I have an exe that calls a PRNG but I don't know the name of this algorithm, could someone take a look at the asm dump of the exe and tell me if it rings a bell?, I know there are many advanced assembly programmers in here that could spot this sort of thing but I don't know how feasible it is since it might take some time to do so and time isn't cheap

But just in case, this is the dump:
http://pastebin.com/f2a720984

To get you started, the decimal number 1000 is passed to the routine, so it all begins under SUB_L00402079.

From my limited experience and point of view, it doesn't look like a mersenne twister to me, but it seems to outperform many other PRNGs in both speed and output.

The reason I want to know what algorithm this is, is simply to read more about it, try to implement my own version of it, see how it goes.

Could it be an extended congruential algorithm?

Any clues onto what algo this is will be greatly appreciated

Thanks.

Might be a Park-Miller variant?
Nah, looks like two LCG's overlapped?

It could just be a custom RNG, not too uncommon

It's uniform spread is not that of a custom half-assed PRNG at all, it's also quite fast!.

That said, it's not very simple to make an efficient PRNG at all, you may get random numbers quite easily by using many known and simple methods, but the generated data at the end could not be uniform enough and/or the algorithm might not be fast enough for most real-world applications.

So I'm still wondering if anyone has a clue. On an IRC channel someone pointed e that this is a variant of the lagged-Fibonacci algorithm, but I'm not so sure about this.

Why not try asking the person who wrote the program you snipped this from?

Isn't lagged-Fibonaci slow? (not that I've actually tested or anything...)