PCG, A Family of Better Random Number Generators
pcg-random.org
PCG, A Family of Better Random Number Generators
1–10 of 57 posts
Re: PCG, A Family of Better Random Number Generators
#2I'm impressed by how fast this algorithm is. I wrote a toy implementation in Go, and it was only ~50ns/op to generate random numbers. And it's very even in distribution, but I'd have to rerun my tests to find the percentage again.
Re: PCG, A Family of Better Random Number Generators
#3Seems perfectly reasonable and worth knowing about. Thanks for submitting.
Re: PCG, A Family of Better Random Number Generators
#4Sounds awesome, if it lives up to its claims. Have the domain experts weighed in on this yet?
Re: PCG, A Family of Better Random Number Generators
#5There is a seminar about PCG from its author available on youtube.
https://www.youtube.com/watch?v=45Oet5qjlms
Its a bit long but I highly recommend it. I learned a lot about RNGs from it.
Re: PCG, A Family of Better Random Number Generators
#6Unless it's secure, I see no motivation to use it.
In their table, they acknowledge that Arc4Random and ChaCha20 as secure. The only negatives against ChaCha20 is that it's not 'fast enough' and k-Dimensional Equidistribution. In my experience it's never showed up in my profiling, so I feel it's fast enough; and I'm not sure why I want k-Dimensional Equidistribution.
So... I'll keeping using ChaCha20 when I need to (it's part of nacl/libsodium if you don't know where to get it from).
Re: PCG, A Family of Better Random Number Generators
#7This is fine for quickly generating random numbers for simulations, just don't use it for cryptographic applications. If you want random numbers for crypto use a CSPRNG.
Re: PCG, A Family of Better Random Number Generators
#8Sounds awesome, if it lives up to its claims. Have the domain experts weighed in on this yet?
The latest word is,
"This paper is currently submitted to ACM Transactions on Mathematical Software, where it is currently under review. "
Re: PCG, A Family of Better Random Number Generators
#9Unless it's secure, I see no motivation to use it. In their table, they acknowledge that Arc4Random and ChaCha20 as secure. The only negatives against ChaCha20 is that it's not 'fast enough' and k-Dimensional Equidistribution. In my experience it's never showed up in my profiling, so I feel it's fast enough; and I'm not sure why I want k-Dimensional Equidistribution. So... I'll keeping using ChaCha20 when I need to (…
I wrote a standalone C++11 (compatible with the standard RNG library) implementation of ChaCha (with the number of rounds as a parameter), available here: https://gist.github.com/orlp/32f5d1b631ab092608b1
This is also the implementation used for the benchmark at the pcg-random.org homepage.
Re: PCG, A Family of Better Random Number Generators
#10Unless it's secure, I see no motivation to use it. In their table, they acknowledge that Arc4Random and ChaCha20 as secure. The only negatives against ChaCha20 is that it's not 'fast enough' and k-Dimensional Equidistribution. In my experience it's never showed up in my profiling, so I feel it's fast enough; and I'm not sure why I want k-Dimensional Equidistribution. So... I'll keeping using ChaCha20 when I need to (…
I wrote a standalone C++11 (compatible with the standard RNG library) implementation of ChaCha (with the number of rounds as a parameter), available here: https://gist.github.com/orlp/32f5d1b631ab092608b1 This is also the implementation used for the benchmark at the pcg-random.org homepage.
How does it compare to the libsodium implementation: https://github.com/jedisct1/libsodium/blob/master/src/libsod...
Btw, libsodium prefers salsa20 instead of chacha20, I don't know why; but I trust they made a good decision.