FM_Jarnis Posted September 19, 2011 Author Share Posted September 19, 2011 (edited) Oh, interesting. Did not know about OB's connection with Nvidia! I don't think there's faul play here, though. If FM was biased like that, they would never have disallowed GPU physics. Trust me, there is no bias. Both AMD and NVIDIA hate us equally Edit: Well, one week one of them hates us more, the next week its the other. Depends a bit on what piece of hardware they are shipping at the time and how it performs. Meanwhile FM just measures the performance and reports the gory details, always striving for unbiased benchmarks. Edited September 19, 2011 by FM_Jarnis Quote Link to comment Share on other sites More sharing options...
FM_Jarnis Posted September 19, 2011 Author Share Posted September 19, 2011 Ok, so no tessellation, no lods.What is overclocking a vga about, then? Pure frequency? Oh, interesting, so the skills will be the amount of binning needed to find the right vga, the right cpu etc... I strongly disagree. I can understand the decision behind prohibiting the mipmap tweaking: it was not a "tweak" it was a blank screen from the beginning to the end of the bench. But the lods... don't change the rules even on that please -.- I'm not saying we are doing anything with LOD tweaking yet. I'm just saying we are open to the idea if OC community thinks that detecting it would be useful. Tessellation thing has to be done - without it, AMD runs could be done with major benchmark components completely disabled and with completely different looking image. That just isn't right and that is why no AMD driver since the addition of tessellation controls have been approved. Our goal is always that all hardware run the exact same work, preferably rendering exactly identical images and then the benchmark tells you which piece of hardware is best at doing it. Changing the rules by changing the workload makes any comparisons flawed. Quote Link to comment Share on other sites More sharing options...
|ron Posted September 19, 2011 Share Posted September 19, 2011 Well, speaking about tessellation, you made the point clear... if nvidia isn't going to release a driver version in which you can control the tessellation, then this feature must be ON for both of them. But for the lods, I disagree even on detecting them... it's not a difficult thing to spot the right lods, but hey, it's time consuming and it's not correct that someone will benefit from my work, except if I want to share... and usually, I always give a hand to whom is asking my lods etc... Hope to have cleared out my point too Quote Link to comment Share on other sites More sharing options...
FM_Jarnis Posted September 19, 2011 Author Share Posted September 19, 2011 Well, speaking about tessellation, you made the point clear... if nvidia isn't going to release a driver version in which you can control the tessellation, then this feature must be ON for both of them. Even if they would, we would probably do the same thing. What is the point of benchmarking DX11 cards for doing major DX11 features (Tessellation) if drivers can silently remove a major feature workload ("disable tessellation"). The score from such a run would be pointless for figuring out the DX11 performance of the hardware. Unless you plan on playing every DX11 game with Tessellation Off as well? Quote Link to comment Share on other sites More sharing options...
|ron Posted September 19, 2011 Share Posted September 19, 2011 Gaming and overclocking are two different things, at least speaking about extreme overclocking, if we could call it so. My purpose in benching a vga is to squeeze out every fps: back in the days of the radeons 9700/9800/x850, the 3dmark03 was a COMPLETE purple, green and blue artifact... who cares about graphic quality, the important thing is the score and being able to let the vga close the bench without crashing. If Tessellation is still there a little, even with tessellation slider at minimum in the CCC, then for me it's ok. It's EXACTLY the same as putting drivers on "high perf." instead of the useless "high quality". That is what 3dmark is used for, not for benching with shining graphics, magix effects etc... at least, not in the overclocking environment Of course, a score must be obtained using the same rules, that is why I temporarily agree with the "blocking" of tessellation-off-scores from ati, at least until nvidia will decide what to do about their drivers. But Lods is a different thing. Quote Link to comment Share on other sites More sharing options...
