- CLN-list - ginac.de

Re: Compiling CLN on Intel Macs
by Christian Fleck 09 Mar '06

09 Mar '06

Hi Richard, I attach the files you asked for. Thanks again for considering my problem. Christian

2 2

Re: Compiling CLN on Intel Macs
by Christian Fleck 07 Mar '06

07 Mar '06

Dear Bruno, actually I don't know how to do this here. I run configure with "-- disable-shared"; what else should I do? I fear that static linking is somewhat not easy, because it says in the man pages for gcc: "-static On systems that support dynamic linking, this prevents linking with the shared libraries. On other systems, this option has no effect. This option will not work on Mac OS X unless all libraries (includ- ing libgcc.a) have also been compiled with -static. Since neither a static version of libSystem.dylib nor crt0.o are provided, this option is not useful to most people." I tried to use the static flag with the result: "/usr/bin/ld: can't locate file for: -lcrt0.o collect2: ld returned 1 exit status" Christian

3 3

Compiling CLN on Intel Macs
by Christian Fleck 06 Mar '06

06 Mar '06

Hi Richard, thank you for your answer. I followed your advice and here is the output from gdb: gdb ./exam GNU gdb 6.1-20040303 (Apple version gdb-437) (Sun Dec 25 08:26:53 GMT 2005) Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "i386-apple-darwin"...Reading symbols for shared libraries .... done Function "cl_abort" not defined. (gdb) run Starting program: /usr/local/cln-1.1.11/tests/exam Reading symbols for shared libraries . done Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: KERN_INVALID_ADDRESS at address: 0x06a0afe1 0xa0b02584 in std::wclog () I attach the config.log Thanks, Christian

1 0

Date error
by Fabrice Marchant 17 Feb '06

17 Feb '06

> Then, my apologies. Thanks. It was not so important. Was angry cos I fear everybody thinks I'm a naughty boy. All right now.

1 0

My "effort" to make my mail keep the first place.
by Fabrice Marchant 16 Feb '06

16 Feb '06

> Mister Byron wrote : > I am deleting your message because the date is set far in advance, > in an apparent effort to make your email stand out over all others. What is "apparent" is not always true. I DID not wanted to do this. The date of this machine was false. It moved several years later indeed but I do not know how it happened. I have fixed this as soon I've noticed the problem. I use CLN lib for years and I love it. I'm NOT so stupid to use such a pitiful trick the first time I post to CLN-list ! But I couldn't care less you removed my msg. The important thing is I received interesting informations about the kind of hardware to build my next machine. Hopefully several persons didn't focused on this erroneous date and gently answered clever advices. Fabrice Marchant

2 1

A good PC hardware for CLN ?
by Fabrice Marchant 14 Feb '06

14 Feb '06

Hi ! I used CLN lib on linux debian Woody and try now libcln3 on debian Sarge ( on a Athlon T-Bird 1GHz 400Mo RAM ). I think to buy a more recent PC computer. What kind of PC for CLN - processor INTEL or AMD - to execute speedly integer computations ? Is motherboard worthy ? Thank you for your help. Fabrice

4 3

how to assign cl_I type to structure number
by mahesh ram 06 Feb '06

06 Feb '06

i am r.seetaram, M.Tech, C.S. i have faced the following problem: cl_I val=10; struct node { cl_I data; struct node * link; }; struct node * start=NULL; start = (struct node *) malloc(sizeof(struct node)); start->data = value; // error in this line; --------------------------------- Yahoo! Autos. Looking for a sweet ride? Get pricing, reviews, & more on new and used cars.

3 2

Re: Re: A good PC hardware for CLN ?
by Fabrice Marchant 28 Jan '06

28 Jan '06

> 1. Re: A good PC hardware for CLN ? (Joerg Arndt) > 2. Re: A good PC hardware for CLN ? (Richard B. Kreckel) Thank you very much for your answers. > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Tue, 24 Jan 2006 16:33:40 +0100 > From: Joerg Arndt <jj(a)suse.de> > I invariably recommend AMD64 these days (socket 939). > > My off the shelf 2.2GHz system is on par with a 3.6GHz intel > equivalent (EMT64(?), ex-pen-sive!) for the interesting stuff. System > is 2-3.5 times faster than Athlon at same clock. Memory throughput is > a dream. > > The timings given (by Granlund) are a true eye opener: > http://swox.com/doc/x86-timing.pdf > > Today's intel 64bit cores are a lousy hack at best. > > Alternative architectures (PPC) might be fine but are surely > much more expensive. I'll follow you and the document : AMD64 ! Are dual core useful ? I looked at the prices (France) in this cheap store : http://www.ldlc.com/navigation/cat.html?sscat=140025&constructeur=C00000080… I noticed on this page : - AMD Athlon 64 3700+ - 2.2 GHz, Cache L2 1 Mo Socket 939 0.09 micron - Core San-Diego (version boîte) 259,01 euros and - AMD Athlon 64 X2 Dual-Core 3800+ Socket 939 0.09 micron (version boîte) 338,99 euros. > > Message: 2 > Date: Tue, 24 Jan 2006 22:45:39 +0100 > From: "Richard B. Kreckel" <kreckel(a)ginac.de> > Even without x86-64 Assembler code, the use of 64 bit Digits gives an > very impressive performance gain. If you buy such a machine, make sure > to run int in 64 bit userland if arbitrary precision arithmetic is > important for what you're doing. Please, what does 64 bit userland mean ? Maybe recompiling your .cc sources for 64 bits will improve the speed too ? Does g++3.3 is able to produce such code ? Arbitrary precision arithmetic is a fun stuff and then very important to soften the hard life ! Sorry for my poor english language. When I wrote : "Is the motherboard worthy ?", that didn't had any meaning. I meant : "Is the motherboard important for speed too ?". I wonder if processor cache size is important for integer computations with CLN ? ( I focus on integers. ) And RAM amount ? Dual channel ? Can we forget hard drive perfs, hoping a huge RAM avoids to swap ? Regards Fabrice

3 2

CLN 1.1.11 released
by Richard B. Kreckel 22 Nov '05

22 Nov '05

Good evening, This is to announce the availability of CLN 1.1.11. It's a maintenance release with only one change: All input of numbers is now much speedier (except when the base is a power of two, in which case the performance was already optimal). As always, it is available from <http://www.ginac.de/CLN/>. Enjoy! -richy. -- Richard B. Kreckel <http://www.ginac.de/~kreckel/>

1 0

CLN 1.1.10 released
by Richard B. Kreckel 23 Oct '05

23 Oct '05

Good evening, This is to announce the availability of CLN 1.1.10. It's a maintenance release that fixes some outstanding issues, most notabliy compilation failures on some new versions of GCC. To quote the NEWS file: Implementation changes * Removed the vptr of modular integer and univariate polynomial classes in order to fix some crashes in that sector of CLN. Code using this needs to be recompiled. * Many more operator signatures make the integration of CLN types with float/double easier. Other changes * Several compilation fixes. * Made it possible to cross-compile CLN. On another note, the GiNaC website has finally moved to a new site. Along with it, CLN's CVS repository and mailing list have moved, too. Without this, most probably there wouldn't be a release yet. If you were using CVS, please check out a new tree. And, please, make sure you use the cln-list(a)ginac.de address, and not one with uni-mainz.de in it. If unsure: the web page has all the relevant information updated. Happy hacking -richy. -- Richard B. Kreckel <http://www.ginac.de/~kreckel/>

1 0