So then why is hwbot allowing scores with high "bug"????
And why do i get this as a response? http://www.hwbot.org/forum/showpost....&postcount=187
Printable View
So then why is hwbot allowing scores with high "bug"????
And why do i get this as a response? http://www.hwbot.org/forum/showpost....&postcount=187
He prefers it even more, when we call it "anomaly" :p: :DQuote:
Originally Posted by dinos22
j/k Gautam ;)
i also encountered a bug with my 8600m-GT
after first test, a window pops up, saying driver is lost or something, not able do display normal things anymore, i should reboot etc.
test continues => higher scores :)
lol at jmke :rolleyes:Quote:
Originally Posted by T_M
Because the high scores are NOT the bugged ones. Check out the post at HWbot and this thread regarding the 4870X2 performance scaling. There's no point in blocking scores that represent the actual performance of these cards just because most of the other scores are bugged.Quote:
Originally Posted by T_M
This is probably why Gautam asked me not to use the term 'bug', as it making it sound bad.
That bug has nothing to do with what we're dealing with here, but is what some people call the 'black screen' bug. Basicly the driver crashes, but 3DMark03 doesn't detect it and continues to count frames, resulting in a much higher score. People experienced it in Vista32/64 with the X2's and blaimed the VPU recover function.Quote:
Originally Posted by Bellisimo
so there's a high bug, low bug and normal ?
Only high bug and low bug, there seems to be nothing in between. It's either GT2/GT3 at a correct level or GT2/GT3 at a subnormal level.
You are saying here that high scores are buggedQuote:
Originally Posted by massman
And here you are saying they are not :shrug:Quote:
Originally Posted by massman
Im not trying to be an arse, its just this :banana::banana::banana::banana:ty situation in which im trying to clarify what hwbot's stance is.
So far im hearing that high runs are actually the real score, and low runs are bugged. As such hwbot accepts them all.
I use the term 'bug' because otherwise people don't know what I'm talking about. I shouldn't because:
Quote:
Originally Posted by Gautam
Therefor, this sentence should be: "It's the other way around: you lost the boost by unplugging the graphics card. If you disconnect your system, or parts from your system, from the power circuit, you lose the boost. The only way to retrieve the boost is by luck."Quote:
Originally Posted by Gautam
Anyway, these are just words to describe the phenomenon. The problem is that people have used the term 'bug' from the beginning, while they shouldn't have.
Your hearing is correct: the runs with GT2/GT3 low are bugged, the runs with GT2/GT3 high are the ones that show the real performance, hence are legit scores.
The scores that are high because of the 'black screen' bug (driver crashes resulting in a higher score) are NOT legit and will not be accepted.
so how do you distinguish the black screen submissions from boosted scores
does black screen submission affect GT1 and Nature
This called "bug" in uncontrollable and most obviously undetectable.
Check below result with drivers 8.8 and PCI frequency @ 100MHz, which seems quite normal to me. MUCH better than my modded 9800GX2 was giving me with 8500 @5,7GHz (around 85k+)...
http://www.imageshack.gr/files/1iqdi65qenxrbl3bxhyf.jpg
Now check this out...
http://www.imageshack.gr/files/8masvtib50m1e0kst17n.jpg
...fresh installation of 8.9 on a very light OS (XP 32Bit), no CCC and PCI @ 120MHz. No black screens no changes in refresh rates, no tweaks (LOD, priority etc). Just a simple run with most of GT lower than the first one, though the outcome is higher by 4000 marks...:confused:
what .... the.... :banana::banana::banana::banana:
El Diablo of Suxxor is Back :D:D:D 666
http://www.ixtremtek.com/boblemagnifique/136666pts.jpg
This is pure ATI power man....:ROTF:Quote:
Originally Posted by T_M
No wonder, in most previous, ATI used 3D03 & Vantage to prove this VGA's strength.
Anyway, I'm looking forward for a nice LN2 session and a 233454597875 3D03 single card score.
