Printable View
personally I thing sth should be done, his scores are just unrealistic, he is making fun of all oc community
If he was cheating then, he has no reason to stop, cause nothing was done to prevent then since and it will go on and I am quite sure that many are thinking of cheating and indeed they are already doing it or they are going to start, because it's just too easy, just cheat a couple of %, your excause is tweak.
I'll place a bet here....:)
I'll bench with single card and get some scores......Next day will be broken!........You want to bet?......:rolleyes:
Good one.Quote:
Originally Posted by hipro5
But I think you should bet on something else...that someone won't reply here :D
lol hipro
Iv'e benn on vaction from all hardware buisness some time know. Will back at the rigg next week. I have been benching 88gtx sli in 05 well over 5GHz cpu and from my experiance 800/1200MHz in SLI is all ou need to max out GT3. Then its all about hte cpu.
I just saw OPBs score and thought what the heck is up with that score. That is bloody imposible on ~5230MHz even with crazy good memory. Then I saw this thread must agree that his score looks strange. There are good LOD tweaks and some bus retalted tweaks for 05 but nothing close to this. From all the tweaks I know I can gain about 200-300 points and that are no real tweaks mainly overclocking the bi16 bus and pcies freqs.
If you ask me from my experience this score aint posible on the cpu clock showing in the compare link, not at all posible.
I havent benched 05 in a long time but even 06, GT1 and GT2 is pure cpu bound after a G80 gpu speed attainable by watercooling... so for 05 I'm quite sure GT1 and GT2 you can't do anything to up it other than cpu power, high cpu clocks number one effect, FSB and memory bandwidth- number 2 effect, and nothing else.So that 6 fps jump can never be from GPU, only CPU, and a lot of mhz at that
The difference in core clock between OPB and KingPin's SLI runs is 0.3% and the difference in score for GT1 is 8% and for GT2 is 4% for reference.
Regards
Andy
Out of interest lets say that OPB is a proud person - now Pedro touched on the subject of the old 2k1 tweak of setting multiprocessor to gain about 2k in the scores. If OPB has discovered something like this, I dunno some sorta software "bug" that allows for greater scores AND he's a proud person - then he's hardly gonna share it straight away is he? From what I gather he has been spending alot of time 3D benching of recent. He's also been in contact with MM and a few other good 3D benchers from OCX so he'll had gotten hold of alot of good info. I don't think the score are impossible but they are very "unique" for the clocks. Unfortunately I don't think XS will be hearing from him anytime soon. I'm not a proud guy at all but if I get scores like that then came here and read this thread I'd probably not say anything either. I understand there's some hostility here but there's a big difference between constructive criticism and blatant accusations.
Just my pov
John
@JB:
Don't expect any gains like these ( real big ) from any software tweak on 2k5 & 2k6 mate.
Unless explained, there's nothing leading to a valid score like this one.
how can you be absolutely sure :confused: (at least you seem to be)
I'm not ABSOLUTELY sure, but 90-95% sure :DQuote:
Originally Posted by dinos22
I don't want to see this getting into a " How do you know? " " I do know!" set of posts with nothing new or point of view brought to the table without at least some backup data to show your point, thanks.
Regards
Andy
Hi :D.
Well, don't know what to say about this one as I don't like to be involved in controversy, but here goes. Not trying to stir the pot, just giving my input from my experience benching the HW. When I saw that 33.9k 05' SLI score, it seemed way off the norm for the clocks he is stating in his description. IMO the single card 05' score in question isn't really that out of the norm for the cpu clocks on 680i platform, but he is obviously mis-stating his gpu clocks in that one, which is a whole other issue in itself. What validity does HWBOT have if guys are going to not put the right card clocks in their submissions?...especially someone at the very top who others will compare their results too. I think I know where he learned that move from :rolleyes:
Hitting the wrong numeric key when typing in a submission is the lamest thing I ever heard (if those are OPB's actual words).
With all due respect to MM who is THE best 3d bencher at ocx period, not quite Jonny.Quote:
Originally Posted by Johnny Bravo
Without giving names, I have asked some of these people directly about his SLI 05' score and the responses I got was:
Not sure whats up...his card clocks were higher than stated at 650, more like 920-1200...could have been caused by a bug...you know how runs are with the 8800's, one minute the score is high next minute low.
Not the kind of response I was expecting.
As for the comparison with my 32k SLI 05' score from november, keep in mind that was done on a pre-production EVGA mobo that had some issues and was done with relatively low bus speed (401x13). On a stryker or retail EVGA with 43*x12 there is definitiely more cpu power to be had in 3d, but it's not even close to the difference between the two scores with the cpu speeds being so close.
Like shamino and some others said WHO KNOW what the HW is capable of, after a certain point gt2 and 3 become very cpu bound and increases in cpu power be it through better timings or tweaks or whatever, results in increases in all game tests. Whatever is making the score seem "artificially" high, is coming from CPU end no doubt regardless of the fact that he is mis-stating his card clocks.
