mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Conjectures 'R Us (https://www.mersenneforum.org/forumdisplay.php?f=81)
-   -   proth 2.0 use (https://www.mersenneforum.org/showthread.php?t=25935)

masser 2020-09-07 15:47

proth 2.0 use
 
Anyone from CRUS using proth 2.0?

There was a brief discussion [URL="https://mersenneforum.org/showthread.php?t=25130"]here[/URL]. I'd like to know how well it works for our project's purposes.

pepi37 2020-09-07 17:15

[QUOTE=masser;556333]Anyone from CRUS using proth 2.0?

There was a brief discussion [URL="https://mersenneforum.org/showthread.php?t=25130"]here[/URL]. I'd like to know how well it works for our project's purposes.[/QUOTE]

Wotk ok and well. Mote info after I back from work

rogue 2020-09-07 21:29

proth20 can only be used for k*2^n+1. In other words b = 2 and c = +1. Hopefully Yves will extend that to other bases and c = -1 in the near future.

My experience is that it is slower than pfgw, but I haven't run it with many different values of n to see how it scales as n increases.

masser 2020-09-07 21:45

[QUOTE=rogue;556368]proth20 can only be used for k*2^n+1. In other words b = 2 and c = +1. Hopefully Yves will extend that to other bases and c = -1 in the near future.

My experience is that it is slower than pfgw, but I haven't run it with many different values of n to see how it scales as n increases.[/QUOTE]

Do both pfgw and proth20 run on Nvidia GPUs?

pepi37 2020-09-08 09:45

[QUOTE=masser;556369]Do both pfgw and proth20 run on Nvidia GPUs?[/QUOTE]
It has no sense to compare PFGW and Proth: one is run under GPU and another is run under CPU.
Proth 2.0 has very nice performance

Table of limits ( when program go to next step)


Red is exponent on k*2^[COLOR=red]n[COLOR=Black]+1[/COLOR][/COLOR]

2^17 [ [COLOR=Red]688,094 - 1,376,221[/COLOR]]
2^18 [ [COLOR=red]1,376,222 - 2,752,477[/COLOR]] PPSE (1.5M)
2^19 [ [COLOR=red]2,752,478 - 5,504,989[/COLOR]] PPS (2.8M)
2^20 [ [COLOR=red]5,504,990 - 10,485,716[/COLOR]] DIV (5.5M)
2^21 [[COLOR=red]10,485,717 - 20,971,486[/COLOR]] ESP (14.1M), 321 (15.5M), Cullen (17.7M)
2^22 [[COLOR=red]20,971,487 - 39,845,845[/COLOR]] PSP (22.1M), SOB (32.5M)
2^23 [[COLOR=red]39,845,846 - 79,691,743[/COLOR]]
2^24 [COLOR=red][79,691,744 - 99,999,999[/COLOR]]

henryzz 2020-09-08 20:31

[QUOTE=pepi37;556403]It has no sense to compare PFGW and Proth: one is run under GPU and another is run under CPU.
Proth 2.0 has very nice performance

Table of limits ( when program go to next step)


Red is exponent on k*2^[COLOR=red]n[COLOR=Black]+1[/COLOR][/COLOR]

2^17 [ [COLOR=Red]688,094 - 1,376,221[/COLOR]]
2^18 [ [COLOR=red]1,376,222 - 2,752,477[/COLOR]] PPSE (1.5M)
2^19 [ [COLOR=red]2,752,478 - 5,504,989[/COLOR]] PPS (2.8M)
2^20 [ [COLOR=red]5,504,990 - 10,485,716[/COLOR]] DIV (5.5M)
2^21 [[COLOR=red]10,485,717 - 20,971,486[/COLOR]] ESP (14.1M), 321 (15.5M), Cullen (17.7M)
2^22 [[COLOR=red]20,971,487 - 39,845,845[/COLOR]] PSP (22.1M), SOB (32.5M)
2^23 [[COLOR=red]39,845,846 - 79,691,743[/COLOR]]
2^24 [COLOR=red][79,691,744 - 99,999,999[/COLOR]][/QUOTE]

How does the performance compare per watt on typical hardware?


All times are UTC. The time now is 08:28.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.