements..

this is weird!
June 24, 2004 7:48:05 PM

It may be 100Hz on your desktop, but are you absolutely sure it's 100Hz in the game? I have to bring up my monitor's settings through the panel buttons on the monitor to ensure it's running at the correct rate. I have my DX override set to 85Hz since I play some games at 1600 x 1200.

:smile:

<font color=red> If you design software that is fool-proof, only a fool will want to use it. </font color=red>
June 24, 2004 7:53:28 PM

I used to run 2700 ram on my 3.0C, then switched to 3200. Didnt notice any improvement in games as far as framerates are concerned.

I think there may be something wrong with your card. I would RMA it anyway and be sure, especially if i spent as much as you probably did on it.

Good luck!

"Who is General Failure, and why is he reading my drive?"
P4 3.0 HT, Intel D865GBF, 512MB Crucial PC3200 DDR, WD 36GB Raptor 10,000RPM, BBA Radeon 9600PRO, SB Audigy, Hauppage WinTV
a b U Graphics card
June 24, 2004 8:12:17 PM

Intel Pentium C's do lose performance running slower ram Async. You want PC3200 minimum, faster if you OC. SO overclocking the 2.8C to 3.1 GHz and using PC2700 will hurt performance for sure.

Also it depends on the type of ram and bios tweaks (settings like PAT or game enhancement, turbo, etc. and mem timings). I doubt the card is bad. It was the AA/AF.


ABIT IS7, P4 2.6C, 512MB Corsair TwinX PC3200LL, Radeon 9800 Pro, Santa Cruz, TruePower 430watt
June 24, 2004 8:13:43 PM

Yes, 100% sure and have checked the monitor panels. Plus I'd know if it was 60 as its so terribly flickery!

I hope it isn't a problem with the card, how can i prove it to the guys I bought it from if it is?? "it gets half the framerate it should" would probably result in them saying I have another hardware issue aside from the gfx card.
June 24, 2004 8:15:53 PM

Yeah I've taken overclocking off and got 9080 on 3dmark. But WHY is it still doing so poorly on UT? This is the main question. I'd expect it to be 20 fps or so less than Toms benchmark because of my ram, certainly not 50 fps less, which is what I'm getting.
a b U Graphics card
June 24, 2004 8:17:09 PM

Are you still getting 30 fps?


ABIT IS7, P4 2.6C, 512MB Corsair TwinX PC3200LL, Radeon 9800 Pro, Santa Cruz, TruePower 430watt
June 24, 2004 8:22:08 PM

Yeah.. I've tried many things now, and the result is always pretty much the same. 33 fps average, sometimes lower on open maps, 70 in non-player populated places which are not open areas.

I cant get it to go faster no matter how much aa I turn off, no matter how many details I reduce in-game.

Anyway I bet I'm being a pain, it just makes no sense. Calling it a night and gonna go get drunk! :) 
June 24, 2004 8:27:06 PM

If you get the fps on lower res its a cpu problem you have.

Athlon 2700xp+ (oc: 3200xp+ with 200fsb)
Radeon 9800pro (oc: 410/360)
1024mb pc3200 (5-3-3-2)
Asus A7N8X-X
June 24, 2004 8:28:19 PM

A gamer who get's drunk? I love you.

<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN</font color=blue>
<font color=yellow>SWEDEN SE SWEDEN SWEDEN</font color=yellow>
<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN
a b U Graphics card
June 24, 2004 8:28:27 PM

Honestly I suspect that their timedemo may be a whole lot less stressful than when you are actually playing during your tests. HardOCP says 1280x1024 4X/16X is the highest playable settings on an X800 pro / P4 3.4C with 1GB PC3200 and Cat 4.5's. They dropped to 37 fps and average 57 fps.

<A HREF="http://www.hardocp.com/article.html?art=NjI4LDk=" target="_new">http://www.hardocp.com/article.html?art=NjI4LDk=&lt;/A>


ABIT IS7, P4 2.6C, 512MB Corsair TwinX PC3200LL, Radeon 9800 Pro, Santa Cruz, TruePower 430watt
June 24, 2004 8:30:13 PM

Make sure the settings are turned off in the driver and not just in game. The ATI drivers may be overriding your game options. Just go to the 3D tab in the ATI control panel and move the slider over to performance just to be sure.

