Radeon X850XT - Slow?

daedal

Gawd
Joined
May 10, 2005
Messages
686
Just ran a few Half-Life 2 benchmarks and compared my results to the ones below only to discover, well I'm only getting a bit more than half of the FPS they are! What gives! Are there any other tests I should run? What do I do now?

http://www.anandtech.com/video/showdoc.aspx?i=2299&p=4

at_canals_08: 85.40 FPS
at_coast_05: 74.78 FPS
at_coast_12: 75.10 FPS
at_prison_05: 80.77 FPS
at_c17_12: 46.84 FPS

It has it's own Molex cable, IE - it's on a single molex line with nothing else plugged so power shouldn't be causing the problem.
I also get some very drastic spikes in-game. I can be running at 130.00 FPS and drop to 40 FPS or so in larger areas.. 40! That's not normal! Running on Catalyst 5.5.
 
Are you running the same settings, including AA, AF, etc.? Is your system EXACTLY set up or roughly similar in speed? Do you have programs running in the background (browsers, antivirus, etc.)?
 
- P4E 3.0Ghz @ 3.6Ghz, Zalman CNPS7000, AS 5
- ASUS P4P800E-Deluxe Motherboard
- ATI Radeon X850XT @ 546/621
- 1 Gig TwinX Dual-Channel Corsair
- Thermaltake Shark Case, 520w OCZ Modstream

Not exactly the same setup, but rough. It's a 40FPS difference.. and yes, ran 1600x1200 no AA/AF.
 
daedal said:
- P4E 3.0Ghz @ 3.6Ghz, Zalman CNPS7000, AS 5
- ASUS P4P800E-Deluxe Motherboard
- ATI Radeon X850XT @ 546/621
- 1 Gig TwinX Dual-Channel Corsair
- Thermaltake Shark Case, 520w OCZ Modstream

Not exactly the same setup, but rough. It's a 40FPS difference.. and yes, ran 1600x1200 no AA/AF.

no wonder.........its a P4 :rolleyes:

Wisper: Try A64 3500+ and turn off all anti-virus, program, webpage and put a VGA cooler on the X850 and add 2 sticks of 512 mb ram.
 
holycpu said:
no wonder.........its a P4 :rolleyes:

Wisper: Try A64 3500+ and turn off all anti-virus, program, webpage and put a VGA cooler on the X850 and add 2 sticks of 512 mb ram.
Hesitant to stick a VGA cooler since it would void warranty and this thing is brand new and it's still 600$. As far as the rest is concerned, I could run CS:S nearly as fast as with my X850XT that I did with my 9800Pro.

Was told it might be the 5.5 Drivers. I tried to downgrade to 5.4 and 4.12 by Omega but both sets didn't get picked up.
 
"Try A64 3500+" lol, and he'll see a 3 fps difference from his overclocked pentium 4..that is a very "idiotic" suggestion..hehe I own an athlon 64, 6800ultra setup..I can be cruising around with 130fps too and drop down to the 40's easily...
 
holycpu said:
no wonder.........its a P4 :rolleyes:

Wisper: Try A64 3500+ and turn off all anti-virus, program, webpage and put a VGA cooler on the X850 and add 2 sticks of 512 mb ram.

"All you need to do to fix your problem is pay another $100 for a new motherboard, $200 for a new processor, $30 for a new heatsink and $150 for more RAM!"

Seriously dude, when you typed that did you honestly think you were being helpful, even in the slighest?

I've been starting to get these sorts of stutters in games, I think it's either heat (not very likely, but possible) or Windows accesssing the page file.
 
Alright, this may sound weird but lower your memory back to 590 or 560 (whatever your default is). I tried overclocking the memory on my X850 XT PE AGP and the thing totally tanks out when I overclock the memory past 600 mhz. Try running it with the default memory speeds and see if your fps increase. I'm not sure of the explanation for this but I'm sure someone on this forum can address this.
 
you made sure youve installed your mobo drivers? they about half your speed if they arent on
 
aZn_plyR said:
"Try A64 3500+" lol, and he'll see a 3 fps difference from his overclocked pentium 4..that is a very "idiotic" suggestion..hehe I own an athlon 64, 6800ultra setup..I can be cruising around with 130fps too and drop down to the 40's easily...

thats y its "wisper" only :) WISPER !!!! jeeze, don't u get that ! obviously its being sacastic. :cool:
 
finalgt said:
"All you need to do to fix your problem is pay another $100 for a new motherboard, $200 for a new processor, $30 for a new heatsink and $150 for more RAM!"