Still nothing from Futuremark????
interestingQuote:
Originally Posted by dread77
i might try no CCC as RT v2.11 does the job for OCing anyways
but for some reason P45 boards i have do NOT like PCI Express frequency going up
that will probably kill my scores a lot actually now that i think of it
Well, Blackops does not like high PCI frequencies when playing with 4870X2 as well. The highest I could get with Blackops was 108MHz, above that... no post at all.
On Rampage Extreme I gained almost 2500 marks playing 3D03 in 120Mhz from 115Mhz. Maybe PCI frequency is also related with the bug.:confused:
I have no idea, I have not yet been able to perform a black screen bug. I reckon it's just one test hitting extremely high FPS rates and if that is the case, it should indeed be fairly easy to detect them.Quote:
Originally Posted by dinos22
The difference between your two runs lies in the Nature test, which increased by about 140FPS, which is A LOT. Did you tweak anything? Or is it just the PCIe link that is giving you such an incredible boost?Quote:
Originally Posted by dread77
That is one of the possibilities. Any chance you could test this?Quote:
Originally Posted by dread77
Quote:
Originally Posted by massman
Already did. It gave me 2500+ marks from 115MHz to 120Mhz. Though, somehow, it's a reasonable gain with these cards. If it was giving me more tnan 10K marks I would say, that's it.
What on earth is going on here ? :eek: What's that boost you got in GT4 ... ? :confused: And that was sudden ? No tweaks, nothing ? :confused:Quote:
Originally Posted by dread77
Share your tweaks with us plz :p: :D :rofl: j/k ;)
http://forum.crazypc.ro/attachment.p...2&d=1222104829
I`m 1000% sure its a normal run, not bugged!
800/1000 i got 88900, no black screen no nothing, and raising the VGA clocks the FPS`s are higher to, visible :).
Hey guys, I am planning on making a video or even multiple videos of various 03 runs at various speeds.
My runs are reliable and I can get them any time.
I am planning on breaking 163k on tape this weekend at about 6320MHz, with the cards at 840/1080. If people want any other CPU speeds or GPU speeds for me to videotape, let me know and I will.
I will do complete runs, and I'm willing to do multiple runs back to back to show consistency too. My plan is to do 4 or so ~163k runs back to back. I'm confident I can do it.
As for the reason some people have much higher runs than others, I mostly believe it has to do with the build of Windows, though I don't understand why Tim's runs used to be high (142k at 4.8GHz is very good) and then suddenly got low (157k at 6.1 is not, I can get 157k probably at about 5.6-5.7)
I want to demonstrate what the refresh change does, but I think I'll do that in a seperate video, at 4.5GHz maybe.
Quote:
Originally Posted by Gautam
The video cannot prove anything now. You know how many great overclockers got "bug low" and how many people results the same as yours?
85% bug low VS 15% real score. I didn't understand why these 85% need to trust 15% real run.
More and more great overclockers achieve bug low now. If only few bug low, you could say it's bug low. But over 80%+ bug low, how did you say it's bug? Or you say these overclockers are NOT good ocer? I saw some results gpu is higher than you and cpu is similar to you, but score is 10000 score difference. We know 03 depend on gpu clock much than cpu. If gpu clock 980 970 960 950 940 cannot be higher than your score, they could not make sense.
Hmm...you wanna get me plane tickets to MI then? :p:
And yes I know, and I don't understand why all of your guys scores are so low. When I install service pack 1 my scores become high EVERY TIME.
http://www.xtremesystems.org/forums/...&postcount=216Quote:
Originally Posted by AndreYang
NO! :DQuote:
Originally Posted by AndreYang
The problem is not due to the incapability of the bencher to get 'bug high' scores, it has nothing to do with the skill of a bencher, it's something that at the moment, no bencher can control. Gautam says he gets the good scores with SP1, but other people don't. I get a good score when I overclock in small steps, other don't. It's like the 1M boost biosses: the hardware/software makes the score better, regardless of how skilled the overclocker is.
Quote:
Originally Posted by massman
Please explain why is ONLY 3D 2003? If I can take very high score(180K+) and no black screen, you could say it's NOT bug run? I also can take video.