im only getting 5878 on 3dmark05!

wakeboarder540

Distinguished
Mar 13, 2006
5
0
18,510
A friend said I should be getting a higher score than 5878 on 3dmark05...
this is my system:

AMD 64 4000 (stock)
Asus A8N SLI (original model)
some cheap RAM (1 stick of 512mb)
evga 7800GTX 256mb with ACS3 (470/1300)
antec "not so stealth" 450W p/s
 
why don't you try upping that cheap ram to 2x512. or if its seriously cheap ram, spend a hundred bucks and get better 2x512 or if you wish 2x1gb. then, see what score you get.
 
First Of All , Its Your Memory Try 1 or 2 Gb and it MAY solve your Problem , Second Try Updating Drivers , Third its maybe your PSU , ...
Lots of Possibility , But I`m 99% Sure Its Memory :twisted:
 
A friend said I should be getting a higher score than 5878 on 3dmark05...
this is my system:

AMD 64 4000 (stock)
Asus A8N SLI (original model)
some cheap RAM (1 stick of 512mb)
evga 7800GTX 256mb with ACS3 (470/1300)
antec "not so stealth" 450W p/s

So you have a Dual Channel DDR motherboard and a Dual Channel DDR Processor... yet only a single stick of 512MB of ram?

It's definatly the ram.

2x512MB as a minimum for serious gaming... ouch.. that's quite the bottleneck you've got there.
 
A friend said I should be getting a higher score than 5878 on 3dmark05...
this is my system:

AMD 64 4000 (stock)
Asus A8N SLI (original model)
some cheap RAM (1 stick of 512mb)
evga 7800GTX 256mb with ACS3 (470/1300)
antec "not so stealth" 450W p/s
512MB of single channel RAM... what a waste of 7800GTX.
 
A friend said I should be getting a higher score than 5878 on 3dmark05...
this is my system:

AMD 64 4000 (stock)
Asus A8N SLI (original model)
some cheap RAM (1 stick of 512mb)
evga 7800GTX 256mb with ACS3 (470/1300)
antec "not so stealth" 450W p/s
512MB of single channel RAM... what a waste of 7800GTX.

just as much of a waste as at 7800 GS with an Athlon XP, yea i spoke with a kid who wondered y his FX 5500 got better FPS than his 7800GS.....
 
I highly doubt it's your ram causing you to be 2000pts lower then u should be.

7800GTX running with 512mb ram 7796 score
bertram928@hotmail.com (if this url doesn't work search by the email address to find it on futuremarks site)

Did you uninstall your old graphics driver from your last card, before you installed your new 7800gtx? Alot of people mistakenly think that you don't have to uninstall the drivers for your previous card when you install the new driver...
It can make a 15-20% difference.
Example: 6800gt and P4 system...
After Uninstalling/Reinstalling drivers...5014
Before...3625
ATI and NVidia drivers fight each other, but getting a different nvidia card (without uninstalling the old drivers) can be just as devestating to your performance. You can't just overinstall new drivers for your graphics card, you must uninstall, reboot, and install new ones. If your old card's drivers are still in, in addition to your new ones, you need to uninstall your current graphics driver (just the nvidia graphics driver...leave the other nforce drivers alone...), shut down, put your old card in, boot up, uninstall your old drivers, shut down, put your new card back in, reboot, and finally, install your new drivers.

Simply uninstalling your old drivers may not work, or may not work completely. You might see a % increase, but not as good as you would if you uninstalled it properly.

If your old card was an nvidia card and you didn't uninstall the old drivers before installing new ones: then simply uninstalling the driver you have in there currently (and reinstalling the new drivers), should do the trick...
-otherwise you may have to go through the earlier steps, only this time, uninstall your current graphics driver (just the nvidia graphics driver...leave the other nforce drivers alone...), shut down, put your old card in, boot up, reinstall your old drivers (since they were improperly uninstalled...), reboot, uninstall your old drivers, shut down, put your new card back in, reboot, and finally, install your new drivers.


Or...It could just be that you have a lot of crap running in the background holding you back. If you ran it with f@h running for instance (no pun intended), that would account for the lower then avg score.

