3DMark99?

[Massman]

There is no way to detect what score is "tricked" with CPU speed and what is not. Anyone who can play with affinity/priority and some tools for slow down cpu can boost score by 1%-10%-100%-... with it and you will never be sure unless score will be too big for the clocks/ranks. All that you can - adding rule for all new submissions/screenshots to show CPU speed (system info window -> processor speed or details window -> processor speed).

 

http://i.imgur.com/tCNCy.png

 

I think both 3dmark99 and 3dmark2000 must be completely removed. Or make "wrapper" for it to monitor cpu speed/usage.

 

That's why I think it's not a good idea to hand out points for these benchmarks. So, just for fun ... I don't suppose anyone would cheat in a 'just-for-fun' ranking?

[GENiEBEN]

Cant believe ppl choose to cheat in a fun bench, i knew the trick since P4 HT era.. Anyway, where is the GIVE POINTS, DISSALLOW TRICK option ehh?

[Chiller]

people WILL cheat, even if it's a little bit.

 

just don't allow this benchmark and there will be no problems.

[Massman]

Well, all software is prone to hackers and cheaters. If we follow that logic, no benchmarks will be on HWBOT .

 

99/00 are just benchmarks we added for fun ... I don't think it's too much to ask for people not to use those cpu mhz tricks?

[Christian Ney]

what about post #25 ?

 

 

Or: Disallow this tweak/Trick/or whatever but SS like this will be mandatory:

 

 

Clearly show Detected CPU Speed

Edited April 18, 2011 by Christian Ney

[Johni5]

No bugs .Control panel V sinc OFF and run.

[Massman]

Seems clear: 66% want to keep it and of that crowd, 68% wants disable points and cpu tricking.

 

So, we'll keep the benchmark for nostalgic sentiment, but won't ever give points to it due to it's buggy nature. Just have fun with it!

Edited April 19, 2011 by Massman

[Massman]

Rules updated, now removing the bugged results.

[Christian Ney]

On my side I will remove the one that are with more tha 3.4 GHz

 

Thank you massman for disallow this trick.

 

Now I have a suggestion, can you change the screenshot exemple and add one that is like the one I provided here with the CPU Speed showed and mandatory ? You can use mine SS of course

 

EDIT: Here is an exemple for a 3DMark 2000 SS:

Showing the CPU Speed at the bottom right

 

 

 

3DMark 99:

 

Edited April 19, 2011 by Christian Ney

[Massman]

Should the detected frequency always be equal to stock frequency?

[Christian Ney]

For my i7 2600K without the Trick, the CPU Speed was always between 3.34 and 3.4 GHz

 

Always under stock Freq (around 2 % under) or equal to stock freq

[Freakezoit]

@ Pieter

 

Well i see that from an different point , it is an trick oky that is correct.

I see it equal as other benchers.

 

But the point is if you show in the details that you stay under your Real Freq.

 

Example : Cpu is running @ 5500.00mhz and test shows 5499.99Mhz why should it be Wrong to allow it.

 

I see it like i run that freq. and only the program has an problem to read it Right so there is nothing wrong by helping the programm read the Freq right.

 

The same thing happend @ FM benches with an old FM systeminfo it reads it wrong and the test is a bit slower (i can confirm it in PCM05 and other benches ). But with an actual update scores are higher. Example PCM05 with an old one , Video encoding is slow around 1xxx. with an new one it is around 2xxx or more than 3xxx depending wich cpu and speed is used. Because it uses only maybe 2 cores without , and even on an 980x with it uses all cores.

 

So what will be if all say there it is oky and on the old benches no that is not oky that it knows the real cpu speed. You didn`t get points only cups or medals , thats it .

 

Oky if someone runs his Cpu @ 5500mhz and that test shows more than it is not oky that is the way how i see it. I can also tell you how i found that out with the Cpu freq detection , i was allways wondering that i had my Cpu @ 5.5ghz but 3DM99 and 2000 reads only cpu Freq without turbo wich is 3400mhz if Blck is 100.00 , you could rename the exec but it will not really know what freq your cpu is really running at. But than i had spi running (forgot to close it) and started 3DM2000 and saw that it reads more than normal (around 4500mhz ) Than i tested it how that happend and i found out that the more load on the Cores where 3DM99 & 2000 is started the more freq is read. Oky some call it an cheat other an trick i see it twice if freq is below real it is an trick , if it is higher it is an Cheat.

 

So maybe we should talk about if it is correct to run an bench @ Freq your cpu is running @ or not.

 

I know that alot guys in the Hall of fame (3DM99 & 2000) have not an freq of 3400 or 34 x Blck thay set. That is oky because some guys think there scores are oky but others that show there freq. and had also more are not oky like my WR wich was done with an Cpu Freq of 5.25ghz and Bench @ 5.05ghz . So we need here Clear rules that all show the Freq what the bench reads in Details. Than we can talk about what ist oky and what isn`t.

[Christian Ney]

Allow the trick:

Always under or equal @ the Real CPU Freq

 

Disallow the trick:

Always under or equal @ the Stock CPU Freq

 

 

Everywhere we need the CPU Speed Detection details mandatory on the SS

[tiborrr]

Thank god the tricked results are being removed. This is nothing more than a windows speed trick used ages ago in SuperPi and PiFast.

 

But unfortunately, as Sam said, there will always be people who will try to cheat, even just by 10MHz.

[Chiller]

and that's more the reason to not use this benchmark.

[Christian Ney]

Making screenshots like the ones I provided here will solve your problems