tiborrr

Hi, in the last few days i've submitted a few 7900GS results in three different categories (01SE, 05 & AM3), but only recieved points for 01SE which gave me a boost to ~ 471 points. I've also submitted 05 & AM3 results yesterday and also an improved 01SE run, but still at ~ 471 points while i should be at around ~ 495 since i recieved another 14.8 + 8.7 for AM3 & 05. http://hwbot.org/user.do?userId=7258 I've noticed similiar problem in the past. I solved it with deleting and re-posting the results, but this time it doesn't work. Please attend to it. Thanks! Greets, N.

Quoted for truth! Signature material, btw! :woot:

XS with two new members again on the 4th place, hunting down fellow benching buddies from MXS Modlabs.

@jmke: I told you to provide ORB, not some screenshot. There's no way for a G92 8800GTS single-card to reach 1000FPS at the clocks given before (1050/1150) without MIPMAP tweaking, let alone 1400! Misleading screen, fake or taken from a wrong benchmark. C'mon, are you serious? I though your knowledge of 2001SE was on a bit higher level. Apparently someone else is living in a wokywonderland...

That ORB link leads to 8600GS, those results look all over the place. Too high nature (look like a SLI setup), too low Car High (that could indicate SLI), too low Drago Low (again, indication of SLI).

:banana: Link please?

Yes, jmke, that's the main reason - if that makes you happy, so be it! :woot: I know it's impossible to hunt down these results even if i put my chip under cascade and run it at 5500-5530MHz and GPU under singlestage which might yield me another 50-100MHz on the GPU. Therefor i say it again - i don't care anymore, please lock it.

Yeah, would really like to see 1100FPS Nature on 8800GTS G92

Newer -> slower with ATi in older benchmarks because of optimization for newer 3000 series cards, it's funny how it's opposite with nVidia, GeForce 6600GT runs like hell on 169.21 Leave it be, I don't care anymore and lock this up as a favor

Not highly suspicious but indeed the result looks too high for the clocks given, compared to the other runs. Best example would be 3DMark03, even 2k1 @ 59k is highly unlikely. But let's leave it as it is, don't wanna argue anymore.

I sincerely hoped you guys have some kind of system that logs all submitted and deleted results. Well, noone knows anything again, it's better to lock this topic before something is said that both parties could and would regret . Cheers, N.

I try to keep this debate on decent civil level. Two of us clearly provided evidence that's imposible to get 1000+ FPS without MIPMAP tweaking (yes, that's not only nvhardpage, but d3dtweaker aswell), you bluntly shut us both down stating the scores are within regulations. Dear hwbot, they're not. Has any of you guys run 8800GTS 512MB @ e.g. 1050/1150MHz, IQ only affected by LOD and got more than 910-915FPS in nature (okay, nVidia 790i platform gets you 5-10FPS extra, but none of the guys in question is using nV790i)? @rich: I've got outruned in many categories and never complained about it, but i hate when people do not provide enough validation. Hwbot.org is flooded with cheaters, it's hard to keep cool mind.

Ha ha, it's funny, just take a look at the log. Those exact scores were on hwbot.org previous week and they got deleted, but this time they don't. Bias or unbiased - You can play ignorant on me, that OK, i'm getting used to it around here

Hipro's has is an GTX, it has two more 64bit mem channels and 8 more ROP units! I show no respect for the HWBot service? It seems your service is showing no respect for other benchers, treating them like they don't know the benchmark and stating "well, you're missing something out". There's no magical science in Nature. C'mon.

Fine for me then, expect 107k 8800GTS singlecard, LOL

It's impossible to run 1000+ FPS on ~ 1050/1150 without MIPMAP tweak. Normal score should be around 910-915FPS at LOD 4.9 (optimal) and very high PCI-E bandwidth. This one is definitely fishy! Besides, check - OPB has NVHardPage shortcut in his tray The only time i did 1000+ was with forced textures to low MIP level @ 1050/1150MHz. It's funny how you make fun of people like they don't know the benchmark.

Ofcourse it's unbiased. Those scores we're once before submitted to ORB, i warned hwbot about the lack of proof, they were removed. Now, a week later, they're on hwbot.org again. How come then the first time they we're removed & now they're not? It's good to know the new rules are a joke then and don't mean s**t after all. Now all the cheating guys who almost ran out of bussines due to new rule enforcement will quickly learn how to adapt the scores to look more 'valid'. LOL... I dare all Photoshopers to continue their legacy, it's as easy as before, just watch out when adapting your scores with the life-like MHz speed, hehe! Kill your scene. Compare this & score, almost identical Nature despite some difference in memory speed. 59k is unrealistic with 2600XT yet with 2600PRO.

