GMP-ECM for Android

yoyo · 2016-01-14, 18:50

Has anybody yet tried to cross compile a version for Android@ARM?

(Please don't start now a discussion about processing power, battery drain and user interface.)

I never tried to compile gmp-ecm. Is there a link how to compile it and which preconditions I need, especially which other libraries are needed?

My background for the question is that I have thousands of Android clients available which might run some curves.

yoyo

RichD · 2016-01-14, 19:18

Just to give you a quick answer before experts can better address your situation.

First you need the GMP arithmetic library here.
I believe it supports ARM processors but I don’t know anything about this family of CPUs. Once GMP is configured, compiled and installed you can move on to ECM.

You can get ECM here. Similarly you can configure, compile and make this product which links to GMP. The final product is known as GMP-ECM.

Some people use a different math library (MPIR) but I’m not familiar with it.

xilman · 2016-01-14, 20:46

Quote:

Originally Posted by yoyo

Has anybody yet tried to cross compile a version for Android@ARM?

(Please don't start now a discussion about processing power, battery drain and user interface.)

I never tried to compile gmp-ecm. Is there a link how to compile it and which preconditions I need, especially which other libraries are needed?

My background for the question is that I have thousands of Android clients available which might run some curves.

yoyo

The simple answer is that I don't know.

Rather more productively, perhaps you could ask questiosn of the relevant dev teams. First ask on gmp-devel@gmplib.org about the availability or otherwise of GMP on ARM Android (note that Android runs on x86 too). Then ask on ecm-discuss@lists.gforge.inria.fr about GMP-ECM on ARM Android.

Needless to say, some initial research via search engines would be highly advisable.

Paul

VictordeHolland · 2016-01-14, 21:55

I've ran BOINC on my smartphone in the past (HTC One M7 with Snapdragon 600) on Enigma@home. From what I remember it was terribly slow compared to a quad core computer CPU.

Do Smartphones/ ARM devices have enough free RAM for the stage2 of ECM?

VBCurtis · 2016-01-15, 00:13

If the curves are small enough, sure. My Nexus 5 (late 2013 production) has 1GB memory and often 600+ MB free. There's plenty of useful curves that can be run in half a GB.

henryzz · 2016-01-19, 18:35

I have run ecm on a raspberry pi. The big problem when I last tried it was compiling and running programs on an android device. Looks like you need the android NDK. I would be interested if you could get it compiled.

yoyo · 2016-01-19, 20:04

I cross compiled already other stuff with android NDK. My concerns are only which other libs I need and how to convince configure to cross configure it and do not a configure for the host where I compile it.

2016-01-14, 18:50	#1
yoyo Oct 2006 Berlin, Germany 607 Posts	GMP-ECM for Android Has anybody yet tried to cross compile a version for Android@ARM? (Please don't start now a discussion about processing power, battery drain and user interface.) I never tried to compile gmp-ecm. Is there a link how to compile it and which preconditions I need, especially which other libraries are needed? My background for the question is that I have thousands of Android clients available which might run some curves. yoyo

2016-01-19, 18:35	#6
henryzz Just call me Henry "David" Sep 2007 Cambridge (GMT/BST) 5820₁₀ Posts	I have run ecm on a raspberry pi. The big problem when I last tried it was compiling and running programs on an android device. Looks like you need the android NDK. I would be interested if you could get it compiled. Last fiddled with by henryzz on 2016-01-19 at 18:47

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Android app recommendations	Uncwilly	Lounge	11	2020-10-07 12:40
GIMPS client on Android?	user115	Information & Answers	14	2017-06-16 03:55
Mlucas (or Prime95) for Android	yurigomes	Mlucas	13	2016-11-25 11:13
handwriting on Android questions	jasong	jasong	2	2013-02-21 00:04
Android	henryzz	Lounge	7	2011-01-19 18:21

2016-01-14, 19:18	#2
RichD Sep 2008 Kansas 2²×821 Posts	Just to give you a quick answer before experts can better address your situation. First you need the GMP arithmetic library here. I believe it supports ARM processors but I don’t know anything about this family of CPUs. Once GMP is configured, compiled and installed you can move on to ECM. You can get ECM here. Similarly you can configure, compile and make this product which links to GMP. The final product is known as GMP-ECM. Some people use a different math library (MPIR) but I’m not familiar with it.

2016-01-14, 21:55	#4
VictordeHolland "Victor de Hollander" Aug 2011 the Netherlands 1176₁₀ Posts	I've ran BOINC on my smartphone in the past (HTC One M7 with Snapdragon 600) on Enigma@home. From what I remember it was terribly slow compared to a quad core computer CPU. Do Smartphones/ ARM devices have enough free RAM for the stage2 of ECM?

2016-01-15, 00:13	#5
VBCurtis "Curtis" Feb 2005 Riverside, CA 5²·11·17 Posts	If the curves are small enough, sure. My Nexus 5 (late 2013 production) has 1GB memory and often 600+ MB free. There's plenty of useful curves that can be run in half a GB.

2016-01-19, 20:04	#7
yoyo Oct 2006 Berlin, Germany 607 Posts	I cross compiled already other stuff with android NDK. My concerns are only which other libs I need and how to convince configure to cross configure it and do not a configure for the host where I compile it.