Seriously dude, when you typed that did you honestly think you were being helpful, even in the slighest?

I've been starting to get these sorts of stutters in games, I think it's either heat (not very likely, but possible) or Windows accesssing the page file.

see my post above? :rolleyes:
There is not much he can do really, since even you can't think of a solid reason why he got such low fps. Then only conclusion is that he can try getting next gen card.
 
neubspeed said:
Alright, this may sound weird but lower your memory back to 590 or 560 (whatever your default is). I tried overclocking the memory on my X850 XT PE AGP and the thing totally tanks out when I overclock the memory past 600 mhz. Try running it with the default memory speeds and see if your fps increase. I'm not sure of the explanation for this but I'm sure someone on this forum can address this.
Just tried it. Still having runs if 70 to 80 FPS.
And yes, latest motherboard drivers installed.

My 3DMarks seem to be about on-par though.
 
might see if the hard drive that the game is on needs a defrag....also, what size is your page file?
 
Hard-Drive is defragged everyday (remote access from work
tongue.gif
) so it's kept clean.
Not sure about the page file though. I never really played with it. Alot of the websites I see have different opinions about it and I'm never really sure what to use.
 
daedal said:
Just tried it. Still having runs if 70 to 80 FPS.
And yes, latest motherboard drivers installed.

My 3DMarks seem to be about on-par though.
runs in what sense? Stress test? dips in game? or that your max? if youre dipping to 70 fps then dont complain, source is very heavy on the cpu when theres lots of people running around, and thats gonna be were your issues are.
 
Herulach said:
runs in what sense? Stress test? dips in game? or that your max? if youre dipping to 70 fps then dont complain, source is very heavy on the cpu when theres lots of people running around, and thats gonna be were your issues are.
The 70 FPS is the benchmark I'm running (see first post), but it dips down to 30-some FPS in open areas. Having many people on-screen doesn't really seem to be affecting it as it runs fine in CS when a dozen people are on screen, but as soon as a wide area appears, then it drops to 40-50FPS and during open firefights dips to 40s and high 30s.
 
Are tehy using a 128mb or 256 mb card in those benches? I dont know if it would make that huge a difference but its something to factor in (I know nothing about ati cards, there may or may not be 128mb/256mb versions.)
 
Herulach said:
Are tehy using a 128mb or 256 mb card in those benches? I dont know if it would make that huge a difference but its something to factor in (I know nothing about ati cards, there may or may not be 128mb/256mb versions.)
Guess you probably mean 256/512mb but yes, I believe they are running the same 256mb card (albeit maybe PCI-E while I'm on AGP). I've read other dozens of reviews and they get about the same framerate I do, but most don't post their custom benches so I can't really test them out.

Thinking it might be a driver issue, I started off with ATI offical 5.5 and benched. I then tried with zer0point 5.4s which lowered my FPS a tiny bit. I then tried to install Omega's 5.4s and 4.12s but both would not recognize my card, or my card would not detect the drivers and refuse to run properly.
 
finalgt said:
I've been starting to get these sorts of stutters in games, I think it's either heat (not very likely, but possible) or Windows accesssing the page file.
If it's an nVidia based card, it's probably overheating. May also want to try different drivers.

To the OP: Dude, they were using an FX-55 & you're on an Intel! ...and probably low-latency Corsair with a system optimized to run benchmarks. What do you expect??

I admit the last benchie does look strange, but there was something funky going on there. The rest look about right, though.

