![]() |
![]() |
#1 |
"Ed Hall"
Dec 2009
Adirondack Mtns
148A16 Posts |
![]()
I'm starting a new thread to address these, rather than clutter existing threads. I haven't contacted the author(s) yet, because I'm not familiar enough with where my issues are and the solutions may well be simple.
Basically, I installed most of the dependencies via the repositories for Ubuntu 20.04, leaving only mpfrcx and cm to install manually. This worked great for the first machine I tried, and almost* great for the second. Both compiled and passed all the checks and are working without major issues. I've tried three machines since then and all three fail the cm make check step due to the inability to load the libmpfrcx.so.1 library. But, the ./configure for cm had no trouble finding it in its check and it is installed to the exact same place on all five machines. All five machines are Ubuntu 20.04 and all meet the dependencies for cm. * Oddly, although running fine, this machine does not produce .primo files, while the first one does. All commands (and options) are identical. |
![]() |
![]() |
![]() |
#2 |
"Oliver"
Sep 2017
Porta Westfalica, DE
2·23·29 Posts |
![]()
IIRC, it has to be compiled manually beforehand. See e.g. http://gitlab.inria.fr/enge/mpfrcx.
|
![]() |
![]() |
![]() |
#3 |
"Ed Hall"
Dec 2009
Adirondack Mtns
2×11×239 Posts |
![]()
I did compile and install mpfrcx before compiling and installing cm, if that's what you mean.
The steps I'm trying to use (hopefully to supply in yet another (How I . . .): (GMP-6.2.0 and PARI/GP already installed) sudo apt install libmpc-dev libmpfr-dev libpari-dev acquire the tars for mpfrcx and cm and unpack ./configure for mpfrcx make make check sudo make install ./configure for cm make make check - failure You have reminded me of something, though. On the first machine, I may have installed more than the minimum (-dev) for one or more package. Maybe that's my issue, although I would expect the ./configure to be accurate in its checking, and I believe the minimum was what I used for the second machine. |
![]() |
![]() |
![]() |
#4 |
"Ed Hall"
Dec 2009
Adirondack Mtns
2·11·239 Posts |
![]()
Interesting partial progress: I installed mpc as an addition to the other mentioned packages and one of the non-working machines passed the check and runs ecpp, but there are still no .primo certificates.
I tried a totally separate machine with all the aforementioned packages, such that it should match the previous configuration and it fails the check. More study is necessary. . . |
![]() |
![]() |
![]() |
#5 |
Bamboozled!
"๐บ๐๐ท๐ท๐ญ"
May 2003
Down not across
266168 Posts |
![]()
Trying desperately to remember what I had to do to get it running on my Ubuntu sytstems...
I wonder if it was this LD_LIBRARY_PATH=/usr/local/lib though that may have been some other package such as Algol 68 Genie. |
![]() |
![]() |
![]() |
#6 |
"Ed Hall"
Dec 2009
Adirondack Mtns
2·11·239 Posts |
![]()
I know I had to work with PATHs to get my GPU packages to work right and I explored the PATHs for these machines for ecpp.
I've been partially successful by installing mpc (rather than just libmpc-dev) on some of the machines via the repository, but only a couple started working then and none of the machines, other than the very first, will generate .primo certs. My next step is going to be uninstalling all the repo versions and manually installing the available packages from multiprecision.org to see what that may accomplish. |
![]() |
![]() |
![]() |
#7 | |
Bamboozled!
"๐บ๐๐ท๐ท๐ญ"
May 2003
Down not across
2D8E16 Posts |
![]() Quote:
pcl@thoth:~/Pictures$ ls /usr/local/lib libcm.a libcm_mpi.a libcm_mpi.so libcm_mpi.so.0.0.0 libcm.so.0 libmpfrcx.a libmpfrcx.so libmpfrcx.so.1.2.3 python3.10 R libcm.la libcm_mpi.la libcm_mpi.so.0 libcm.so libcm.so.0.0.0 libmpfrcx.la libmpfrcx.so.1 pypy2.7 python3.7 x86_64-linux-gnu pcl@thoth:~/Pictures$ |
|
![]() |
![]() |
![]() |
#8 |
"Ed Hall"
Dec 2009
Adirondack Mtns
2·11·239 Posts |
![]()
Thanks!
I "think" I have found the problem, but all is not well and my attention span is diminishing. What I HAVE discovered is that while libmpc-dev is of sufficient version in the repository, mpc is NOT. Neither is libmpfr-dev. To manually install mpc requires me to manually install mpfr, which is OK, but mpc's make check fails: Code:
Error, mpc_mul returns an out-of-range exponent: 0.10010110101000101010000000001001011011000011101100000E-1085!!!<!!! Bug most probably in MPFR, please upgrade to MPFR 4.1.0 or later FAIL tmul (exit status: 1) The most annoying thing is that the first machine worked fine with all the repository versions, but each of the other machines fail in various manners. Later today I will try this again with a yet un-fiddled with machine in my set and see what happens with it. Perhaps these others have a bit of left over code somewhere. |
![]() |
![]() |
![]() |
#9 |
"Ed Hall"
Dec 2009
Adirondack Mtns
2×11×239 Posts |
![]() |
![]() |
![]() |
![]() |
#10 | |
If I May
"Chris Halsall"
Sep 2002
Barbados
11,087 Posts |
![]() Quote:
Have you tried attacking this problem space using a different OS substrate? Personally, I have given up on Ubuntu. In all of my spaces. For LTS Linux contexts, I'm currently deploying either on pure Debian (when I *have to*), or Rocky Linux when I'm allowed to future-proof. |
|
![]() |
![]() |
![]() |
#11 |
Sep 2002
Database er0rr
5×29×31 Posts |
![]()
@EdH, did you get mpc from https://www.multiprecision.org/mpc/download.html ?
I installed the headers only through apt Code:
apt list mpc Listing... Done mpc/kali-rolling 0.34-1+b1 amd64 Code:
apt list libmpc-dev Listing... Done libmpc-dev/kali-rolling,now 1.2.1-2 amd64 [installed] Code:
ls /usr/local/lib libcm.a libcm_mpi.la libcm_mpi.so.0.0.0 libcm.so.0.0.0 libgmp.a libgmp.so.10 libmpfrcx.a libmpfrcx.so.1 libprimesieve.a pkgconfig libcm.la libcm_mpi.so libcm.so libecm.a libgmp.la libgmp.so.10.4.1 libmpfrcx.la libmpfrcx.so.1.2.3 libysieve.a python2.7 libcm_mpi.a libcm_mpi.so.0 libcm.so.0 libecm.la libgmp.so libgwnum.a libmpfrcx.so libmsieve.a libytools.a python3.10 Last fiddled with by paulunderwood on 2022-11-30 at 06:09 |
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Gpuowl Windows builds | kriesel | GpuOwl | 34 | 2022-12-10 19:18 |
Would Personal Builds run win-11? | a1call | Hardware | 29 | 2021-11-27 14:42 |
Apparent Ubuntu Version Troubles with ecmpi | EdH | GMP-ECM | 1 | 2020-05-04 13:16 |
ARM builds and SIMD-assembler prospects | ewmayer | Mlucas | 183 | 2019-02-25 08:17 |
Running 32-bit builds on a Win7 system | ewmayer | Programming | 34 | 2010-10-18 22:36 |