I just picked up another stryker as I blew up the vreg on the one I set 100k with :D. If I had run 05' with the 100k setup, the score would have been closer to 33k, but nowhere near 34k. The 100k clocks were around 950/1220 on cards and 5288mhz(406x13) on cpu.
05 and 03 will get the treatment from me in SLI this week with the stryker using higher bus speed with 12x, better tweaking, and much higher card clocks than early launch runs from Novemeber so I should have more of an apples to apples comparison of whats going on with his 34k SLI score to make a better judgement on it, as I can match the clocks "stated" in that run easily. As for the others like am3, hipro has the peronal experience now to comment on that.
The flaming has to stop though, especially from those who really have no clue what goes on when benching at this level. Anyone who would question Hipro's intentions about bringing this up is out of their :banana::banana::banana::banana:ing mind. Show some respect. The fact is the scores don't add up atm, for me particulary the 05' SLI score. This is the opinion of not just Hipro but MANY top guys. These things need to be discussed to preserve the integrity of what we are all doing here. Nothing personal against opb as I am sure if any of us posted some of the runs that he has recently, they would be in question.
Fair enough, shows what I know :rolleyes:Quote:
Originally Posted by k|ngp|n
it sad to see that guys are using tweaks and tweaks for getting more and more points... and if they cant get the results they are begining to cheat. imho extrem tweaking is lame because you have to know the right persons.. so a normal one cant get those scores... even if they run exactly the same clocks..
Let's assume that everyone who thinks OPB is a cheater is right. Who would be that stupid to cheat just by entering the gpu clocks wrong? That would be the most idiot way of cheating imho.
The core clock isn't correct, we know, but I don't think this is an example of cheating.
The 'hitting the wrong numbers' was not a statement by OPB, not at all, it was one way to explain the weird core clock.
@Xtacy: Xtreme tweaking is a part of extreme benching.
Quote:
Originally Posted by Xtasy
I`ve found plenty of tweaks by myself. I`m unlikely to be the first one to find them, but a lot of people dont need to have everything spoon-fed to them- you get out what you put in, the key to good benching is understanding. Maybe tweaking isnt a 100% fair comparison to an everyday, fully-loaded computer, so with that principle, better stop using Cascade, Dice and LN2.
yeah that's true.. look at how many top 50 ranked pi or 3dmark results have their cpu clocks at default by accident while folks immediately above and below them are in 4+ ghz speed arenaQuote:
Originally Posted by massman
@massman
I am going to enter this SPi result today on hwbot.org as I figured OPB's secret tweak.
here is my cpu validation link at higher freq than SPi was run, so I have the same level od proof as OPB has, so that must be a new SPi 1M world record.
you could expect a couple of WRs from 3DMarks and Aquamark soon
Are you trying to boycot our database?Quote:
Originally Posted by zbogorgon
I'm very disappointed that you would even think of submitting a fake result. I've lost the respect I had for you, mate. I don't mind if you think someone is cheating at all, if you have the proof for it, but I DO mind if you try to cheat yourself. TBH, it wouldn't be strange if you did some great PI times and nobody believes you, just 'cause you once had the intention to sumbit a fake result ...
Sad.
ups, I forgot to wrote "sarcasm on" at the begining of previous thread.
as I said before, If i wanted to cheat, I would already do so when I submited cpuz WR and I am 99% sure that I would be holding SPi 1M WR today, not OPB.
With boycoting I would achieve nothing, I appreciate what you guys have done with hwbot, it is by far the best database of OC achievements on this planet and I am fighting today becasue I don't want everything go to waste.
p.s. but wise man once said, If you can't beat them, join them.
Thats is the problem: you make a mistake at XS and somebody is bound to notice. Whether the mistake was intentional or accidental, something should be done.
What most of the amature overclockers here don't realize is that running benches is like playing pool. After playing for a long time, you get to see things other people can't. Meaning, a slight variation in scores, especially when running almost identical hardware, needs explanation or something is up.
We got hipro, kinc, kingpin, pedro, zak, eva, shamino all commenting on this. If are not at that level, you need not give your opinion ;)
I would like it if opb said something. Usually remaining silent is a good sign of something..
@ahmad, he commented accusation for his Spi 1M WR all-right. We all where a bunch of idiots that cannot even comprehend the level of his tweaks, sth like that.
Seriously? I would have expected more from OPB considering all the work he has done.Quote:
Originally Posted by zbogorgon
Sarcasm or not, I don't like the attitude.Quote:
Originally Posted by zbogorgon
I really feel a bit lost now. When I first started with benchmarks, I never saw anyone being accused of cheating. Now we're about 4 years later and the word 'cheat' is said 'bout every month. I really like what I do at HWBot, because I have the opportunity to spot cheaters and ban/block them. HWBot is becoming a huge database with, thanks to the contest, even more and more older hardware, but somewhere down the road, this huge database is being used as a way to spot possible cheaters. I will ban every single cheater if there's proof that 100% correct, not if there are just a bunch of insinuations, correct or not.
p.s. your wise man was not at all pugnacious.