On May 5, 2018, at 3:23 AM, KHMan wrote:
On 5/5/2018 12:08 PM, Albert Chan wrote:
Reading Vigna latest xoshiro paper (section 11, conclusion),
next version of Lua will use xoshiro256** for math.random.
Is it true ?
Lua 5.4 ?
http://vigna.di.unimi.it/ftp/papers/ScrambledLinear.pdf
IMHO, math.random is similar in purpose to C standard library's
random function. It's pseudo-random, that's about it. It does
not promise any quality specifications.
Are there serious flaws that disqualifies the current
implementation from this purpose?
I just want to check if Vigna's claim is correct.
It would be bad if he just use Lua to promote xoshiro256**
This is from his xoshiro paper, conclusion (section 11)
http://vigna.di.unimi.it/ftp/papers/ScrambledLinear.pdf
The combination of xoroshiro/xoshiro and suitable scramblers
provides a wide range of high- quality and fast solutions for
pseudorandom number generation. Parallax has embedded
xoroshiro128** and xoroshiro32++ in their recently designed
Propeller 2 microcontroller; xoroshiro116+ is the stock generator
of Erlang, and the next version of the popular embedded language
Lua will sport xoshiro256** as stock generator.