Main things you can do to improve performance are to scan for viruses, disable P2P programs, disable indexing and disable system restore.
 
sleepeeg3 said:
To the OP: Dude, they were using an FX-55 & you're on an Intel! ...and probably low-latency Corsair with a system optimized to run benchmarks. What do you expect??
Thanks for the input. I do have to admit, they are using a far superior Processor but do you really think it would make THAT MUCH of a difference? I also compared my benches to other sources and I just don't match up.

Re-installed my chipset drivers and my last run (the one that got 59.9 FPS earlier) went up to 72.20 FPS. I'll run the other ones and see if there's a difference.

[EDIT] My first run of benches ran at 72.85 with a 10FPS variable. Pretty much what I got in the first run but I guess it might just be my setup and not the card although I don't see my CPU activity rise much.
 
Yes, I do.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2330&p=2

Chipset would definitely help, because the nForce IDE driver alone improves performance about 5-10%. The default MS one blows and has not been updated since like 2001. Those weird slowdowns are almost always due to your drives being accessed, which is what disabling all those things I mentioned should drastically improve.
 
sleepeeg3 said:
Yes, I do.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2330&p=2

Chipset would definitely help, because the nForce IDE driver alone improves performance about 5-10%. The default MS one blows and has not been updated since like 2001. Those weird slowdowns are almost always due to your drives being accessed, which is what disabling all those things I mentioned should drastically improve.
Thanks for the link; very informative. Definitly will pick up an AMD in my next build. ;)
Oh well, guess that settles it then: CPU's holding me back.. *Cry*
 
i think it just the game in some situations u have 100+ FPS and then u drop to the 50's.........
 
Don't give up so easily. In that anand review there's a 10 FPS difference between a 3.6 Pressy and a 3500+. That gap should be less for you - a lot of people here are forgetting that Northwood > Prescott.

Try benching with other games and snythetic benchmarks, see how they fall into line. Tell us how it goes. Source is a very CPU-dependent engine.

Shame on all you who posted "AMD pwns Intel, duh" as a knee-jerk response. That's not helpful. Else, the root of everyone's problem here would be that they don't have an FX-57 (or a 4800+, depending on your flavor).
 
I would format and install latest drivers. ALSO..overclock the damn thing..they oc jsut fine. :D
 
LstOfTheBrunnenG said:
Don't give up so easily. In that anand review there's a 10 FPS difference between a 3.6 Pressy and a 3500+. That gap should[/]i] be less for you - a lot of people here are forgetting that Northwood > Prescott.

Try benching with other games and snythetic benchmarks, see how they fall into line. Tell us how it goes. Source is a very CPU-dependent engine.

Shame on all you who posted "AMD pwns Intel, duh" as a knee-jerk response. That's not helpful. Else, the root of everyone's problem here would be that they don't have an FX-57 (or a 4800+, depending on your flavor).


I agree, but at those settings with the 10fps diff...was probably @ 800x600 low details or something..upping the resolution should give minimal to non gains.
 
sleepeeg3 said:
Yes, I do.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2330&p=2

Chipset would definitely help, because the nForce IDE driver alone improves performance about 5-10%. The default MS one blows and has not been updated since like 2001. Those weird slowdowns are almost always due to your drives being accessed, which is what disabling all those things I mentioned should drastically improve.


I tried the Nvidia IDE drivers, but it would cause the system most of the time to hang during Windows load, or when a disc was inserted into my DVD rewriter. I updated the firmware of the DVD rewriter and it did not fix it, only uninstalling the Nvidia IDE drivers and using the MS default ones that come with WinXP SP2 worked. A shame considering the performance gains I have heard of...
 
LstOfTheBrunnenG said:
Don't give up so easily. In that anand review there's a 10 FPS difference between a 3.6 Pressy and a 3500+. That gap should[/]i] be less for you - a lot of people here are forgetting that Northwood > Prescott.

Try benching with other games and snythetic benchmarks, see how they fall into line. Tell us how it goes. Source is a very CPU-dependent engine.

Shame on all you who posted "AMD pwns Intel, duh" as a knee-jerk response. That's not helpful. Else, the root of everyone's problem here would be that they don't have an FX-57 (or a 4800+, depending on your flavor).
Actually, I am running a Prescott (S478). Other benchmarks gave me good/normal results.

