Onepagebook in Cheating Scandal?

Jokester

Jokester

Jokester

Don
Joined
7 Aug 2003
Posts
44,880
Location
Aberdeenshire
http://www.xtremesystems.org/forums/showthread.php?t=126560

A lot of people in this forum will be aware of OPB, especially from his RAM overclocking, but it would seem things don't appear to be as they seem. Reserving judgement at the moment, but I feel there's definitely a case to answer as his result is a good 4-5% more efficient than every other overclocker out there.

Coolaler might not be looking to clever either.

Will be watching to see how it develops.

Jokester
 
Interesting.

I'm surprised that ye olde 'hand picked hardware' subject hasn't been put forward for arguement. Me for one believes that such "top overclockers" have access to the best equipment and as such we the public should only believe the law of averages. But...I have had high overclocks on "poor" equipment so why shouldn't these guys be ahead with inside knowledge and or tweaks?
 
weescott said:
But...I have had high overclocks on "poor" equipment so why shouldn't these guys be ahead with inside knowledge and or tweaks?
Click to expand...

The suggestion though is that their score is actually faster the even the best hardware/tweaks could possibly achieve at those clock rates. If it's a tweak, it gives them the equivalent boost of a couple of hundred MHz extra on the CPU which would be unheard of.

Jokester
 
Good point, but could the 'hand picked' debate contribute to this? Or even insider knowledge? Either or may be ligitimit but either way unfair. Dare I say...are they paid by the industry? :eek: I won't sit on the fence! Cheat Cheat Cheat Cheat Cheat Cheat :p
 
Some of these guys do get hand picked components...

its also possible to find some mad tweaks several months down the line that you overlooked before... i.e. with the RAM on my P4 I just happened one day to accidentally select one of the reserved values instead of the value that research and logic would have suggested was the most optimal, and suddenly I was seeing a 500MB/s increase in RAM read speeds...

also the highest settings aren't always the best for performance... i.e. again my old P4 it would overclock to 4.7gig stable with high end cooling - but something was causing it to throttle back to almost stock speeds - 4.05gig was giving much higher benchmarks.

I'm not an amazing overclocker, too impatient to read and study it, but I have a certain amount of natural ability and instinct, I will often squeeze an extra 5% out of what my friends are capable of without seeming to do much difference settings wise, etc. tho I do seem to always get the luck of the draw with component quality.
 
Last edited:
Rroff said:
hand picked components...
Click to expand...

Hand picked components might let you reach higher clock speeds but they won't let you run quicker at the same speed. Everyone else is hitting the same speeds as them but are miles off the pace.

As for tweaks, it must be some tweak to give them that level of boost.

Jokester
 
Jokester said:
Hand picked components might let you reach higher clock speeds but they won't let you run quicker at the same speed. Everyone else is hitting the same speeds as them but are miles off the pace.

As for tweaks, it must be some tweak to give them that level of boost.

Jokester
Click to expand...

true
 
From reading they're not being accused of cheating. It's just that their results don't match "the norm". You'd also assume the "score" (multiply speed by time) would be relatively consistent if they were using the same or at least similar tweaks. The "scores" seem to vary quite a bit from the "average" or "ideal" scores which i think is the issue being raised.
 
yah... I don't think using time x cpu frequency is a good indicator tho - as I've noted upto 0.5second differences on the same frequency when using different RAM, etc. settings.

Also I've seen a spread between 15 and 20 seconds for 1Mb SPI on a E6600 @ 3gig, from different people, so you can't really go just on frequency.
 
excess heat and voltage can cause CPUs to throttle, etc. so that lower frequencies can provide higher scores - however they are being caught out here because there are other times at those frequencies which don't agree with what they are getting... ofcourse its possible that those scores were done with less than optimal heat or voltage, etc. etc. etc.

its also possible for different revisions of the same product to perform very differently...
 
Rroff said:
Also I've seen a spread between 15 and 20 seconds for 1Mb SPI on a E6600 @ 3gig, from different people, so you can't really go just on frequency.
Click to expand...
On different chipsets with different strap latencies, then 5 seconds easily. There's a long list of variables which may not be immediately apparent. I'm sure those guys have tweaks they're not sharing, thats why they have world records.
 
Rroff said:
excess heat and voltage can cause CPUs to throttle, etc. so that lower frequencies can provide higher scores - however they are being caught out here because there are other times at those frequencies which don't agree with what they are getting... ofcourse its possible that those scores were done with less than optimal heat or voltage, etc. etc. etc.

its also possible for different revisions of the same product to perform very differently...
Click to expand...


Erm, no, each product it 99.99999% the same, performance is 99.999999999% the same. Even the smallest error in the chip and it won't work/work as well.

As far as I can see, software mods can't do anything that drastic.

Must be timings or hardware that does it...
 
don't think I'd answer people like that either if I was either of them tbh... I find it interesting that raju is sticking up for them, as I've noticed from other forums he doesn't tend to stick his nose out without something solid to back it up...
 
Concorde Rules said:
Erm, no, each product it 99.99999% the same, performance is 99.999999999% the same. Even the smallest error in the chip and it won't work/work as well.

As far as I can see, software mods can't do anything that drastic.

Must be timings or hardware that does it...
Click to expand...


uh no... I'm not talking about 2 different products from the same batch...
 
does bring up one question tho... the validity of freak results... I've seen it not uncommonly with overclocking that one run will out of the blue return a very high score that is impossible to replicate again, and totally outside of any common result curve... can you call that a world record if you can't replicate it?
 
Rroff said:
does bring up one question tho... the validity of freak results... I've seen it not uncommonly with overclocking that one run will out of the blue return a very high score that is impossible to replicate again, and totally outside of any common result curve... can you call that a world record if you can't replicate it?
Click to expand...
If its well off the curve? No. If you can't repeat it within a percent or two, then I'd say its an aberration and not valid.
 
You must log in or register to reply here.
Back
Top Bottom