Massman Posted September 19, 2011 Share Posted September 19, 2011 If tesselation can be reduced for Nvidia (which is possible), then what would you prefer, |ron ? Quote Link to comment Share on other sites More sharing options...
|ron Posted September 19, 2011 Share Posted September 19, 2011 Hi Pieter, according to me the solution is a MINIMUM level of tessellation mandatory for both ati and nvidia... just like when you set drivers to high perf. Tessellation to 0x is like a sort of mipmap tweaking, so if mipmap is contrary to the rules, then tessellation totally disabled should be as well. Don't you agree pieter? what is your opinion about that issue? And jarnis also, what is your opinion (not futuremark one, I mean your thought)? Quote Link to comment Share on other sites More sharing options...
Massman Posted September 19, 2011 Share Posted September 19, 2011 My thoughts on this were already posted here: http://hwbot.org/forum/showthread.php?p=129486#post129486 Quote Link to comment Share on other sites More sharing options...
|ron Posted September 19, 2011 Share Posted September 19, 2011 Mmhh the link is to the latest post by Chew, so you basically agree to say that tessellation is not THE feature, or at least you agree with chew about using a minimum amount of this effect (of course, given that nvidia will add this feature to their driver some day) because that doesn't hurt visual quality even being at 5%? Seems more or less what I said, if I'm not mistaken, no? Quote Link to comment Share on other sites More sharing options...
Massman Posted September 19, 2011 Share Posted September 19, 2011 I posted earlier in the thread. Quote Link to comment Share on other sites More sharing options...
Maxi Posted September 19, 2011 Share Posted September 19, 2011 There is no difference between LOD and tesselation on a basic level, both effect image quality. Also maybe I'm wrong, but I don't think it's possible to completely disable Tesselation with the slider, just reduce it. Quote Link to comment Share on other sites More sharing options...
SF3D Posted September 20, 2011 Share Posted September 20, 2011 But reducing tesselation level have a lot higher impact on image quality than just LOD. Like Maxi said, they are same kind of things on basic level, but tesselated objects are far more advanced. If we could have tool to set tesselation to 0 on both vendors there is no issue. If only vendor A can do it, but B can't, it is the issue of vendor B or their users. Then we also have to decide do we allow it to be used after all or does it take the whole idea of a benchmark away. If I could decide, reducing tesselation would not be allowed. LOD was the thing of it's time, but we should move forward. Quote Link to comment Share on other sites More sharing options...
Massman Posted September 20, 2011 Share Posted September 20, 2011 Assuming we want to disallow altering the Tesselation level, there are still the following problems: - we can't enforce this policy on Heaven DX11 (unless we add detection mechanisms to the wrapper) - we can't see it in the benchmark GUI (so screenshots will be meaningless) - we can't see it in the ORB submission page (at least not for a few weeks) Assuming we don't want to allow it, for instance by figuring out how to reduce tesselation on Nvidia cards too, we have potential problems if: - FM's implementation of the Tesselation detection in the benchmark GUI is to remove E/P/X score indications or something similar that makes it impossible for us to verify the score Either solution has its problems. // edit // Assuming we can find a solution for the Heaven benchmark, the only 'sensible' thing to do is disallow any AMD-based submissions for both Heaven and 3DMark11 untill the necesarry fixes have been set in place. This is only a good solution, though, if we will eventually be able to detect the tesselation level in both Heaven and 3D11 and assuming we will not be able to reduce tesselation level on Nvidia-based graphics cards. This would mean 1777 results of 438 users would be considered invalid. Quote Link to comment Share on other sites More sharing options...
knopflerbruce Posted September 20, 2011 Share Posted September 20, 2011 Erm, to disallow subs for weeks shouldn't be an option. Still we have to deal with existing submissions - my suggestion is to disallow this for the 7xxx series and up, assuming we can verify the tesselation level without clicking the orb link. Quote Link to comment Share on other sites More sharing options...
Christian Ney Posted September 20, 2011 Share Posted September 20, 2011 nvidia's phone number ? Quote Link to comment Share on other sites More sharing options...
