gmp-4.3.0 is out

[Batalov]

http://gmplib.org/#DOWNLOAD

mpir-1.1 is also out
http://mpir.org/

Choices, choices!

[Yamato]

Nice. Does Gaudry's patch work with this?

[ET_]

I found a bug

While the macros

— Macro: __GNU_MP_VERSION
— Macro: __GNU_MP_VERSION_MINOR
— Macro: __GNU_MP_VERSION_PATCHLEVEL

work as should, the Global Constant

— Global Constant: const char * const gmp_version

still refers to 4.2.2

Luigi

[jasonp]

Are you sure you don't have another version of GMP somewhere on your system? You can be using the new header files but an older precompiled library may be in your path...

(this has bitten me too, and is why GMP-ECM's configure script looks for conflicts)

[CRGreathouse]

Has anyone done a speed comparison?

[jrk]

This is a comparison between MPIR 0.9.0 and MPIR 1.1.0.

0.9.0:

$ time ( echo "4917400029751399370968199585848845965968509843013458776241467726948301116079944821945421874145871641363912960613047045854615665784286241"|gmp-ecm -c 1 1e6)
GMP-ECM 6.2.1 [powered by GMP 4.2.1] [ECM]
Input number is 4917400029751399370968199585848845965968509843013458776241467726948301116079944821945421874145871641363912960613047045854615665784286241 (136 digits)
Using B1=1000000, B2=1045563762, polynomial Dickson(6), sigma=2899847002
Step 1 took 5063ms
Step 2 took 2266ms

real    0m11.978s
user    0m7.331s
sys    0m0.016s

1.1.0:

$ time ( echo "4917400029751399370968199585848845965968509843013458776241467726948301116079944821945421874145871641363912960613047045854615665784286241"|gmp-ecm -c 1 1e6)
GMP-ECM 6.2.1 [powered by GMP 4.2.1] [ECM]
Input number is 4917400029751399370968199585848845965968509843013458776241467726948301116079944821945421874145871641363912960613047045854615665784286241 (136 digits)
Using B1=1000000, B2=1045563762, polynomial Dickson(6), sigma=1199160155
Step 1 took 4205ms
Step 2 took 2155ms

real    0m9.971s
user    0m6.361s
sys    0m0.031s

There is about a 15% speedup, judging by user time.

[CRGreathouse]

I was more interested in GMP vs MPIR, but that's good to know. 'Word on the street' was that the last versions of GMP and MPIR were about the same speed; will that still hold, or will GMP pull ahead?

[ET_]

Originally Posted by jasonp

Are you sure you don't have another version of GMP somewhere on your system? You can be using the new header files but an older precompiled library may be in your path...

(this has bitten me too, and is why GMP-ECM's configure script looks for conflicts)

I am sure. I reformatted my laptop last week and had no previous installation of GMP (in fact I was waiting for 4.3.0 to install)

Luigi

[akruppa]

Luigi,

I can't reproduce this. The file version.c in GMP refers to the macro VERSION which is defined as
#define VERSION "4.3.0"
in config.h. The definitions of __GNU_MP_VERSION, __GNU_MP_VERSION_MINOR, and __GNU_MP_VERSION_PATCHLEVEL in gmp.h also have 4, 3, and 0, repectively.

A test program

#include <stdio.h>
#include <gmp.h>

int main ()
{
  printf ("From header: version %d.%d.%d\n",
          __GNU_MP_VERSION, __GNU_MP_VERSION_MINOR,
          __GNU_MP_VERSION_PATCHLEVEL);
  printf ("From library string: %s\n", gmp_version);
  return 0;
}

prints

From header: version 4.3.0
From library string: 4.3.0

Can you compile your (or my) test program with the -H flag for gcc? It'll show the full path of files that are included by the preprocessor, so you may be able to find a well-hidden (but not well enough for cpp!) stale header file. Similarly, giving the option -Wl,-t to gcc (so that gcc passes the option -t to ld) causes linking to show the full path of libraries that are used.

Alex

[ET_]

Strange enough, after having turned off and on again the laptop, everything seems to work fine.

Thanks anyway for the help.

Alex, I wrote the same sample program to test the version number before posting on the forum, and the results were different. Oh, well...

Luigi

[M0CZY]

I have been trying to compile GMP-ECM using the new GMP-4.3.0 under Linux, but I have hit a snag.

The GMP installs correctly, and passes 'make check', but in the ECM stage, when I get to './configure --with-gmp=/usr/local', it outputs some stuff then finishes with

checking if gmp.h version and libgmp version are the same... (4.3/4.3.0) no
configure: error: 'gmp.h' and 'libgmp' have different versions, you have to reinstall GMP properly.
user@user-desktop:~/ecm-6.2.2$

I received the same difference in versions when I compiled under Windows using MinGW/Msys, but the program ignored it and continued working.

Could this be a bug in the new version of GMP?