3DMark05 : 5877
3DMark03 : 13173 (OCed a bit)
3DMark03 : 12388 (Stock)

I thought picking up a new video card would have improved my framerate, not lower it..
 
:( Sry about that. Well, then, erm, you get what you deserve. Okay, just kidding, but Prescott & 865 is still > Prescott & 915/925.

You're actually seeing a *decrease* vs. the 9800 Pro? Odd indeed.

Bah humbug. I can't find any 3DMark scores to compare yours too without delving into the ORB. Must hurt for Kyle, Anand, and others to not even use them anymore. Anyway, my advice? Try reinstalling HL2. What can it hurt? Certainly won't hurt you in the pocketbook like buying an FX-57 would.
 
LstOfTheBrunnenG said:
:( Sry about that. Well, then, erm, you get what you deserve. Okay, just kidding, but Prescott & 865 is still > Prescott & 915/925.

You're actually seeing a *decrease* vs. the 9800 Pro? Odd indeed.

Bah humbug. I can't find any 3DMark scores to compare yours too without delving into the ORB. Must hurt for Kyle, Anand, and others to not even use them anymore. Anyway, my advice? Try reinstalling HL2. What can it hurt? Certainly won't hurt you in the pocketbook like buying an FX-57 would.
Not really a decrease but the performance increase is negligible. Guess I'll re-format when I get a minute and see.
 
well im running a 3500 amd and a x850xt pe...and in cs:source with 1600x1200 and 6aa/16af the lowest my frames drop to are 50s..and i play half-life2 at the same res/candy and the lowest i drop to is 60 and my highest is about 175 or summit cant remember havent played the game in ages..
 
Mayhs said:
well im running a 3500 amd and a x850xt pe...and in cs:source with 1600x1200 and 6aa/16af the lowest my frames drop to are 50s..and i play half-life2 at the same res/candy and the lowest i drop to is 60 and my highest is about 175 or summit cant remember havent played the game in ages..
Then again, your Processor eats mine for lunch and still has room for more. ;)
But my FPS averages around 70 to 80 (100+ in small rooms) and this at 1024x768 w/ 4xAA and 4xAF and oftenly tips under 40. At this setting, the CS:S video stress test gave me 130FPS or so average which is lower than what I was pulling with my 9800 Pro.

Maybe I should go back to BestBuy and swap it but the weird thing is, the 3DMark runs are straight on.
 
well at least I know now, kinda..that it's not only nvidia cards that dips to the 40's :)
 
CS: Video Stress Test Results

1024x768 - 4xAA, 16xAF
All other settings at highest level
DirectX level is at 8.1 (mat_dxlevel 81)

Run 1: 125.72 FPS
Run 2: 123.03 FPS
Run 3: 120.50 FPS
Run 4: 120.51 FPS
Run 5: 119.83 FPS
Avg: 121.92 FPS


1024x768 - 4xAA, 16xAF
All other settings at highest level
DirectX level is at 9.1 (mat_dxlevel 91)

Run 1: 110.03 FPS
Run 2: 109.02 FPS
Run 3: 107.89 FPS
Run 4: 108.39 FPS
Run 5: 108.55 FPS
Avg: 108.77 FPS


1024x768 - 4xAA, 8xAF
All other settings at highest level
DirectX level is at 9.0 (mat_dxlevel 90)

Run 1: 106.48 FPS
Run 2: 105.47 FPS
Run 3: 104.15 FPS
Run 4: 104.07 FPS
Run 5: 102.83 FPS
Avg: 104.60 FPS

Weird. My FPS keeps dropping and dropping the more runs I make.

Check out these guys' benchmark runs. See the Gamespot timedemo FPS? Guess it might not be so abnormal for me to dip down to 40 or so. Hell, they're getting 40-45FPS average on a X800XT-PE
http://hardware.gamespot.com/Story-ST-102-789-9-9-x&body_pagenum=2

Here's a couple of other VST runs from other people which I compared to. I seem to be in the average so I guess it really is my processor bottlenecking.
http://www.nvnews.net/vbulletin/showthread.php?t=40278
 
Back
Top