Benchmarking 101 - Gatt Edition.

[Synth]

ITS OFFICIAL.  Just make your own benchmark thread and post pics as evidence.


Rev 05

Got a hot rig you want to brag about? Or maybe just want to see how your rig stacks up to others? Here's your chance.

We shall record/post top 20 scores in each category. You may enter as many times you like in any of the benchmarks
providing the rules are followed for each benchmark. You are not required to submit scores for all benches, but you may if you wish.

Additionally we might have a contest for LOWEST scores in the near future



The Tools:[/font]

[ Video Card Benching ]

#1 3dmark 2001SE 330
#2 3dmark 2003 340
#3 3dmark 2005 1.1

All found at http://www.futuremark.com/





[ Processing Power Benching ]

#4 CPUMARK 2.0 -
http://www.addict3d.org/index.php?page=downloadfile&ID=2794

#5 Gatt PiFast - to be attached.

#6 Super PI_Mod - attached below.




[ EVIDENCE ]

#1 cpuz 1.27 or higher - http://www.cpuid.com/

Screenshot of cpuz with score screen in same pic.










If you can't keep or use a particular o/ced speed FOR AT LEAST A FEW HOURS, THEN DON'T POST BENCHES PERIOD!!!


Suggestions and comments are welcomed.





Cheating:
All persons found cheating will be humiliated publically.                    

[TrinireturnofGamez]

AMEN!
like everyone else does it .    

 i also recommend PCmark 2004 for over all system benchmark , it takes into consideration , HDD read/write , application performance , and 3d performance . Even if you have an older video card , you can still manage very impressive scores .
                       

[Crixx_Creww]

...
this was my idea.......


synth is reading minds and stalking people...

this was mine... i had a GRAND plan for this....
baka                    

[MASTER_RAGE3]

...
this was my idea.......


synth is reading minds and stalking people...

this was mine... i had a GRAND plan for this....
baka

that happens alot u have an idea and then the next day u see it out there that does get me vex                    

[Synth]

AMEN!
like everyone else does it .    

 i also recommend PCmark 2004 for over all system benchmark , it takes into consideration , HDD read/write , application performance , and 3d performance . Even if you have an older video card , you can still manage very impressive scores .

Well the older PCMark was utter crap. I do have 2004 but haven't tested it yet.

That said the reason I avoid "all-in-ones" is that the scores produced by such is just a number and is not really reflective of any area in particular. Since the scores are cumulative an extremely high score in one or two areas can skew the whole score making your system seem alot faster that it really is (overall).

I will add some separate tests for other perfromance related hardware like Hard Drives soon.

That said, testing overall application performance is important too but will take more research effort and bigger file downloads.

There are millions of benchmarks out there, some "better" than others. I don't want people needlessly wasting precious bandwidth to download multi-MB files only to say its crap and choose another.
And I want to keep process as simple as possible.

Keep the suggestions coming...                    

[Synth]

...
this was my idea.......


synth is reading minds and stalking people...

this was mine... i had a GRAND plan for this....
baka

lol @ stalking.

Yes I can read minds...sometimes even my own                    

[TriniXaeno]

wonderful initiative.

glad to see my personal favs making it on the list. 3dmark iwmc, insanely popular and not without its detractors.

besides all the ati/nvidia controversy, they have delivered a solid and more importantly, simple benchmarking and burn in experience for our PCs over the  years.



mad onion iymc                    

[unforgiven]

ok so whos first?                    

[W1nTry]

What about Sisandra 2004 SP2? there are a whole suite of benchmarking utilities in there with the benefit of cross referencing various manufacturers etc. IT also gives numbers like memory bandwidth etc??? also it has information module instead of using CPU-Z etc??? And even though ppl said they were against the all ni ones the overal coverage matrix so ppl can see how ur platform compares against its reference performance?                    

[Synth]

Sorry about this thread "dying" for a while. Just a little while longer until we can start accepting results officially.

Sandra 2005 out now actually. If the 2004 scores dont differ significantly from 2005 we will use that.

Sandra mem scores shall be used. Both in buffed and unbuffed mode.

Unbuffed mode is the true representation of chipset-memory performance; In other words, a better reflection of general real-world performance.

Buffed is a reflection of peak possible bandwidth, it takes into consideration cpu mhz and cpu and mobo optimisations.

Cpu-z is harder to be faked. Cpu-z is technically more correct than sandra.                    

[W1nTry]

This may seem like a real noob question... but I am assuming that I am supposed to use Everest for a CPU view only... cause I figure it could be used for other things like mobo details, memory etc.... which I think contributes to the overall performance of the system and thus should be included... that said...  What pics do u need to validate the rest of the platform specs? since CPU-Z covers the CPU and can cover the mobo, ram, etc?                    

[Synth]

Well validating results is a complex task. The more effort taken to preventing cheating the more effort and resources it will take for user to set up and submit. I am trying to keep it as simple as possible yet mantain some kind of reproducible standard.

Fact is anything can be faked...or one could simple use anothers valid results for that matter. So there is only so much we can do as far as checking.

That said, knowing cpu speed, fsb and few other details can be enough for all but the extreme cheat.

Will adapt as problems are encountered.                    

[TrinireturnofGamez]

How bout having all people with insanely high scores go to HQ and benchmark there?  
  note : i'll be re doing all my benchmark scores in a month or so, getting a 6600gt and vantec tornado . Will be tough for anyone to beat my benchmark scores unless they have 6800 / X-800 and an a64 .                      

[W1nTry]

YUP, pplz I think that was a blatant  and direct challenge... so who's gonna take him up on this offer? hard to beat, I am sure there's someone out there with a 3500+ A64 or better and a X800 or 6800.... come on ppl step up to the plate!                    

[TrinireturnofGamez]

let the countdown BEGIN !  2 weeks till 6600gt and pwnage . 5 days till the tornado .                      

[TriniXaeno]

recently, I am finding the agp 6800 is just a few dollars more than the agp 6600GT.

is it worth it?                    

[Synth]

For 50us more its a far better deal getting the 6800 non ultra; If just so it can mantain a certain level of fps as new, more demanding games comes out. A must if you need to max out AA and AF by a default.

For that 50us you get 256bit memory bus (vs 128 in 6600) which will help alot if AA and AF is turned on/up.

All of which help contribute to enhancing the useful lifetime of the card.                    

[TrinireturnofGamez]

i got my 6600gt for 220 , the cheapest 6800 goes for 260-270 . IF you plan to mod it to 6800gt [which i hear is possible] it IS worth it . The 128 bit bus doesn't hold back the 6600 much, remember this is GDDR @ 1ghz , same speed as on a X-800 , thats as much bandwidth as a 9700 pro has .                      

[TriniXaeno]

so for us$40 more, I can get a 6800 with 256mbit memory bus which gives me better performance with AA and AF turned up.

hmmm

sounds sweet

thats only TT$240 more. If I don't eat KFC for this month, I could afford that.

6800 immc!                    

[TrinireturnofGamez]

check www.outpost.com , they have a 6800 BFG OC going for 250 , thats if you can get it before it sells out .