New Mersenne Software For Test Mersenne Prime Numbers On Android

[Prime95]

Originally Posted by thorken

this program run in android, in android i dont know a app that run more faster that this. if you know any app that runs faster say me please.

First off, welcome to the forum.

Congratulations on creating a working Lucas-Lehmer test program. I hope you have found the experience enjoyable and informative.

The skepticism here is no reflection on you personally. There are two reasons your announcement has not received the enthusiasm you hoped for. 1) The processor in a phone is not powerful enough to contribute to the search for new Mersenne primes. 2) People here are reluctant to download *any* software from the Internet.

Instead, I suggest posting your code. Members here can suggest improvements or marvel at your ingenuity.

[thorken]

Originally Posted by paulunderwood

Try testing this one which is on the wave front of GIMPS testing: M82589933.

How long will it take? Does it fit in the RAM of your smart phone?

in play store a app called lucas lehmer its for android very very slow. this app its about twenty times more faster.

[science_man_88]

Originally Posted by thorken

in play store a app called lucas lehmer its for android very very slow. this app its about twenty times more faster.

Okay, how does it perform in Big O notation ?

https://rob-bell.net/2009/06/a-begin...ig-o-notation/

[kriesel]

Originally Posted by thorken

the app have been executed in my three moviles without problems its a program to calculate mersenne primes nothing more... and its make it work very good, if dude execute in emulator android in virtual machine.

How many of the criteria does it pass? How can we know, independent of taking the word of a stranger or gambling the health of our personal phones? (I'm not inclined to buy a burner phone to test it out myself, because I know from the power budget of a phone its throughput will not be competitive.) https://www.mersenneforum.org/showpo...98&postcount=4

What are its exponent limits versus phone memory size or other parameters? https://www.mersenneforum.org/showpo...91&postcount=2

[kriesel]

Originally Posted by thorken

Hello ...

Since my grasp of Spanish is almost nonexistent, I keyed in an American-English-keyboard approximation of the first screen shot, ran it through Google Translate, and got:

MvIiIaX - Mersenne Cousins ​​Numbers Calculation - MvIiIaX
The Mersenne Prime Numbers, Are Numbers Cousins ​​Such That ... 2 ^ n-1 is a Prime Number, n Being Cousin. Example: If We Give An Interval From 0 To 10000 When n Be 9973 We Will Obtain A Number Of Some 3003 Digits To Analyze And See If It Is A Prime Number Or No. The Mersenne Prime Numbers Are Represented As: M2, M3, M5, M7, M13, M17, M19 ... M82589933 That's the 51st Mersenne Prime Number Calculated By January 2019, Which Has More Than 24 Million Digits, Much Much Calculation. The purpose of this program is mathematical and Benchmark mode in a given interval, for these calculations it takes much more power than a mobile gives to January 2019, although you will know ... within a few years this program to be surprised ...

At the end of the calculation, the program will give us a rate of volume, this rate of speed of your mobile generating cousins ​​of Mersenne, will have a maximum of 100 points, the higher you are, the more powerful your mobile, for example a S4 Mini Year 2013, Gives An Index Of 0.176 Points, Calculating From 0 To 10000 Of Exponent,? And yours ?. You Can Serve In Way Of BenchMark In Your Mobile Equipment Or To Compete With Your Friends ... To See Who Has The Best Mobile. This Program Does Intensive Use Of The CPU, So The Quicker It Is, The Quicker The Calculations ... The File M8AX - Mersenne.TXT Will Be Created With The Results Of The Calculations In The Folder In Which You Execute The Program.

Enter the Exponent from which the calculations will start.
0

Enter The Exponent In Which The Calculations Will End.
 10000

Anybody know what "MvIiIaX" is?

[pinhodecarlos]

I thought your second language learnt at school was Spanish.

[science_man_88]

Originally Posted by kriesel

Anybody know what "MvIiIaX" is?

I don't know but it seems close to melilax ( appears to be a honey based enema)

[kriesel]

Originally Posted by pinhodecarlos

I thought your second language learnt at school was Spanish.

German. Before PCs were a thing. Just a touch of Spanish for travel purposes (please, thank you, where is the bathroom, etc)

~p^2.45, about a century for M(82859933) from a very long extrapolation made from only 3 data points and limited precision.

What I've seen for other codes range from 2.03 up. 2+ is the minimum, One for iteration count, one for operand size, plus a bit of change for fewer bits/word in the fft transform at larger sizes. Or n * n log n log log n. CUDALucas 2.094, prime95 2.094. https://www.mersenneforum.org/showpo...78&postcount=2
https://www.mersenneforum.org/showpo...23&postcount=2


(Folks in the know, don't take that as talking down to you. It's for those looking up at the learning curve still ahead.)

Instead, I suggest posting your code.

Meaning, source code.

[firejuggler]

no need for 'burner' phone.. there are android emulator out there...( well for windows at least)
Bluestack, Memu,LDplayer, Andy and Nox.

Since I have a dumbphone, if I want to play some mobile game, I have to use an emulator.

I personnally prefer bluestack.

[ewmayer]

Originally Posted by thorken

a mobile phone in 2019 its more powerfull that a pc about 10 years

That may be so, but in that case your program is not making very good use of your hardware, because on my 2009-vintage Core2Duo Macbook classic I can do the LL test of 216091 an order of magnitude faster than the timing you posted.

I also notice that your runtime for 216091 is more than 7x that for 86243, which is somewhat above the O(n^2 log n) scaling one expects from a fast-transform-based large-integer-multiply algorithm. What algorithm are you using for the modular mutiply needed by the LL test?

[ATH]

Originally Posted by thorken

M216091 in 23min 7 sec in s4 mini gt-i9195

The smallest exponent needed for the project is around M47,000,000 for double checking.

If it takes 23 min for M216091 it will take months / years for those exponents if the phone can handle them at all.


Also consider that this project have been running for 23 years! since january 1996. It is very unlikely anyone just joining the project can come up with something ground breaking that no one else thought about in all those years.

A few very intelligent and clever people have made huge changes and contributions to this project over the years, but I do not think any of them had just joined when they did it.

To quote an old man you may or may not know: "Stay awhile and listen..."