Doubt about 3dmark Fire Strike>tessellation

Hi,

I have a doubt. I configured the catalyst for 3dmark Fire Strike, I put off tesselleation, it´s not valid in hwbot??

see the bench please....

http://www.3dmark.com/3dm/270043?

http://www.3dmark.com/3dm/270079?

Thanks

I read this:

http://hwbot.org/forum/showthread.php?t=67413

But don´t say if is ok or not for hwbot

I've asked at least three times in threads to which staff are subscribed but have no answer from a staff member. Non-staff said it's ok to have tess off, and I would imagine that's the case considering the number of submissions without it. It would still be nice to have an actual answer though.

It would still be nice to have an actual answer though.

Yes!!

I got my reported 3dmark no-tess result marked as "checked" by a moderator. That's good enough for me.

I got my reported 3dmark no-tess result marked as "checked" by a moderator. That's good enough for me.

Yes, that's true, but what if tomorrow is not ok and you don´t have the gpu, lost bench and don´t can repeat

Ha that can happen even if there is an official decision right now

Difference between Tesselation on / off :

http://www.3dmark.com/3dm/275249

http://www.3dmark.com/3dm/275168

I prefer activate Tesselation , only it is valid from Futuremarks aproval.

Difference between Tesselation on / off :

 

http://www.3dmark.com/3dm/275249

 

http://www.3dmark.com/3dm/275168

 

I prefer activate Tesselation , only it is valid from Futuremarks aproval.

I prefer it too!

If they allow it off in 3DM11 I don't see why not on this one

If they allow it off in 3DM11 and Heaven I don't see why not on this one

fixed

Let's make a poll from this thread. As far as I can see, the problem is:

- tesselation on/off allowed

- only go for "this is a valid run" in the 3dmark gui or not

Main problems:

- tesselation cannot be detected in screenshot, only if the result has been uploaded to the ORB. That either requires a functional internet connection of the benchmark rig or a verification link.

- if we want to enforce the "valid run only" (for tesselation), know that Nvidia LOD adjustments are not detected and are classified as valid runs in the GUI. Why disallow tesselation tweaking, but not disallow LOD tuning (as that is not detected)

- disallowing tesselation would break consistency of rules between the various DX11 benchmarks

So, I guess the poll options are:

- allow tesselation and lod tuning

- disallow tesselation, allow lod, but not enforce "valid run" according to the 3DMark ORB

- disallow tesselation, allow lod and enforce "valid run" policy according to 3DMark ORB

If we disallow tesselation and rely on the 3DM ORB for decision whether or not it's a valid run, I don't see how we can award points for the benchmark though. As said, LOD is still seen as valid and chances are the tesselation detection might break too with future drivers.

Let's make a poll from this thread. As far as I can see, the problem is:

 

- tesselation on/off allowed

- only go for "this is a valid run" in the 3dmark gui or not

 

Main problems:

 

- tesselation cannot be detected in screenshot, only if the result has been uploaded to the ORB. That either requires a functional internet connection of the benchmark rig or a verification link.

- if we want to enforce the "valid run only" (for tesselation), know that Nvidia LOD adjustments are not detected and are classified as valid runs in the GUI. Why disallow tesselation tweaking, but not disallow LOD tuning (as that is not detected)

- disallowing tesselation would break consistency of rules between the various DX11 benchmarks

 

So, I guess the poll options are:

 

- allow tesselation and lod tuning

- disallow tesselation, allow lod, but not enforce "valid run" according to the 3DMark ORB

- disallow tesselation, allow lod and enforce "valid run" policy according to 3DMark ORB

 

If we disallow tesselation and rely on the 3DM ORB for decision whether or not it's a valid run, I don't see how we can award points for the benchmark though. As said, LOD is still seen as valid and chances are the tesselation detection might break too with future drivers.

Poll sounds good.

Is mipmap use detectable in a screenshot or does it's use cause an "invalid run" ?

I ask, because it's use was before my time benching.

Doesn't make a lot of sense to me to allow LOD and Tess adjustments in every bench except 3dmark. Definitely more confusing for new people.

The valid-stuff-only thing would be a major PITA when it comes to drivers too, it takes futuremark ages to approve drivers.

Not true. Non-beta drivers are generally approved within 48 hours now.

Also if the LOD thing is a major sticking point, I can bring it up as potentially something we could go and detect. Can't offer instant fix, but we can investigate what can be done related to it.

Overclockers -love- beta drivers

//edit: how can you not see LOD as a major sticking point for the 3DMark benchmark? Tesselation "optimisations" are clearly invalid for 3DMark submissions, but the LOD adjustments are not? Why?

Poll sounds good.

 

Is mipmap use detectable in a screenshot or does it's use cause an "invalid run" ?

I ask, because it's use was before my time benching.

The practical difference between mipmap and tesselation is that mipmap is an on/off switch whereas tesselation (and lod) is a slider. Because it's a slider, you can finetune the tweak to near legit quite easily without anyone noticing. Mipmap was upping the score quite a lot, so it could be detected with a fair amount of certainty.

But to answer your question: no, mipmap can not be detected on the screenshot.

Fyi, if we enforce "only valid runs", we also can no longer use beta drivers.

Imo HWBot does not have to follow FM rules. I would set the same rules for 3DMark "12" like we did for 3DMark 11.

Makes life much easier for benchers and we have the same rules for the different benchmarks.

LOD is quite tricky as you have to find the correct value for each benchmark and it has (apart from 3dm2001) only a small impact on the result.

The suggested poll by PJ sounds good tho.

Not true. Non-beta drivers are generally approved within 48 hours now.

 

Also if the LOD thing is a major sticking point, I can bring it up as potentially something we could go and detect. Can't offer instant fix, but we can investigate what can be done related to it.

That's not true I was on 13.1 when 3dmark came out and was they weren't approved but the 13.2 beta where lol

Just to be clear from my own perspective, I think it should be the same as 3DMark 11 & Heaven. The only reason I've been asking is to have a staff member tell me whether that is correct so I can run the bench and submit without fear of having the score removed.

Is there anyone that is opposed to having the same rules like 3DMark11 and Heaven? If not, a poll isn't really worth the trouble even

Keeping things consistent sounds good to me! If Futuremark reported LOD and Tess in the screenshot, it might be worth it to keep things legit, but since it does not (currently), let's just stick with the tried and true cheat-the-benchmark-into-higher-scores method!

Keeping it simple makes far more sense to me.

So either ban it in everything and wipe the current 3d11/heaven scores out or allow it in 3d12

Overclockers -love- beta drivers

 

//edit: how can you not see LOD as a major sticking point for the 3DMark benchmark? Tesselation "optimisations" are clearly invalid for 3DMark submissions, but the LOD adjustments are not? Why?

Tessellation adjustments are in default driver UI. You could do these without realizing that you are messing up the benchmark load -> high priority to recognize and flag.

LOD tweaks require third party tools mucking up with internal settings that the vast majority of people have no idea about -> not so high priority.

That's not true I was on 13.1 when 3dmark came out and was they weren't approved but the 13.2 beta where lol

13.1 did go through approval process - it actually did so before the launch.

It was specifically NOT approved because it didn't render the Fire Strike test correctly (part of the particle effects were missing). 13.2 betas fixed this problem. Approval is not just a rubber stamp; the driver actually has to render things correctly.