numpy performance and random numbers

[Rami Chowdhury]

On Dec 20, 2009, at 17:41 , Peter Pearson wrote:

> On Sun, 20 Dec 2009 07:26:03 +1100, Lie Ryan <lie.1296 at gmail.com> wrote:
>> On 12/20/2009 4:02 AM, Carl Johan Rehn wrote:
>> 
>>>>>> Parallel PRNGs are an unsolved problem in computer science.
>>> 
>>> Thanks again for sharing your knowledge. I had no idea. This means
>>> that if I want to speed up my application I have to go for the fastest
>>> random generator and focus on other parts of my code that can be
>>> vectorized.
>> 
>> If you don't care about "repeatability" (which is already extremely 
>> difficult in parallel processing even without random number generators), 
>> you can just start two PRNG at two distinct states (and probably from 
>> two different algorithms) and they will each spews out two independent 
>> streams of random numbers. What was "unsolved" was the "pseudo-" part of 
>> the random number generation, which guarantee perfect replayability in 
>> all conditions.
> 
> Why not use a good cipher, such as AES, to generate a pseudorandom
> bit stream by encrypting successive integers?  

Isn't the Fortuna PRNG based around that approximate concept?

-------------
Rami Chowdhury
"Never assume malice when stupidity will suffice." -- Hanlon's Razor
408-597-7068 (US) / 07875-841-046 (UK) / 0189-245544 (BD)