With all that said...your gaming is severely limited (in terms of new games...) by that 512... There are two games out now that i personally play that use more then 1024mb to play (and almost every game i play uses more then 512...)...EQ2 and BF2. I've never seen where BF2 uses more then 850mb or so, but it runs sooo much smoother with 2048 then with 1024. EQ2 also no longer lags in heavy zones as tons of stuff are loaded in (but i have ctrl-alt-del to chk the processes, and it does use more then 1024mb during gameplay with the way i have it setup).
 
TH has tested 3dmark05 with 512mb to 2gigs with little difference in performance. But only using one stick has eliminated dual channel. Which is reducing your total performance. Ideally if you would get at least 2x512mb of faster memory, you would see a nice improvement. AMDs love ram with fast timings. You have a nice system I would get some nice ram. Patriot makes good ram for the price. I have included a link.

http://www.newegg.com/Product/Product.asp?Item=N82E16820220006

If you decide to get more that 1gig don't go the 4x512 route go with 2x1gig. If you use 4 sticks of ram it will raise your timing from 1T to 2T. If you stay with using only 2 memory banks you can still use the 1T timing. Which is a signicant improvement in memory performance.

And lastly if your system is still not running as fast as you think it should. I would look at the PSU. Make sure that you have a strong enough 12v rail or rails. And make sure that you have the power connected correctly to it.

3500+ winchester stock speed
MSI Neo4 Plat SLI
2x1gig OCZ Platinum 2.2.3.5.1T
BFG 7800GT OC 499/1200
Ultra 500w PSU

3dMark05 score 8,025
 
If the 7800 isn't getting enough juice he will get an error message stating that it's clocking the 7800 down. It will do that on boot, or during the middle of a game, wherever it becomes a problem. It's built in to the driver (it does something similar when it gets too hot). Then anytime you look at the gpu/ram speeds of the graphics card they'll show that they've been clocked down. It won't even boot if he doesn't have the power cable plugged in...

Since he hasn't said that he's been getting any kind of errors, and he hasn't said anything about it straight up locking up, then i'd have to argue against it being a power issue.

edit: Off-topic...How do you like that ocz plat 2x1024? Is it the 3200? I've got the ocz plat 3200 2x1024...oc'd to 250fsb 1:1, have you tried oc'n with it?
 
Please stop posting that the issue with 3dmark is due to the ram... (sorry to pick on u handy =D). We all know he needs more ram, but that isn't the issue here. His low score is from something else.

http://service.futuremark.com/compare?3dm05=1851335 Whats that? 17000pts in 3dmark 05 you say? Yes...thats right...on 512mb of tccd 2x256. There is another url i have linked above that points to the same system he has, with a different person who has 512mb and is scoring 2000pts higher then he is with the same basic system.

At best you're going to see a 3-5% increase in your 3dmark05 score by utilizing dual channel. And there is maybe 1% in having 1024 over 512.
 
I was just throwing the power issue as a small possibility I am pretty confident that the issue is ram.
I don't mean to hijack your thread. But yes I like the ram. It is the 3200. This is the second pair I have owned also had 2x512. Have done a little OCing nothing major I am selling this processor and mobo to a friend so I'm going easy on it. Getting the Opty 170 and MSI diamond so we will see.
 
Meh...you guys and your dual channel fetishes...

Dual channel is almost a marketing gimmick imo...It's does about the same % difference as running cl2 vs cl3. I can back up the cl2 vs cl3, since i've got tccd which does 2 2 2 5, hyperX which does 2 3 3 8, and ocz 2 3 2 8...as well as some crappy older 2700 ram. I recently tested all 3 i listed above on a couple of bench programs with less then 1% change between the different ram types. I think i have links to them...Eh i didn't have the ocz on there...but i have HyperX 2 3 3 1T and on a divider 400:333 and 2T timings, and my patriot xblk tccd 2 2 2 1T

3DMark 06: 2498 3000+ (s) 6800gs (s) 490 1100, 1:1 2 3 3 8 12 16 1T (HyperX)
3DMark 06: 2497 3000+ (s) 6800gs (s) 490 1100, 1:1 2 2 2 5 12 16 1T, (Patriot XBLK)
3DMark 06: 2495 3000+ (s) 6800gs (s) 490 1100, 4x512 400:333 2 3 3 10 12 16 2T

That really isn't the best example...If i had a bunch of runs on my 3dmark2001 it'd show more of a difference, since it's more cpu/memory limited then 06...I've got runs on it, but mostly oc's...I'll see what i can dig up...

