Futuremark Publishes Guidelines for Driver Optimizations

[Nautis]

You wont find me making any excuses for nvidia.. there is no doubt it makes it easier for doing some of the nefarious things they are doing.

Is it worth it? Seems to have been so far.

How much of a problem are modified drivers? Not much but its a shot in the ego when more and more people are using modified drivers rather than your own. Alot of hard work goes into drivers to get them to work the way they do and they dont like people messing with their work. They also dont like people fixing their problems.

 

[kyleb]

so what is the deal with this anyway? i just checked futuremark's homepage and right there is a link to the get the 45.23 drivers which i belive break the rules futuremark set forth. do the rules really have any value then our are they just for show?

 

[Pete]

worm (or any other FutureMark employee who is qualified to answer),

How do nVidia's current drivers square with your newly clarified policy?

 

[kyleb]

weird, the last three posts do not show in this thread for me, but they do show in the topic view of the reply page.

hu, fixed now.

 

[bloodbob]

I think the ultimate answer here is gonna be to test we are going to have to either A) use the DVI B) not travel to USA or risk prosecution for violation of the DMCA. Because after last time nvidia got busted they got shamed and I don't think they plan on getting stuffed again. Most likely the will check if the camera has left the camera path way and all screen captures will be rendered at full quality and they will simple say screen captures take a preformance hit.

 

[madshi]

legitimate? To keep people from modifying them and releasing them (ex. omega).

If I remember right, omega said that he would not be modifying any driver executables/dlls. He just puts files from different driver versions together and changes a lot of settings. So as far as I can see, encryption should not be a problem for such modified drivers.

I think the main reason for the encryption is to hide the internal workings like application detection etc...

 

[Nick[FM]]

We are in the process of finalizing the QA now. It will be posted later today in PDF format. It took us a day longer than expected due to so many questions and we wanted to answer as many as possible. Thanks for all the questions and interest in the new guidelines!

so what is the deal with this anyway? i just checked futuremark's homepage and right there is a link to the get the 45.23 drivers which i belive break the rules futuremark set forth. do the rules really have any value then our are they just for show?

Err, we have always posted links to the latest drivers. Even to beta and leaked drivers. I'm not really sure what you have against us posting news about new drivers, and link to them?

So what has Fururemark done now after the guidelines have been released? Have they droped scores off the ORB that used "illegal" drivers? Have they done anything?

We have been writing the QA and doing a lot of other things besides that. We haven't dropped any results from the ORB.

It seems that many of you think this is something so easy to do, and could be done in an instant. It's not exactly a "walk in the park". We are dealing with a large amount of BIG companies (and websites ), and that simply takes time.

Anyway, I'll give you guys a hint when the QA is online!

 

[PatrickL]

Worn, you really don't see the irony to link drivers that are breaking FM own rules ?

 

[Quitch]

Uh... it is easy to do. What work? There have been countless threads analysing the latest drivers, showing the cheats. I say again... what work? Fire up the drivers, confirm it's true, drop the results. You don't even need to find more than one of the cheats.

What were your tech people doing while your wrote the statement? I mean, did you really have NO idea what the guidelines will be. The longer you leave it before taking action, the longer they look like window dressing.

 

[kyleb]

ya worm, do you really thing that you should distribute drivers that do not meet your own guidelines without even commenting on that fact? do you not see how deceptive it is do do otherwise? please don't use lack of time as an excuse either as this has been going on for the better part of the year now and we have got nothing but lip-service from you all.

 

[Joe DeFuria]

ya worm, do you really thing that you should distribute drivers that do not meet your own guidelines without even commenting on that fact? do you not see how deceptive it is do do otherwise? please don't use lack of time as an excuse either as this has been going on for the better part of the year now and we have got nothing but lip-service from you all.

Just give them until Nov 1 before we start demanding things like this.

FM has set a very good and respectable course of action at this time. Give them a chance to execute on it.

I was just as hard as you, if not moreso, on how FM handled the aftermath of their reaction to the original cheat PDF. I am also disappointed that their response took this long in coming...but they have in fact now responded.

