Forum Home
Press F1
Thread ID: 70277	2006-06-27 21:15:00	7800GS underperforming in certain games?	Zygar (29)	Press F1

Post ID	Timestamp	Content	User
466732	2006-06-27 21:15:00	My AGP 6800GT blew up a while ago, and the store replaced it with an AGP 7800GS . However, it doesn't seem to be performing as well as it should, in some games . I have it running at 400/625 (400/1250 effective) and I've done some comparisons to benchmarks on tomshardware ( . tomshardware . com/2006/02/02/nvidia_7800gs_confirms_agp_aint_dead_yet/index . html" target="_blank">www . tomshardware . com), and some of my previous results on my old card . The tomshardware test rig is about identical to my own, only a different motherboard) TH 3dmark05 result, default settings: 5986 3dmarks My 3dmark05 result: 5832 3dmarks . (This is close enough to be an anomaly due to different motherboard chipset) TH FEAR result, 1024x768, 4xAA, 8xAF: 46FPS average . My FEAR result, same settings: 47FPS average . So far, so good, right? I took a look at the Doom 3 results: TH Doom 3 result, 1280x1024, no AA, 8xAF, high quality, Demo1: 97fps average . My Doom 3 result: 75FPS . A 22fps difference . While it doesn't make a huge difference when vsync is enabled, it doesn't bode well for future games, and it does indicate some kind of problem . Before anyone asks, yes, that was on the second run, after the textures had been cached . I got like 59fps the first time around . edit: I just looked through an old chatlog of mine, and I was telling my friend that I got 72FPS at max settings 1280x1024 on my 6800GT at ultra speeds, and both tomshardware and anandtech are showing higher results . Possibly there's another problem causing my low Doom 3 benchmarks, but I shouldn't be getting a lower performance in HL2 than before . I did the CS source stress test at 1280x1024, max settings, with 4xAA and 8xAF on, and I got 78FPS . An old chatlog of mine indicates that I got 86 . 2fps, at the same settings, on my 6800GT at Ultra speeds . This seems kind of abnormal . I wasn't really recording my frame rate all the way through HL2 Episode 1, but I'd get a varied framerate, but I'm sure it was lower than the TH HL2 Lost Coast benchmark . I'm going to download Lost Coast and test it at some point, when my bandwidth cap allows for it . (curse you 3rd world new zealand internet) My specs: - AMD Athlon 64 3200+ S754 (Newcastle, runs at 2 . 2ghz stock) - MSI K8N Neo Platinum (NForce3 250gb) - latest BIOS and nforce drivers . - 7800GS, obviously . Running at 400/1250 for testing purposes, as all the benchmarks online seem to be testing a BFG model that runs at that speed . - 1GB of Geil RAM . - Audigy 2 ZS I'm running a fresh install of Windows XP Professional, and I have the latest drivers for everything, Forceware 91 . 31, etc . Has anyone else had this problem? Could other 7800GS owners possibly do some Doom 3/CS Source stress test results so I could do some comparisons? Any help or solutions would be much appreciated . Thanks .	Zygar (29)
466733	2006-06-27 22:13:00	My AGP 6800GT blew up a while ago, and the store replaced it with an AGP 7800GS . The tomshardware test rig is about identical to my own, only a different motherboard) The Toms PC maybe similar but that doesn't mean identical . And what background processes have you got happening while you are playing/benchmarking? All of thiose things can make the small difference you have . What patches for the games, what drivers? Etc etc . . . . And what brand card do you have? If you are that picky then do some research on the different brands of your card - that can make a difference too .	pctek (84)
466734	2006-06-28 04:38:00	I have nothing but the essential windows services running, I already checked that . Also, I would hardly call 20fps difference small - my card should apparently push the framerate I'm getting in the CS source stress test at 1600x1200! It's not about me being picky, it's about my card majorly underachieving in HL2 and Doom 3 . This problem is really weird . It looks like for Doom 3, my computer has always been underperforming, for Source, I'm getting worse performance, and for other games, it's running fine . (I still need to test BF2, Far Cry, etc, but . . . )	Zygar (29)
466735	2006-06-28 05:14:00	Okay, what the ****. I put AF up to 16x and ran the CS Source stress test at 1280x1024. 82.1fps. I put it down to bilinear filtering. 91fps. Surely it would make more of a difference than 9fps. There was absolutely no difference in framerate between 8xAF and 16xAF. Turning off AA made more of a difference, sending it up to 105fps (I'd expect a bigger difference though, still)	Zygar (29)
466736	2006-06-28 06:36:00	Apologies for the triple post, but apparently editing your post is some cardinal sin or something and it's disabled on pressf1 . I overclocked the card to 450/670 (450/1340 effective), up from 400/625 and stock of 375/1200, and I got . . . nada . FEAR: 50fps - a 3fps boost . Doom 3: 76fps - around the same . HL2: 86fps - a very slight boost - now it's running the same as my old overclocked 6800GT! BUT, 3dmark05 went up to 6400 - this is a 1000 point boost from stock speeds . I'm starting to think my graphics card might not be the problem here - but what else could it be? Is there some motherboard setting I should play with? My CPU shouldn't be limiting me this badly, I've seen inferior CPUs get better results in benchmarks . And I've made sure I don't have vsync on or anything . . .	Zygar (29)
466737	2006-06-28 07:16:00	Use older drivers, those the new 9 series control panel driver that only just came out, so will be a bit bugger . The 2 first game use DX API where Doom uses OGL, I would imagine sightly older driver would have better openGL code . Edit Old card blew up?, what PSU are you running?	SolMiester (139)
466738	2006-06-28 07:20:00	Apologies for the triple post, but apparently editing your post is some cardinal sin or something and it's disabled on pressf1. You can only edit your post for 15 mins, after that you can’t edit your post any more. Hope this helps :)	stu161204 (123)
1