Faking Benchmarks, The Good Benchmark List, How To Spot Cheats and More...

[poke349]

Here's a new multi-threaded Pi-program that I wrote.

Main Thread:
http://www.xtremesystems.org/forums/...d.php?t=221773

It's faster than SuperPi and PiFast. But I'm a little shaky on how to implement anti-cheat protection.

It already has some anti-cheat protection, but I'm no expert in cheating benchmarks so I have no idea how easy it can be cheated. So far, I haven't been able it cheat it myself using all the tactics I know.

Can someone try to cheat this? I'm interested in making it more solid.


For now, I know that the CPU frequency that it reports isn't always accurate and can probably be manipulated. I'm interested to see if there are other (bigger) problems that make it easy to cheat.

[c1rcu1tburn3r]

Anyone know how this was possible???!

Validation Here

[NEOAethyr]

Here's a new multi-threaded Pi-program that I wrote.

Main Thread:
http://www.xtremesystems.org/forums/...d.php?t=221773

It's faster than SuperPi and PiFast. But I'm a little shaky on how to implement anti-cheat protection.

It already has some anti-cheat protection, but I'm no expert in cheating benchmarks so I have no idea how easy it can be cheated. So far, I haven't been able it cheat it myself using all the tactics I know.

Can someone try to cheat this? I'm interested in making it more solid.


For now, I know that the CPU frequency that it reports isn't always accurate and can probably be manipulated. I'm interested to see if there are other (bigger) problems that make it easy to cheat.

I know this is an old thread but I just noticed it today and your post.
Please don't!!!!

I cannot run 32m superpi mod vers (any mod ver including the xs) because of the anti cheat.
It thinks that most of my systems are cheating when I'm only trying to make a run at 32m for stability testing.
I have to run the original superpi.

It might be the diff's between my irq timing and my cpu timing, I don't know, I haven't tried it on an unmodded bios (pic timing mod).
But I really don't care about bench's lol, I just need the stability tests to run .

Anyways, leave the stability option alone in terms of cheat protection, add as much cheat protection as you want to the bench option.

Edit:
However, I supposed could do some messing around with the pic's, pic0 1 and apic timings to see if I can cheat your program later on...
If you wanted me to, not sure if I can get access to the apic multiplier in win7 though, haven't tried yet.
I don't mod those for cheating, I mod them for latency and hd controller timings (via chipsets I can mod the timings directly, other chips I have no idea how to so I mod the master timing).