PRP testing on non-x86 CPUs

rogue · 2021-07-02, 19:28

Years ago, Phil Carmody wrote a program called phrot. This program can be used to PRP Sierpinski/Riesel numbers on non-x86 CPUs. My intention is to get this running on one of Apple's new CPUs but I do not have one yet and probably won't get one until next year.

If you have an M1 CPU, download the attached file and follow the readme to build and run the program. Please post updates here.

Uncwilly · 2021-07-02, 19:57

mLucas is getting PRP (with proof file generation and GEC, IIRC). That might be the best choice.

Viliam Furik · 2021-07-02, 20:34

Quote:

Originally Posted by Uncwilly

mLucas is getting PRP (with proof file generation and GEC, IIRC). That might be the best choice.

I don't think Mlucas can handle Sierpinski and Riesel numbers, can it?

Uncwilly · 2021-07-02, 21:09

Quote:

Originally Posted by Viliam Furik

I don't think Mlucas can handle Sierpinski and Riesel numbers, can it?

No mention of it in the readme.
https://www.mersenneforum.org/mayer/README.html

It does have PRP with GEC.

Viliam Furik · 2021-07-02, 23:39

Quote:

Originally Posted by rogue

This program can be used to PRP Sierpinski/Riesel numbers on non-x86 CPUs.

OP was talking about a program for Riesel and Sierpinski numbers. Because of this, Mlucas, even though having PRP with GEC, comes out of play.

axn · 2021-07-03, 03:43

IIRC, phrot was designed for non-base-2 numbers (to support SR5), and was not recommended on base-2 on x86 because it was much slower than LLR.

IIRC, there is a plain C LLR, that could be compiled on any CPU.

Ergo, you should compare phrot with plain C LLR (regardless of the CPU), if you're testing base 2 numbers.

rogue · 2021-07-03, 13:32

Quote:

Originally Posted by axn

IIRC, phrot was designed for non-base-2 numbers (to support SR5), and was not recommended on base-2 on x86 because it was much slower than LLR.

IIRC, there is a plain C LLR, that could be compiled on any CPU.

Ergo, you should compare phrot with plain C LLR (regardless of the CPU), if you're testing base 2 numbers.

If there is a "plain C LLR", can you provide link?

axn · 2021-07-03, 13:48

Found a really old thread - https://www.mersenneforum.org/showthread.php?t=15584

Not sure it is the right one. You could PM Jean and check.

2021-07-02, 19:57	#2
Uncwilly 6809 > 6502 """"""""""""""""""" Aug 2003 101×103 Posts 10101100111011₂ Posts	mLucas is getting PRP (with proof file generation and GEC, IIRC). That might be the best choice. Last fiddled with by ewmayer on 2021-07-02 at 20:34 Reason: Not for Sierpinski/Riesels

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Status of Wagstaff testing? and testing Mersenne primes for Wagstaff-ness	GP2	Wagstaff PRP Search	414	2020-12-27 08:11
Anti-poverty drug testing vs "high" tax deduction testing	kladner	Soap Box	3	2016-10-14 18:43
CPUs assigned	houding	Hardware	10	2015-03-25 00:33
A tale of 3 CPUs	chris2be8	Hardware	7	2010-07-20 23:12
Speed of P-1 testing vs. Trial Factoring testing	eepiccolo	Math	6	2006-03-28 20:53

2021-07-03, 03:43	#6
axn Jun 2003 2³·683 Posts	IIRC, phrot was designed for non-base-2 numbers (to support SR5), and was not recommended on base-2 on x86 because it was much slower than LLR. IIRC, there is a plain C LLR, that could be compiled on any CPU. Ergo, you should compare phrot with plain C LLR (regardless of the CPU), if you're testing base 2 numbers.

2021-07-03, 13:48	#8
axn Jun 2003 12530₈ Posts	Found a really old thread - https://www.mersenneforum.org/showthread.php?t=15584 Not sure it is the right one. You could PM Jean and check.