Maxi Posted September 20, 2011 Share Posted September 20, 2011 (edited) How someone can argue for LOD yet want to ban this is beyond me. Nvidia benefits far greater than ATI with LOD tuning, even now...yet when ATI could not use LOD everyone running green cards laughed. Now that ATI is benefiting, as some would argue unfairly, it's unacceptable IMO if we sanction this, we also have to ban every D3D Overrider result @ hwbot. This would mean 1777 results of 438 users would be considered invalid. Cool! Edited September 20, 2011 by Maxi Quote Link to comment Share on other sites More sharing options...
Massman Posted September 21, 2011 Share Posted September 21, 2011 FYI, tesselation can be disabled completely and it'll make the benchmark rediculous. So allowing it is no option either. See attachment. Quote Link to comment Share on other sites More sharing options...
|ron Posted September 21, 2011 Share Posted September 21, 2011 Using the slider in the CCC and setting tessellation to 0x, wasn't resulting into this, to me! It was all there and visible! How did you manage to get this one pieter? by disabling tessellation in the benchmark itself? Quote Link to comment Share on other sites More sharing options...
Massman Posted September 21, 2011 Share Posted September 21, 2011 The official driver doesn't allow this, but there are unofficial tweak applications that do allow turning it off completely. It pretty much makes any AMD D3D11 scores suspicious now. Quote Link to comment Share on other sites More sharing options...
|ron Posted September 21, 2011 Share Posted September 21, 2011 How about modifying the benchmark making that it automatically takes a SS in a precise frame of the scene. Then making mandatory to attach this SS to the submission. Of course, it would also be necessary to hide some infos inside the jpg, to let you make a validation in the submitting process. That way, by looking at the SS, we could easly understand if tessellation was disabled. Maybe a bit too complicated? Quote Link to comment Share on other sites More sharing options...
Massman Posted September 21, 2011 Share Posted September 21, 2011 That's a good idea, actually. We could make a reference screenshot ourself and check the the image quality against the screenshot taken by the application. Certainly not the nicest solution, but 'doable' unless we get information on how to access the atimgpud.dll to read the tesselation level. Quote Link to comment Share on other sites More sharing options...
K404 Posted September 21, 2011 Share Posted September 21, 2011 Nvidia benefits far greater than ATI with LOD tuning, even now...yet when ATI could not use LOD everyone running green cards laughed. Now that ATI is benefiting, as some would argue unfairly, it's unacceptable It's not quite the same situation Both nV and ATI had the option to modify LOD and they both reacted to changes. It's not nVidias fault that ATI didn't scale so much with LOD changes. Now we have a situation where ATI have control over something nV have not yet implemented. I hope HWBs solution to this takes into consideration that nVidia will probably include the same tesselation level tools in their drivers at some point Quote Link to comment Share on other sites More sharing options...
Maxi Posted September 21, 2011 Share Posted September 21, 2011 There was a time when LOD was Nvidia only, mainly because there was no common app people could use to tweak it. Then W1zzard came along with it built into ATI Tool. Massman is it also not the same with Nvidia, the ability to disable it? I thought I remember something like that going around. Quote Link to comment Share on other sites More sharing options...
K404 Posted September 21, 2011 Share Posted September 21, 2011 That might have been before my time. Has LOD for ATI been around as long as HWB? Has it impacted competitive benchmarking? Quote Link to comment Share on other sites More sharing options...
Maxi Posted September 21, 2011 Share Posted September 21, 2011 I don't recall exactly, probably around the same time. Before that, yes it definitely effected the ORB rankings. Often what we do with a benchmark contradicts the basic intent of the software. Tweaking blurs the results and removes uniformity, so you wind up with results that often do not have real and meaningful numbers, which is what the test is designed to provide. With this situation, it would probably be best considering everything to enable it on both platforms. The question is how. Quote Link to comment Share on other sites More sharing options...
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.