Here are some PCMark 05 scores:

3486 Jetway 3000+ (s) 6800gs(s) 490 1100, 1:1 2 2 2 5 12 16 1T (TCCD)
3465 Jetway 3000+ (s) 6800gs (s) 490 1100, 1:1 2 3 3 5 12 16 1T (HyperX)
3409 Jetway 3000+ (s) 6800gs (s) 490 1100, 4x512 400:333 2.5 3 3 8 12 16 2T (HyperX)(Patriot Xblk)

That should show a little better what kind of difference memory timings make...Of course that still doesn't say anything about dual channel...But i don't have anything in single channel that is worth posting...I have dc vs sc 9500/9700 oc'd scores, but i'd like to post something a little more up-to-date.
 
...Of course that still doesn't say anything about dual channel...But i don't have anything in single channel that is worth posting...I have dc vs sc 9500/9700 oc'd scores, but i'd like to post something a little more up-to-date.

so then how can you say that dual channel is bogus if you dont have anything to back you up?

Most reviews and such are clear that timings are a marginal improvement but all are very clear that dual channel is much faster. When I went from single to dual channel on my old althlonXP 2700 i noticed the diff. Most benchmarks of intel and amd going to dual channel showed improvement on many angles.

True, there is prb many things that could cause his slowdown, and his system should perform higher w/ that proc and gpu such that memory may not be causing a 2000 point hit in perf... BUT I would agree with most everyone else on here that he NEEDS to go dual channel. He can use 2x256 if he wants but however much ram he has needs to be in two equal sized sticks and he will see an improvement.

I would recomend just buying another value ram stick at 512 if $ is tight and you should see better 3dMark #s. If you plan on oc'ing then for sure buy a matched set (2x512 min) from ocz/corsair or other reputable brand (dont know how patriot is...?) and then oc and notice a bit more...
 
thanks for the suggestions 😀
It's not the power supply... my card is running at 470/1200 fine.
I am going to get more memory... probably OCZ 2-3-2-5 2Gb (2x1gb) dual channel kit, but it takes me a while to get the money for this kind of stuff, lol I have a crappy job 🙁

I forgot to tell you guys... I am using a very old beastie 19Gb hard drive 🙁
Im not sure how fast it is, but it's a FUJITSU MPF3204AT, maybe it's a 5400 rpm drive and thats causing the problem!

would a slow hard drive cause about 2000 difference?
 
wow thanks for your hahaha's, but that didnt answer if the hard drive could be causing about 2000points difference!!!!!!!!!!!!!! :cry:

wow thanks for telling us this late, and if u couldnt figure out what i was implying, then why even bother. and as for your "could this be causing about 2000 points", who the hell knows. maybe its only screwing you over for 1000. but its definitely screwing you over. <-- in case you still havent figured it out
 
well then mr smarty... it was screwing me over almost exactly 2000 points!!! I must have installed some drivers over top of some old ones... funny cause I don't remember downloading any drivers at all, just the evga one from the disk that came with my card.

so theres no point in getting mad at me for telling you guys late about my crap hard drive, because crap hard drive or no crap hard drive it doesn't seem to matter, cause im pullin in about 7800 marks now 😀 yeeeeeeeeeehaaaaaaaawwwwwwwwwww 😀

thanks who ever suggested it was the driver overlap problem, you made my day! 😀
 
I'll see what i can do about that. You did see that i said 'IMO'. So it's just opinion, i didn't say it was in fact. Anyway, my %'s are based on reviews that other people have done...Anandtech, [H]ard, etc...But you're right, never trust someone else to do your own work. So i'll try to run a couple of benches this week on one of my machines in both dc and sc. I'll post pics of scores in pcmark05 and 3dmark 05, with 2 instances of cpu-z for pseudo proof(showing the memory timings and sc/dc, and the cpu info).

Or if you want the info faster, you could test it yourself on your own machine. That way we'd have something to compare against when i do mine. Take it ez, laterz.
 
that is cool man, I wasn't meaning to sound like an a$$, just giving my thoughts... I did see the "imo" so I wasn't thinking you were being a troll or nothin.

I don't think I will have the time to get #s to compare to yours, but i am interested to see what you get. Like I said, the perception i had with my old system "seemed" to be faster... but that could just be me and my bad memory... lol