Ideas wanted!!!

[flexkill]

OK, I want to get some sort of bench mark comp going. If you have any Ideas that we could use please post and let us know.

Remember they have to be as "cheat proof" as possible. This is what makes it so hard. So hopefully we can get a lot of ideas here and find one or two that will work.:beer:

[top game 22]

we could use a combo of benches like chrystalmark and some 3 d marks. 03 05 and comebine the scores..

[dr_hardware]

Anything with online verification system, like CPUz , and 3D Mark & PC Mark series.
Aaa, and wPrime as well :)

[flexkill]

Keep them coming guys.....we will get tis going.:thumbs: :rock:

[Snafu]

I seem to recall a contest that used certain scores from crystalmark and other scores from another bench. Could only use valid driver (no beta, homemade or whatever). Links to the scores had to be provided and a panel of judges ruled if the score was legit.

[flexkill]

Also, we need to discuss what classes you would like to see. Stock, superstock, air , water....etc......

[DaveT]

Also, we need to discuss what classes you would like to see. Stock, superstock, air , water....etc......

It would be nice to have air water and extreme cooling categories:chit:
Also would be nice to have single and dual card:yup:

[Snafu]

Good points!

May have to be careful how many categories there are (depending on the benchmark used).

[DaveT]

Good points!

May have to be careful how many categories there are (depending on the benchmark used).

Do you think 3 main categories (those being the cooling) and 2 sub categories (single card or dual) would be too many?

[flexkill]

Do you think 3 main categories (those being the cooling) and 2 sub categories (single card or dual) would be too many?
It's a good start;) Lets see what some other people have to offer as well.

We need and want our members inputs. It's a chance for you to make the things you want happen....and for the crew to make your BE experience the best it can be for you all.

Thanks guys....keep the brain waves flowing;)

[bf2nut]

Do you think 3 main categories (those being the cooling) and 2 sub categories (single card or dual) would be too many?

i agree with dave

++++++++ Cooling++++++++

1.extreme air

2.extreme water

3.extreme phase:D <--makin good money,think im going phase so0n.would love to see this topic blow up.boschuter lets see some killer case shots plz!!phasing cpu,gpus,and chipsets are possible on teh edge!!!

*extreme makes it sound mean!!:rock: :smash: :thumbs: :beat:
++++++++ VGA ++++++++++++

1. single

2. dual(sli\crossfire)

few case shots represtin OCZ on a sticky header on the topic page!!!!!!!!!!

http://i76.photobucket.com/albums/j11/bf2nut/IMG_1460.jpg

ocz platium ram represtin!!!the champion of ddr1 !!!!
my 4 gigs pushing new 8800gts series!!

http://i76.photobucket.com/albums/j11/bf2nut/purty.jpg

pushing 7950gx2 series

http://i76.photobucket.com/albums/j11/bf2nut/bf2nuts7950.jpg

pushing 7900gtx series!! ocz ram pwnz your Gaming!!!!!!!!

http://i76.photobucket.com/albums/j11/bf2nut/wo0t-1.jpg

[flexkill]

Well....we need something for the not so Xtreme members....so we will need a super stock or something along those guidelines.;) :beer:

[Snafu]

bump

[boshuter]

While I'm all in favor of a contest like this :yup: I really don't think you can make it happen, there are just too many different combinations of cpu/gpu/cooling to make it fair to everyone without a huge number of classes. Without a LOT of categories, whoever spends the most money wins; someone with an E6300 and a 7900 gpu can't compete fairly with someone else running an E6800 and 8800GTX... not to mention the people who aren't running Conroes.

I've seen this tried at different places, and it's always a problem trying to find a fair way for anyone to be competitive with a manageable number of classes... I just can't think of any way to pull it off. Even something limited to one bench, like SuperPi, would need a LOT of catagories to make it fair, when you start talking about system and 3D benches it quickly becomes a real mess.

