Mlucas version 17.1

[ewmayer]

Originally Posted by Xyzzy

All Ernst needs to do is (consistently) use relative links instead of absolute links.

http://www.mersenneforum.org/mayer/

Thanks, Mike - I will do the needful (RIP Mally) when next I get round to updating my readme.html file.

[ET_]

I successfully installed Gentoo 64bits on my Raspberry PI 3.

The system works like a charm with my code, and does not like threads on PERL

This Perl not built to support threads
Compilation failed in require at Pgap_seg.pl line 8.
BEGIN failed--compilation aborted at Pgap_seg.pl line 8.

I downloaded the 3.4 MB tar.gz on Ernst page (the code worked fine on Raspbian 32bit) and ./configured it.

But when I try to make, I get te following:

demouser@pi64 ~/luigi/mlucas-14.1 $ make
make  all-am
make[1]: Entering directory '/home/demouser/luigi/mlucas-14.1'
  CC       $NORMAL_O $THREADS_O
make[1]: *** [Makefile:2986: NORMAL_O-THREADS_O.stamp] Error 1
make[1]: Leaving directory '/home/demouser/luigi/mlucas-14.1'
make: *** [Makefile:2084: all] Error 2

uname -a gives the following:

Linux pi64 4.10.0-rc5-v8 #1 SMP PREEMPT Wed Jan 25 20:13:50 GMT 2017 aarch64 GNU/Linux

The config files are attached.

Luigi

[Mark Rose]

Originally Posted by ET_

I successfully installed Gentoo 64bits on my Raspberry PI 3.

How many months of compiling did that take?

[ET_]

Originally Posted by Mark Rose

How many months of compiling did that take?

Not much. The 64bit OS makes the PI run about 4-5 times slower than a normal PC. One core of a PC nearly equals 4 PI cores.
Consider that I have a Gentoo image installed in a matter of minutes on the micrcoSD.

[ewmayer]

I see several errors in config.log:

configure:3324: gcc -V >&5
gcc: error: unrecognized command line option '-V'
gcc: fatal error: no input files
compilation terminated.
configure:3324: gcc -qversion >&5
gcc: error: unrecognized command line option '-qversion'
gcc: fatal error: no input files
compilation terminated.

Not sure what those are about ... I need to ping config-script author Alex Vong what input the '>&5' are intended to feed to the above 2 gcc flags. Nonetheless the config script continues, here is the next batch of errors:

configure:4053: checking for library containing ceil, log, pow, sqrt, sincos, floor, lrint, atan
configure:4084: gcc -o conftest conftest.c >&5
conftest.c:18:6: warning: built-in function 'ceil' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:12: warning: built-in function 'log' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:17: warning: built-in function 'pow' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:22: warning: built-in function 'sqrt' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:28: warning: built-in function 'sincos' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:36: warning: built-in function 'floor' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:43: warning: built-in function 'lrint' declared as non-function
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
conftest.c:18:50: warning: conflicting types for built-in function 'atan'
char ceil, log, pow, sqrt, sincos, floor, lrint, atan ();
^
/tmp/ccSv91Yj.o: In function `main':
conftest.c:(.text+0x8): undefined reference to `atan'
collect2: error: ld returned 1 exit status

That looks like something related to the math libs. Again, I'll ask Alex to have a look. At some point in the not-too-distant future I need to take ownership (or at least curatorship) of the config/make system, but now is not a good time for that effort.

[ldesnogu]

Ernst, if you run configure on your x86 machine you'll get the same errors in your config.log file ;)

[ET_]

Originally Posted by ewmayer

I see several errors in config.log:
[...]
That looks like something related to the math libs. Again, I'll ask Alex to have a look. At some point in the not-too-distant future I need to take ownership (or at least curatorship) of the config/make system, but now is not a good time for that effort.

I will be in New York and Boston for vacation and come back after May 7th.

My PI system is actually sieving for the DoubleMersennes project, and the Pico-Cluster is still in its box.

Take your time...

Luigi

[ewmayer]

Originally Posted by ldesnogu

Ernst, if you run configure on your x86 machine you'll get the same errors in your config.log file ;)

Indeed - before posting them to the ftp site I used the install scripts to install on my Intel Broadwell NUC, but since that install was successful I never looked at the resulting config.log until now. As you say, same errors, all the way down [there are around a dozen in total.]

[ldesnogu]

My understanding is that it's the normal process: configure scripts test various ways to do things until they find one that works. So I guess there's nothing wrong to be found in these logs.

[ewmayer]

Originally Posted by ldesnogu

My understanding is that it's the normal process: configure scripts test various ways to do things until they find one that works. So I guess there's nothing wrong to be found in these logs.

Yes, that makes sense.

Luigi, could you try going directly into the src directory of the install tree and trying a manual build as per the README page?

[ET_]

Originally Posted by ewmayer

Yes, that makes sense.

Luigi, could you try going directly into the src directory of the install tree and trying a manual build as per the README page?

Sure, as Soon as I get Back home next week.