"Who is General Failure, and why is he reading my drive?"
P4 3.0 HT, Intel D865GBF, 512MB Crucial PC3200 DDR, WD 36GB Raptor 10,000RPM, BBA Radeon 9600PRO, SB Audigy, Hauppage WinTV
June 24, 2004 8:51:49 PM

That's what I feared. But surely a new gfx card should be a lot better than my old gf4? The only current benefit I'm getting is that on certain areas that my gf4 used to grind to a halt in, my new card atleast does 28 fps in. Other than that it plays at about the same framerates!

If I could get an average 57 I'd be happy. But I can't, my average is literally 33, and I think that's too low considering my overall spec.
June 24, 2004 8:59:34 PM

Don't forget the fact that games will look a lot better with the new card.

<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN</font color=blue>
<font color=yellow>SWEDEN SE SWEDEN SWEDEN</font color=yellow>
<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN
a b U Graphics card
June 24, 2004 10:54:08 PM

For 3Dmk03 it's best to run all tests so we can see everything, BTW, compare links don't show up with both systems, I thinnk we are only getting the links you are comparing to, not yours. Post your 'published' link that should be visible in 'details'

Also post 3Dmk01 results and even PCmark results for a complete picture if you want people to tweak by proxy so to speak.

I saw UT2K4 but didn't realize that 'crossfire' was the map therin, don't use old maps, newer maps only for me with onslaught.

As for AA/AF, yep that happens alot. :wink:

And no UT2K4 is not a DX9 graphics class title. IT may have some DX9 aspects for audio, communications, etc. like we've discussed before, but it is strictly a DX8/8.1 class game graphics wise. No use of PS2.0 there, so the GF4ti should do just fine, not sure if there are PS1.4 effects which it couldn't do, but for the most part that would make it faster on average. Also the sliders will automatically adjust changing texture quality etc. so make sure all are set the same.

The other thing is remember the maps are different, and different maps have different strains, Torlan is actually quite easy, most people were surprised that they played the Demo with Torlan without a hiccup on higher settings and suddenly with the new maps of the actual game they found themselves chugging. With some of the newer user created maps I will find that while I never chugged before suddenly a certain map has noticeable slowdowns for me.

Also, to improve load times, disable the pre-load character maps (or whatever it's called) option, which gives you about a 5-10 second delay at the start of each round.

As for the questions;

1) Map, Detail levels, and remember <b>Lars didn't use bots in his review for THG</b>. So it's alot less stressful than an actual game.

2) Settings or drivers, but like I said, the GF4ti is a very very good DX8 card, so it's not stressed by this game or low quality settings (kick the AA/AF up, then it will cripple a GF4ti while the newer cards wouldn't feel a huge impact). And like you noticed, the truely stressful areas the GF4ti just craps out while the newer cards chug a bit, most people are concerned about min fps in FPS. And even with everything cranked to max, UT2K4 may have limited the settings and dumbed it down a bit for the GF4ti. Don't know for sure though that's just a guess.

3) I assume this has been rectified and brought within reason with the AA/AF issue solved. :lol: 

Just some thoughts, not sure if you will see things working as LARS review, or even your old card (if they are set differently behind the scenes or aren't rendering the same effects). The closest you can get is trying to benchmark with no bots on Torlan (a very light map like I said).


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil: 
June 25, 2004 1:48:05 AM

Female? I bet your ugly.lol, just messing.

My dick is so big, that my dick has a dick. And my dicks' dick is bigger than yours.
June 25, 2004 2:14:05 AM

You need to change your sig. Think of the children.