If you can pull it off, I'm willing to be a judge.... I've killed so much hardware in the last few months that I doubt I'll be entering anyway:smash:

[Snafu]

Sorry to hear about your hardware Len :(

Agree that it is tough to figure out the classes. With the old memory contest we had to run amd vs intel but today the choices for categories are fairly complicated.

Then if we have multiple categories then the question becomes how many folks could compete in each class.

SuperPi would be good for cpu/mem - maybe a run what you brung but limit the categories to air/water/other cooling.

Graphics benches are harder but could be limited to single vs multi gpu categories.

There is even the possiblity of multiple benches but only using certain scores from each to come up with a total score.

What would be good is a list of benchmarks and what they measure.


A completely different idea is to make teams to try and level the playing field. The teams then compete amongst themselves and the overall winners can go on to compete against other sites (like a site challenge).

Just trying to get ideas going.

[BGP Spook]

I feel that the days of a single truly objective benchmarking have gone.

Just as the answer to more performance is no longer solely CPU cycles, the answer to benchmarking is a complicated one.

What is likely needed is multiple suits of individual benchmarks.

Then the key would be to choose each benchmark program within a suit to complement the other programs within the same suit and assigning a unique and fair scoring system for each suit.

Each suit would be used to test broad categories.

I am not even sure if there are enough benchmark programs out there to form objective suits.

This would likely take huge amount of time to form and test.

[Snafu]

Any takers interested?

[BGP Spook]

Any takers interested?

Bump.

I think you were headed in the right direction when you said:


What would be good is a list of benchmarks and what they measure.

We should try to post links to where a given benchmark can be downloaded for free and speculation about what it tests.

After we have several candidates we can start evaluating them and forming a suit.

[flexkill]

I will start with these here....if you have mor to add post up and I will add to this list.:thumbs:

Future Mark Downloads Here (http://www.futuremark.com/download/)

Super Pi Mod Here (http://www.xtremesystems.com/pi/)


AquaMark Here (http://downloads.guru3d.com/download.php?det=673)


Cinebench 1.5 Here (http://www.softpedia.com/get/System/Benchmarks/CINEBENCH.shtml)

[BGP Spook]

OpenSourcemark (https://sourceforge.net/projects/opensourcemark/)

HINT (http://en.wikipedia.org/wiki/HINT)*
*I am looking into this one (http://www.ima.umn.edu/~coult/soft.html).

EDIT:
HINT is intended for server or workstation benchmarking in a UNIX environment and requires porting to function in a Windows or Linux environment. It is intended to be edited and compiled by the user in ANSI C to be compatible with their own hardware and operating system.

I know only a little C++(similar in some ways to C) and would not be confident in my ability to make this work. Nor do I think most people on this forum would be able to make heads or tails of it. But please, be my guest to give it a shot and try it out. :smash:

[Snafu]

Some have been using part of crystalmark (sry no link :o)

[flexkill]

Some have been using part of crystalmark (sry no link :o)

CrystalMark (http://www.softpedia.com/get/Tweak/System-Tweak/CrystalMark.shtml);)

[Just Learnin']

CPU-Z (http://www.cpuid.com/cpuz.php) Highest FSB boot into windows.

I understand this 3DMark®2001 SE (http://www.futuremark.com/download/3dmark2001/) to be one of those that can't be cheated at.

[paulxmn]

I would like to see a stock DB and OCDB. An approved template for forum users to list their Bios Settings. Orthos minimum of 8 hours ( the screen shot of orthos should not be IDLE). 3D Mark01 SE 3D Mark 03 and optional between 3D Mark 06 or 05. CPUZ Validated. One thread for Intel and the other for AMD. The data bases should be free of comments and debates. I know this is done at other forums, but it works and makes a great resource for users who want to achieve similar results and a great guide for choosing what hardware you want to buy.

The first image below the Bios Template should be like this.
http://i44.photobucket.com/albums/f18/paulxmn/DFI%20Infinity/Orthos3205.jpg

Bios Settings have to confirmed by any reporting utility similar to Everest.

Best!

