• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Thinking about a new challenge!

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Super Nade

† SU(3) Moderator  †
Joined
Aug 30, 2004
Location
Santa Barbara, CA
Hey all,

I am thinking about starting a 3DMark "Full optimization contest". The idea is to bench the card with and without software tweaks and quantify the jump in scores. The nice thing is, it would bring out the importance of software tweaks and for once, the playing field may be levelled for all. I'm not sure if people would like to share their tweaks and that is perfectly fine, but atleast the awareness about this aspect of benching among the more casual members will increase.

I'm not sure what the prizes are going to be, but that will come later.

What do you think about the idea?

:)

Ideas in no particular order:

  1. Use only one kind of video card. Fixed clocks for all.
  2. Use any Video card.
  3. Limit Benchware to 3DM'01
  4. No AGP
  5. Multiple entries/cards per member.
  6. Slow GPU sub-contest.
 
Last edited:
So basically, your suggesting that the same "glitch" that bugged the last contest (7600gs reffering too ?) would be allowed in this contest, even encouraged ?

Sounds FUN !!! :D

Possible card ideas ? How about a X1300XT :p
 
Hmm...I'll have to look more carefully into that. Well, why don't you ping me with ideas? Different card classes to please both the NV and ATI folks?

I'm pretty much open to any idea as long as it proves to be fun. One idea I'm dead set on implimenting is to have a SLOWEST GPU sub-contest as well. What do you think? :D
 
What is being looked at is how software tweaks can help boost scores, keeping clocks the same, right? So, the specific card being used should not matter if it is from within a certain class of cards. For example, if I pick the x1xxx series, apart from clocks and pipes/core these cards are transparent to software tweaks, right?

Since clocks are going to be constant, and any gains accrued would by via software tweaking only, the card should not matter.

I'd like to hear more from you guys about this. :)
 
Sounds fun , so it would be a stock test vs a tweaked test or would it be a stock test vs a overclocked tweek test ?

I would be intrested in this sounds fun . especial since the ppl with the fastest cards wont be walking all over every one . Plus im always game to learn new tweaks that work =)
 
Any way you want it. In the end the scores with and without software optimizations will be subtracted. Clocks will remain the same for both tests. :)
 
I like the idea. The only problem is that my cards don't seem to benefit much from tweaking because they are so slow already, heh.
 
I think this contest might not be quite possible because of this-

electrorcamd said:
I like the idea. The only problem is that my cards don't seem to benefit much from tweaking because they are so slow already, heh.

Unfortunately I don't think there's any way to even the plane; I'll need to sleep on this one.
 
random thoughts on what you might be able to do.

try 01 and figure out something that factors in natural abilitys, IE the inagrated 6100 in my back up rig definetly won't give me the same performance boost from tweaks as 2 x1900xtxs in CF.

You could maybe try a percentage based thing in 03 or aquamark though becuase they seem to be more balanced between cpu/gpu for everyone.

If we can't do contest, maybe somebody will find the time to do a test for any tweak someone can think up and see how each does.

What else might be an interesting just to see what happens comparision is how big of a difference you can get between two scores at the same clocks.
 
'01 would be the best (most suceptible to tweaks), and Speed's idea of using the same clocks is also what I would suggest. Go for everyones % score increase, that would work fine. Hey if you can't figure this out maybe you should ban yourself ;)
 
Points noted Maxi and Speed. I'll tell you what, I am going to make a note of all your ideas on the first post. Then convert this to a poll with those options. The final decesion will rest with YOU folks who will be benching (ofcourse I will participate and try to use what little stuff I know).

Sound o.k?

Nobody likes my % slow GPU idea? :cry:

Ban myself eh? :p
 
Euuuuhhhh this would require running a bare stock OS for the non-software tweaked runs to be there lowest for the highest % gain.
I can't stand running a bare stock MS OS :p

Just had a thought.. there might be some OS tweaks that could be reversed to make the runs slower?
Could that be potentialy harmfull? Just wondering.
 
Super Nade said:
I'm pretty much open to any idea as long as it proves to be fun. One idea I'm dead set on implimenting is to have a SLOWEST GPU sub-contest as well. What do you think? :D

Count me in for that, I keep meaning to get round to benching the TNT 8mb, 16mb and a Voodoo 3 16mb cards I have around here.
Maybe in the sub-contest you could have the fastest 8, 16 and 32mb cards. Open to any card per MB size.
It would stop the conroe's running away with everything as it would have to be AGP

Any how just my two cents.
 
greenmaji said:
Just had a thought.. there might be some OS tweaks that could be reversed to make the runs slower?

bingo! Not even tweaks just open a couple of instances of Orthos ;)
 
A couple of instances of orthose has got to kill a 3D score.. I think we are above that right?
 
Back