Random Number Generators

All of the Torque Game Engine derivatives use the same two algorithm for generating random numbers. You have the code so you can look in engine/math/mRandom.cc/h.

Linear Congruential Method

A Linear Congruential Generator (LCG) is a simple but well known algorithm for random number generation.

"Fast, fairly good random numbers (better than using rand())."
Park & Miller, 1988, Comm of the ACM, 31(10), pp. 1192-1201

R250 Algorithm

R250 is a higher quality RNG algorithm, that is comparable in speed and quality to the 'Mersenne Twister'. Torque's R250 algorithm appears to use the LCG implementation(above) to seed itself.

See also: R250 Algorithm

"Fast, very good random numbers"
Kirkpatrick, S., and E. Stoll, 1981; A Very Fast Shift-Register
Sequence Random Number Generator, Journal of Computational Physics,V. 40.
Maier, W.L., 1991;

A Fast Pseudo Random Number Generator,
Dr. Dobb´s Journal, May, pp. 152 - 157
Period = 2^249

Other Considerations

- "I would say that it would only be a problem if you were relying on a some kind of determinism based upon a seeded value although there may be other scenarios I´ve not thought of." - Melv.

Using the RNGs in Torque

(stub: Inro)

MRandomLCG

(stub: short tutorial and code samples of how to correctly used Torque's built in RNG's - needed here)

MRandomR250

(stub: short tutorial and code samples needed; I started it but could be expanded! ~SP)

Judging from other examples in the code the R250 may be the easier/better RNG to use. It's a two step process:

1. Instantiate an R250 object,
2. Generate a random number,

The MRandom250 is seeded and initlized for you by the constructor. If passed no arguments, it generates a low-quality random seed to initialize itself. You can optionally pass it a seed to your own liking. (Personally, i'm not sure which is preferable ~SP).

MRandomR250 mRandom1;      // Default Constructor called.
U32 x = mRandom1.randI();

U32 seed = Platform::getRealMilliseconds();
MRandomR250 mRandom2(seed);      // Use the seed I give you.

U32 x = mRandom2.randI();