Those scores does not comply to the new rules which forces users to show the subtest results. Nothing personal, KtP, but the moderator's explanation was 'score not suspicious'. How come it isn't? Many runs with similiarly clocked cards (obvious in 3DM03) get much lower results, no ORB or subtest provided, but moderators are still claiming there's nothing suspicious. So, even by box-checking that you agree to the new rules doesn't mean a crap around here? http://www.hwbot.org/result.do?resultId=741315 http://www.hwbot.org/result.do?resultId=741316 http://www.hwbot.org/result.do?resultId=741317 http://www.hwbot.org/result.do?resultId=741320

Just don't give me the bullcr_p about the flashed 8500GT card. He had like ton of hardware in his hand, he should have spotted the difference between the 128-bit card & 64-bit card in the first second he took it in his hands. Besides, all of his 8400GS runs lacked validation, so that was the main reason for my suspicion when i first saw too high scores with too low clocks (and - ultimately - too high bw & pixel/texel fillrate). @lunatic@98: I understand you want defend you teammate. But, because of that your stance is bias, you're risking credibility for a person that lost his in the eyes of the many a while ago. @knopfler: I do not know how to react when i see someone's making a joke out of everyone else who's benching the same videocard expect of how i reacted. I am like the district attorney, hwbot.org is the judge. I can make accusation with proper proof (which i provided) and take him to hwbot.org court, hwbot.org is the judge who gives the ultimate decision. He was found guilty, perhaps not by all points of verdict, but guilty anyway. I did not rushed into the forums and stated 'look, here's the cheater with too high score', i explained every bit & part of my concerns and only then stated my belief'.

Bwanasoft is like OJ Simpson then Re-check my first post, have i named him cheater? I did not! I have just simply explained my concerns and gave proof and also possible explanations of how this results could be obtained. Bwanasoft had never ever tried to defend himself although i did not attack him personally, i just wanted him to explain it. He failed at it, as it is obvious. This is not the first time Bwanasoft was accused and also proven guilty of photoshoping. He lost his reputation a while ago.

Indeed as they should be. I'm just trying to point out that we ain't such crying newbies as he thinks. He didn't present not a single contra-argument in his defense, only usual whiners-talk-how-evil-we-are-to-always-pick-on-his-results...

No, Bwanasoft, it's not wrong to benchmark the videocards, it's just plain wrong when you wanna make idiots of us when you try to sell me some shitty excuse. Aldi videocard is made by MSI, so no excuse there. 8500GT and 8400GS is literrally the same chip that only lacks 4 ROP units and one 64-bit channel. You can persuade the drivers very easily with some BIOS editing and *.ini modding. Just stop it, dude.

For instance - 8400GS Aquamark3 run - http://www.hwbot.org/result.do?resultId=722059 - Bwanasoft scores pretty darn high with very low clock speeds - after careful observation of GPU-z, i spotted a very weird "Bandwidth" - twice as high as it should be at these memory frequencies on one 64-bit memory channel. If that's true, bwanasoft's 8400GS memory should run @ 926MHz (DDR-1852). - Secondly, Pixel & Texel Fillrate are way off! Before you start with excuses such as "yeah, GPU-z lies" i give you my result with pretty much higher frequencies with the SAME GPU-z version and it shows both bandwith & fillrate correctly. Here: http://www.hwbot.org/result.do?resultId=736311 I have a very bad feeling that bwanasoft deliberately photoshoped the 8500GT 128bit result because of the following reasons: - Pixel & Texel Fillrate is higher due to 8 ROPs on 8500GT (note that on nVidia based card 4 ROP units are binded to one 64-bit memory channel) - fits the profile at this clockspeeds. - Bandwidth is off the scale for 64-bit card, it fits 128-bit (ergo, twice the width) perfectly The same suspicion goes on his 2001SE result - he's scoring 27427 points with card clocked at 570/463MHz, Conroe@4.4GHz. I've ran mine @ 710/510MHz (maximum stable frequency with no voltmods) with Penryn @5.15GHz scoring 23k and up. Only affected IQ with LOD bias range, hell not even MIPMAP can give you that kind of a boost on a 64-bit card (tried that aswell)... 100MHz on shaders brings about 4-5FPS in nature, 10MHz on memory gains you more than 50MHz on GPU. 27k with 64bit mem @ 463MHz? Ha ha ha. His 3DM05 run - http://www.hwbot.org/result.do?resultId=722056 - looks faked aswell, but i have no solid proof other than he provided no ORB since i haven't tested it yet thoroughly. Isn't it about time for him to stop doing that? :woot:

Do not add it, or - if you must - add without points.

F**k you bwana, why the hell did you delete my 6600GT 2k1? And what the hell is wrong with AXP 2800+ SPi1M? Digital photo of the screen is far more of a proof to me than printscreen screenshots. Check the f**king nature, car, dragothic, lobby high scores for the resolution comparison, jesus f**king christ! Where in the whole wide rules there is a statement that the verification screenshot cannot be a still photography image? It's still screen-shot. Hey, my Nikon D40x packaging says "Nikon - at the heart of image"