View Full Version : Who TRUELY states OC speeds?
10-12-04, 10:50 PM
I, by far am not the best overclocker in the world, but I am happy with my results.
I know video cards vary in performance in responce to ocing, but some results certain individuals post seem to be almost unbelievable.
I am not talking about the guy running a hi core on h2o, he put the time and effort into getting great speeds and deserves it.
I'm talking about the guy with a 6800gt running 450+ core on default voltage, cooling, etc.
I would believe the general concenous is to post your highest STABLE overclock, not your highest or imaginary overclock, which perhaps some people do.
I am not trying to start anything here, just wondering what your thoughts would be on some of the incredible results some people have, when the majority of others with the same setup can even come close to that particualr oc.
10-12-04, 10:59 PM
Well im still testing mine out. It looks like the highest completely stable OC I can acheive with mine so far is 440. Thats with no artifacts or anything. Now I can run @ 450 with no messups in games but some benchs dont run right. My load temps were low so Im thinking now it just might be a voltage/power issue and Im trying to figure this out. This is running a 6800GT @ 1.5V on a NV5 Silencer. Looks like I have one of the better cores out there but everyone doesnt luck out like I did.
450/1.25 on my ultra. That is what I run 24/7 and have 0 issues. It has a VGA Silencer on it. Stock volts
I have an XFX GT in another machine clocked at 400/1150, rock stable. I could run it at 410 without artifacts, but it was getting a bit warm for my taste and I'm not going for extreme performance, so I left it at 400. Ultra in the main machine, I don't even attempt to overclock, but I did run benchies at 435 without problems.
10-12-04, 11:43 PM
425/1200 on a Gainward GT Golden Sample, all stock, no mods or anything, and completely stable and artifact free in any application/test.
10-13-04, 05:19 AM
I have an XFX GT in another machine clocked at 400/1150, rock stable. I could run it at 410 without artifacts, but it was getting a bit warm for my taste and I'm not going for extreme performance, so I left it at 400.
Same vid card here, same story except I'm at a round 400/1100. Can go faster, but the Aussie summer is well on its way.
My Gainward golden sample GT runs 400/1100 stable ; 3DMarko3, Cinema4D and Maya and both their benchmark tests and some DX8.1 games.
I see no reason why it shouldn't get any better on the upcoming cames. It's just that I haven't tried going up farther.
10-13-04, 05:54 AM
my clocks in the sig are the clocks that i actually run at... totally stable...
10-13-04, 06:16 AM
It just depends on the core some chips can overclock better then other,I believe that some of the GT can run at 450 on air since my ultra will run 475/1250 all day long on stock HSF.But since the card is so fast to begain with I see no point in running it that high and take a chance and have a $500 papper weight.
10-13-04, 06:21 AM
If i post an OC it's a 100% stable OC i tested with several benchmarks AND games. Same goes for my 3dmarks and aquamark scores.
It's utterly stupid to post a score you obtained while having tearing or artifacts.
10-13-04, 06:45 AM
Well im using an asus v9999ge which is running all vertex shaders but unfortunately ony 12 pipes :/...
Im getting 420 core and 1.2 memory. No artifacts in any benchmarks or games, no stutter. Idle is 48c core and 43c memory. load (highest i ever seen) is 71c core and 69c memory
Runs like butter with no graphical glitches or stability issues
10-13-04, 07:09 AM
yeah 450 1.25 on an msi 6800ultra i dont run it like that all the time but since jakup said what he runs at i tried it. i just got a nv5 silencer and its ok, i just tested a few rounds of BFV and all seems fine on these settings
soon as it crashes ill go back to 425 1.15
Well since I tweaked the core voltage to 1.4V, I can run my GT at 451/1120 completely fine - no issues nor artefacts whatsover.
I have decided to run 440/1100 normally, just to minimise any risk of future problems.
Shame the ram won't go higher than 1120, but I guess I mustn't grumble.
I can run my Pov 6800 GT at 440/1200 with the NV5 Silencer, and that is with 1,3v.
10-13-04, 07:48 AM
BFG 6800GT OC 1.4vmod NV5 email@example.com everyday but it does 3DMark05 at firstname.lastname@example.org w/no problems but on warm days it will get artifacts at email@example.com.
Mine doesnt want to go over 420 core stock, memory i havent tried over 1.1
My sig clocks are my "everyday" speeds....for maxxed out benching I can squeeze 3.84ghz out of the CPU and 440mhz from the GPU.
10-13-04, 09:36 AM
People have different definitions of what constitutes a successful overclock.
In my books, a successful overclock must be 110% rock solid stable. No artifacts, no crashing, no bsod.
Some people boot up once, manage to get off a benchmark before it flakes out and call that a good OC.
Now people who come to the forums are looking for some idea as to how fast they can expect to push a certain piece of hardware. It's pretty counterproductive if everyone posts imaginary OC "results". When I read other peoples OC claims unless I know the person I generally subtract 10 - 15% from their scores for the BS factor. I find once I do that, their results are almost always right in line with what I am seeing with my hardware.
10-13-04, 10:01 AM
Can bench and game my (1st Gen) BFG 6800GT OC @ 425/1200, stock voltage. But this is on Water.
I was much more apt to stating my OC speeds back when I was working with GF3's... I had a 260/560 GF3 for a while, sitting in a kt266a rig with tcb3 samsung, and a 2400 @ 2.3, 182fsb... I'm certainly more enthusiastic with equipment that's not necessarilly supposed to OC that well... Nowadays I don't do anything outta the ordinary, probably because of my rig's lack of PCI/AGP locks, and the fact that my ram and vidcards (5800U, 5900U dev. sample) don't OC too great... When I finally get my hands on an S939 setup with a 90nm 3200, and either a 6800LE or a V9999Ge, then it's volt-modding time again... I never did do a peltier setup :firedevil ... I'd like to do something like that by this time next year...
10-13-04, 10:51 AM
Mine are true, and i have the scores to prove it....sigh, yet, i feel like i could do so much better.
420 and 1.2 for me... can get a few more mhz.. but this is what i run 24/7
10-14-04, 11:58 AM
i can only seem to get 415/1100 completely stable. at 420, it takes anywhere between a few minutes to almost an hour of playing a game like doom3 and then the screen goes completely bogus. this is with a lapped nv5 silencer at 1.4v. something tells me with better case cooling it wouldn't do that...
My BFG everyday OC is 435/1100 .. will do 450/1150 but I backed it down to a "safer" speed..
The issue is that most people check only a few things for artifacts and then call it a successfull overclock.
To find a truly stable overclock you have to get the GPU to use as many of its transistors as possible.
When I want to test the true stability of an overclock, before I flash my BIOS with the new speed, I run about 20 different games and benchmarks running the gamut from DX6 through DX9C functions. It is interesting to see that sometimes, even when all the "modern" games run fine, there will be a lockup or crash in an an older game or benchmark. For instance, I had to back off what I thought was a stable overclock when the multitexure fillrate test on 3DMark 2000 crashed. That was the ONLY crash I could find in my slew of tests, however it showed me that I was not 100% stable. Backing off slightly cured that crash.
Hell, If I just tried to run 3DMark03 and a couple of current games I could probably overclock 30 or 40 more MHz on my core and call it stable. But I would get crashes when I least expect it, and that would be annoying.
I agree with the originator of this thread. I would love to see what people are doing 100% stable, really stable.
I personally am running my PNY6800GT at 400/1150 at stock voltage. Completely, 100% stable. That is sufficient for me. :)
vBulletin® v3.7.1, Copyright ©2000-2014, Jelsoft Enterprises Ltd.