 

[digitalwanderer]

I'm with you on that one Joe.

From what they've put out in their guidelines and the way they followed up and expanded on it in their Q&A, (in a pretty timely fashion too...even by my standards ), I'm more than willing to give them 'til Nov 1st before I start complaining.

After Nov 1st though if it hasn't been instituted and acted on...

 

[Pete]

Optimization Guidelines QA

Edit: A more appropriate link.

The Q&A also answered my Q regarding allowed drivers: FM will post what drivers they do and don't approve of by Oct. 30th.

 

[kyleb]

Just give them until Nov 1 before we start demanding things like this.

ya my initial question was asked prior to the publication of the q&a, but demeanor of worm's response as well as the the fact that they are now apparently postponing the actual exposure of nvidia's cheats, until we get closer to the crhistmas season, absolutely wreaks of foul play to me and i can't rightly see much changing by october 31st either.

 

[Quitch]

So we need to wait till 1st November? When did this all start? February? March? Wow, so we can really count on FM to act like lightning.

 

[WaltC]

What I'm interested in learning here is whether or not FM is going to include vendor code paths in its software in the future. I asked Worm about it in his other thread here but so far he hasn't replied. It seems to me a question in need of a definitive answer--because of statements they made earlier about considering such code paths in the future. I can't see how they might ever do that without invalidating all of their other rules and guidelines, but I'd like to hear a declarative statement by them to that effect--one way or the other. My concern is that inclusion of such code paths would move the "optimization" from the drivers into the 3dMk program itself, thereby completely invalidating its results when used to compare differing IHV products. Maybe they said something about this earlier and I missed it. Anyone?

I also think that releasing further recompile patches on a regular basis might be by far the easiest way to prevent driver ident going forward. For instance--if they released such patches quarterly--then it really wouldn't matter whether an IHV built in detection routines for a specific driver, as the next recompile patch would break them. Seems to me that would be *much easier* for them than worrying over and rigorously testing each and every driver the IHVs put out, beta and otherwise. In fact, it seems to me it would be so much easier that if they don't do something like this going forward I'd really like to know why.

In fact, they could parley it into a positive PR spin for the company--and call it something like "Cheat Guard" (TM), "Designed to protect the credibility and impartiality of our software each and every quarter." But really, none of this means anything if they are going to code in vendor paths into the benchmark...I'd really like an answer on that one.

 

[digitalwanderer]

What I'm interested in learning here is whether or not FM is going to include vendor code paths in its software in the future. I asked Worm about it in his other thread here but so far he hasn't replied. It seems to me a question in need of a definitive answer--because of statements they made earlier about considering such code paths in the future. I can't see how they might ever do that without invalidating all of their other rules and guidelines, but I'd like to hear a declarative statement by them to that effect--one way or the other.

(bolding mine, duh. )

I totally missed that bit, did Worm ever indicate if they will be incorporating seperate vendor code paths in 3dm2k3?!?!

Oh, and I agree with the rest of what you said too Walt and very well put...as usual.

 

[Pete]

(Very OT: You all see how long it took for some corporations to compromise for the customer's benefit? Maybe this'll put the Iraqi reconstruction or the North Korean nuclear talks in a different light for some people. And it's not just money that's at stake in those situations, but power.)

 

[Neeyik]

ya worm, do you really thing that you should distribute drivers that do not meet your own guidelines without even commenting on that fact? do you not see how deceptive it is do do otherwise? please don't use lack of time as an excuse either as this has been going on for the better part of the year now and we have got nothing but lip-service from you all.

Firstly, FM do not distribute the drivers in any shape or form; they simply pass on the news that other sites have them available for download. Secondly, whether such drivers meet the new guidelines or not is irrelevant, as these "rules" are for their use in publishing 3DMark scores and nothing else. Thirdly, you've actually received free software from FM (re: your comments about "lip service") but hey, why the hell should that matter?

 

[Pete]

Firstly, FM do not distribute the drivers in any shape or form; they simply pass on the news that other sites have them available for download.

Honestly, FM's placing link on their site is an implicit acceptance of the linked material, especially something as relevent as drivers (which are required to run FM's products).