TASOS Posted January 30, 2017 Posted January 30, 2017 I see miracles happening in some RealBench Multithreading scores. I hope this miracles are legit. Quote
Mr.Scott Posted January 30, 2017 Posted January 30, 2017 (edited) I see miracles happening in some RealBench Multithreading scores. I hope this miracles are legit. I've been experimenting. I can tell you right now that some of these miracles are not legit. Problem will be, as always, proving it. Asus's own tweak list for this bench is not legal for HWB under the current HWB rules. IMO, Realbench is very flawed and should not be given points under any circumstance. It's PCM05 II. Edited January 30, 2017 by Mr.Scott Quote
TASOS Posted January 31, 2017 Posted January 31, 2017 (edited) I've been experimenting.... I've done some work too Scotty. Tell me your opinion about asking the participants , to show the whole (and more maximized ... and not screenshot cut) window of the log output , as a beggining. Any other thoughts are welcome. ./ Rule reminder Disallowed tweaks/cheats Altering benchmark files or the rendering Any software or human interaction altering the perceived speed of the benchmark program, tricking it to believe it ran faster Edited January 31, 2017 by TASOS Quote
lanbonden Posted January 31, 2017 Posted January 31, 2017 I've been experimenting. I can tell you right now that some of these miracles are not legit. Problem will be, as always, proving it. Asus's own tweak list for this bench is not legal for HWB under the current HWB rules.IMO, Realbench is very flawed and should not be given points under any circumstance. It's PCM05 II. Yea this boosted "heavy multitasking" really have got me wondering as well as the boost there is worth some 1000+ higher total score Quote
Aleslammer Posted January 31, 2017 Posted January 31, 2017 I've had one test run where Encoding & Heavy Multitasking were both over 18000 on a fresh OS install (about the 3rd or 4th). Had some BIOS problems and it will not reproduce at present. I'm guessing it wasn't tested for 939 platforms so could have a hiccup or two or our systems maybe a little tried after all the runs on them. After this is over would be interesting to know what was used to kick up the scores. Never have liked the bench that much, dislike advertisements although the newer one is better than the first. Quote
Mr.Scott Posted January 31, 2017 Posted January 31, 2017 I've done some work too Scotty. Tell me your opinion about asking the participants , to show the whole (and more maximized ... and not screenshot cut) window of the log output , as a beggining. Any other thoughts are welcome. ./ Rule reminder Rule reminder tells me that we are on the same page my friend. Quote
TAGG Posted January 31, 2017 Posted January 31, 2017 (edited) The tweak is not all that complicated... it's basic priority stuff Just think about XTU and you'll find the tweak... I can assure that all files are 100% stock And yes Realbench is a crap benchmark, just like XTU and PCM imo... Edited January 31, 2017 by TAGG Quote
Mr.Scott Posted January 31, 2017 Posted January 31, 2017 The tweak is not all that complicated... it's basic priority stuff Just think about XTU and you'll find the tweak... I can assure that all files are 100% stock And yes Realbench is a crap benchmark, just like XTU and PCM imo... Changing resolution changes the rendering also. No files involved, and no way to prove that it was changed. Quote
TAGG Posted January 31, 2017 Posted January 31, 2017 Everything is run at 1050p without any driver tweaks... And yes I know there is no way to prove that... If nobody else publishes the tweak, i'll make a tutorial on it After the challenge of course Quote
Mr.Scott Posted January 31, 2017 Posted January 31, 2017 I think I have it. Thank you for being accommodating though. It is appreciated. Quote
Administrators websmile Posted January 31, 2017 Author Administrators Posted January 31, 2017 When Frank suggested the benchmark to be used we had more or less a promise it would be reworked because we found tons of holes in it especially at realbench challenge - nothing happened, one of the rare occasions hwbot is not to blame... Quote
lanbonden Posted February 1, 2017 Posted February 1, 2017 When Frank suggested the benchmark to be used we had more or less a promise it would be reworked because we found tons of holes in it especially at realbench challenge - nothing happened, one of the rare occasions hwbot is not to blame... So we are free to abuse the "heavy multitasking" score? Or should I focus on geting a legit run as the abusing scores will be deleted? Quote
Administrators websmile Posted February 1, 2017 Author Administrators Posted February 1, 2017 This is a good question, wanted to post anyway - you all can tweak the heavy mt for this contest as we now have already a lot of subs with this, I want same chances for all and easy work for me, if I see some "tweaks" at test 1 or 2, I will block because on these there is not even a grey zone. On test 4, I will accept certain limits bc there are several ways to boost this Quote
TAGG Posted February 1, 2017 Posted February 1, 2017 I may have discovered a tweak for #1 which has to be 100% legal, since it's basically hardware, not software Will post more about this once i have verified my theory Quote
TAGG Posted February 2, 2017 Posted February 2, 2017 Why would you want quad sli anyways? The framerate is mostly limited by memory bandwith in this bench i think Quote
lanbonden Posted February 2, 2017 Posted February 2, 2017 Anyone here having trouble getting 3dmark to run with two 7950 gx2 in sli?Tells me no benchmark can be ran on my system. But it will run with sli off. I understand that quad sli is fun, but this bench doesnt even suport sli so no use trying Quote
Alpi Posted February 2, 2017 Posted February 2, 2017 I understand that quad sli is fun, but this bench doesnt even suport sli so no use trying Do You think this java thing doesn't scaling good for multi gpus ? Quote
pipes Posted February 3, 2017 Posted February 3, 2017 i can't submit in this old school season 3? Quote
Administrators websmile Posted February 3, 2017 Author Administrators Posted February 3, 2017 Depends on if you use correct hardware and for realbench you need to bench with internet enabled so you can submit directly after benchmark, save file doesnt work. Quote
Administrators websmile Posted February 3, 2017 Author Administrators Posted February 3, 2017 For me there is, a big red button on top that says "open for submissions" and when you move the cursor to it, it says "submit" - can you make a screenshot of your eocsports stage page for oldschool ? We also have plenty of submissions already, so it might be a bug of your rig or browser... Quote
pipes Posted February 3, 2017 Posted February 3, 2017 there was also yesterday the submission button, but when I pressed it gave me the popup window without any additional button for 3 benchmark now there is. Thanks Quote
TAGG Posted February 4, 2017 Posted February 4, 2017 Got me some more stuff to throw at the round Quote
Plasma_Carrot Posted February 6, 2017 Posted February 6, 2017 Can we not ever never never use Real Bench again for Old School competitions. It's over 30 minutes of mind numbing boredom with the hardest test at the end, so you have to wait a long time for the inevitable system crash. Quote
Plasma_Carrot Posted February 6, 2017 Posted February 6, 2017 Have to be online for real bench? Crap my system does better with the Nic off. I already only have only 1gb of ram to use. Anyway I'm finding out that the damage to my motherboard includes some dead PCI-e lanes. Stuck a 8x and cant use 16x pcie speeds. I'm going to get reamed here. But hay I did find one of my gx2's can do 700mhz core and 800mhz memory on stock voltage and hardly even artifacts. So maybe I will not come in last Got to love broken parts. My other gx2 is covered in rust, yes rust. Its amazing I can even get a score with my parts. Anyway this is looking like a pretty cool round, can't wait to see the last of the subs that tend to come in at the last moments. I made the mistake of getting 2 GX2's when I saw the rules on multi GPU's only to find 3D Mark Ice Storm hates Crossfire. Took a while to get them working in 4 way as well, then I saw the score :-( Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.