[glorfy]

CPU-Z (http://www.cpuid.com/cpuz.php) Highest FSB boot into windows.

I understand this 3DMark®2001 SE (http://www.futuremark.com/download/3dmark2001/) to be one of those that can't be cheated at.

So 03, 05 and 06 can be cheated but 01 can't :scratch:

[Just Learnin']

Sorry Glorfy, seems from further reading it is cheatable too.

[Snafu]

This is why it helps to have judges as well as ensuring folks only use certified drivers :thumbs:

The score gets tossed if:
> it is way out of line with other similar systems (or the person can constantly just beat out the top scorer).
> the drivers are not recognised by the benchmark and are not certified (no betas, or specially tweaked drivers).
> any info is missing from a submitted score

[BGP Spook]

Should we add Everest to the list, even though the newer versions are no longer freely distributed? You can still find Everest 2.2 if you look for it.

There is also a 30 day trial period for the current Everest 3.5.

Anyone know of a replacement instead?

EDIT:
Shamelessly stolen from Pete's "Failing SNM L1 cache test (http://www.bleedinedge.com/forum/showthread.php?p=217589&posted=1#post217589)" thread.
http://www.benchmarkhq.ru/english.html?/be_cpu.html

[paulxmn]

This is why it helps to have judges as well as ensuring folks only use certified drivers :thumbs:

The score gets tossed if:
> it is way out of line with other similar systems (or the person can constantly just beat out the top scorer).
> the drivers are not recognised by the benchmark and are not certified (no betas, or specially tweaked drivers).
> any info is missing from a submitted score

I think Snafu has made three good points. From my reading and site participation experienced moderator's and others have been able to find the flaws in over clocks that have been altered by comparing their benchmarks to the hardware configuration and bios settings. Orthos cannot be submitted in the "IDLE" condition, I have not found out why but that is requirement from others who take their data bases seriously.

All the "cheating" aside, from my forum experience I find BE to be well managed and a polite forum compared to others; I have been told that a data base without comments and follows a strict guideline for submissions has the lowest percentage of altered submissions or none at all. You have to consider that those who "cheat" are usually the most verbal.

Best!

[Snafu]

:lol: good points paulxmn.

One idea is to have the competition for fun. This means no prizes except for the title and accolades from your fellow members. It also provides an opportunity to help each other get better scores (sharing tweaks and tricks).

This will help on two fronts. First it will help our members have better benchmarks that will help the hwbot scores and better prepare ourselves for team competition against other sites.

It will also generate interest amongst the members to participate more and could attract a few other folks (we do have a good rep now and this may picque more interest).

As the contests become smoother and smoother running The Crew can look into possible prizes. The downside with prizes (at least those of good value) is that they can bring out the worst folks on the net who are more willing to try and cheat just to win the prize (as compared to prizes with no real monetary value).

In any case more for the fodder ;)

Keep them all coming :thumbs:

[Snafu]

Another thought to add to this...

In addition to the above thoughts what if the score had to be a legit hwbot score? Does hwbot validate the scores to ensure they are legit or is this limited to the top 10 scores?

[benboy80]

Hey guys nice idea , in response to boshuter's issues on unfairness of cpu's ( more money better chip) , maybe make that into one of the classes to make it fairer ?

[flexkill]

OK guys,

Thanks for all the responses and ideas. The BE crew are going to put our collective heads together and see if we can make this happen. It will only happen if we can find a way to make it fair. We shall see;) :beer:

[Phandinhart]

wa.a.a... very cool

[DaveT]

If you want to have judges I would like to volunteer as one but I think if you decide to use judges they should not be eligable for the competition.

[ol'norton]

While wer'e waiting for this super comp, we could have a quickie simple one, just Super PI 32M scores with the processor on 2700Mhz. It would give a good indication of the difference in real world speed of the board/ram combos. It would be interesting to see if core 2 duo's are any faster than AM2's at the same speed.

[wh1tew0lf]

why not just use percentage of overclocking?

[wh1tew0lf]

that would be fair, wouldn't it?