<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN</font color=blue>
<font color=yellow>SWEDEN SE SWEDEN SWEDEN</font color=yellow>
<font color=blue>SWEDEN</font color=blue> <font color=yellow>SE</font color=yellow> <font color=blue>SWEDEN</font color=blue> <font color=blue>SWEDEN
June 25, 2004 6:21:38 AM

Torlan is also having the same problem. Lets say I knock my in game settings to medium.. if I'm having a bottleneck things should improve right? Or if I know my res to 1024.. that should give me more fps too right? How about if I force the aa etc to be the lowest settings?

All this I have tried, to no avail. I can play with my settings all I want, I can try any map I want. The same problem ensues - fps dropping to sub 30 when a little fighting goes on. I just want someone to confirm that this is normal because I don't think it is. If it IS normal I've just wasted my money basically.

By the way, I've asked a couple of 9800 users what their frame rates in ut are like - and they are higher than mine. The thread is here http://www.respawned.co.uk/forums/showthread.php?t=1637...


<P ID="edit"><FONT SIZE=-1><EM>Edited by rageyTH on 06/25/04 07:27 AM.</EM></FONT></P>
a b U Graphics card
June 25, 2004 11:47:20 PM

The main thing is that it will also be BOT dependant, which is VERY system dependant. When benchmarking UT2K4 depending on the benchmark botmatch/flyby, you see that the bot match (more representative of face-to-face gameplay IMO, is very system dependant. And you may not see much improvement with your setup. The improvement you may have seen when switching may have been due to other factors involved in rendering the players. It's hard to tell how much impact your memory is having on your system. I will see whhat has the greatest impact on my rig when I get home. If I can glean some system boosting settings from my tests I;ll let you know. I'll also see if I can dig up the tweak guide I saw when I first bought UT2K4, and see if there's some settings that may help.

Quote:
The same problem ensues - fps dropping to sub 30 when a little fighting goes on. I just want someone to confirm that this is normal because I don't think it is. If it IS normal I've just wasted my money basically.

Well it depends. If this is the ONLY game you play then maybe so, the latest cards aren't needed other than to add a little sparkle IMO. However in other games you may see some benifit, although that means little if this is your primary concern. Also you may simply need to do a little system tweaking if your problem if system dependant and not graphics dependant, in which case an upgrade to an X800 may have had less impact than new memory or such.

I know it's not much help but if you system is a clean install and you're still experiencing issues on the exact same quality settings, then it does sound to be more of a system issue than a graphics issue.


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil: 
June 26, 2004 3:47:27 AM

That fixed the problem.
June 26, 2004 3:56:27 AM

Now I have 2.4C OCed to 3.0. Using HyperX PC3500 2X512. I had a BBA 9800XT. I OCed it. It kicked the crap out of my TI4200 which I still use. There is no way that your gf4 is faster than your X800Pro.
a b U Graphics card
June 26, 2004 5:52:12 AM

Ok just finished a little frag session on the old maps.

First and foremost if you wanna improve FPS turn the shadows to BLOBs (seriously who needs detailed shadows in UT2K4?), unless it's set to blobs my card chugs, all other settings should be ok, but shadows default higher on better cards. Also adjust the fog distance it will also default higher.

And the feature to turn off for quicker load times, "preload all player skins".

I did find a few tweak guides;

<A HREF="http://www.tweakguides.com/UT2004_1.html" target="_new">http://www.tweakguides.com/UT2004_1.html&lt;/A>

<A HREF="http://www.firingsquad.com/guides/unreal_tournament_200..." target="_new">http://www.firingsquad.com/guides/unreal_tournament_200...;/A>(which has a similar CPU/Memory)

Notice the following in the conclusion <font color=purple><i>"As you can see from the previous pages, most of the numbers are in the 30-45 FPS region, and this is due to CPU limitations and in some instances, GPU limitations. Some of UT2K4's attributes are very CPU dependent, so know this before starting your tweak trials."</i></font color=purple>



And you might want to add <A HREF="http://ut.filefront.com/file/ATI_Startup_Logo;25125" target="_new">THIS TWEAK</A> in light of your recent purchase. :cool:


- You need a licence to buy a gun, but they'll sell anyone a stamp <i>(or internet account)</i> ! - <font color=green>RED </font color=green> <font color=red> GREEN</font color=red> GA to SK :evil: 
!