Re: atlas depends on gcc43 *and* gcc44 ???
Hi, Is there a downside to having atlas depend on gcc44 while everything else depends on gcc43? Its just a little annoying that a package like atlas, which is a dependency of a dependency drags in a big build like gcc when I have no other reason for it... I understand that some packages may need a newer compiler than the standard OSX ones, but it would be nice if they could agree on which one ;) Chris ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On Jun 2, 2010, at 3:35 PM, Chris Jones wrote: Hi, Is there a downside to having atlas depend on gcc44 while everything else depends on gcc43? Its just a little annoying that a package like atlas, which is a dependency of a dependency drags in a big build like gcc when I have no other reason for it... I understand that some packages may need a newer compiler than the standard OSX ones, but it would be nice if they could agree on which one ;) I was wondering why anything depends on atlas at all. Is atlas noticeably better than the lapack and blas routines in the Accelerate framework? I couldn't find any comparisons on-line. -- Steve -- -- stephen.lan...@nist.govTel: (301) 975-5423 -- -- http://math.nist.gov/mcsd/Staff/SLanger/ Fax: (301) 975-3553 -- -- NIST, 100 Bureau Drive, Stop 8910, Gaithersburg, Md 20899-8910 -- -- I don't think this will work. That's why it's science. -- -- Naomi Langer (age 6), 17 Feb 2003 -- ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On 06/02/2010 11:13 PM, Stephen Langer wrote: On Jun 2, 2010, at 3:35 PM, Chris Jones wrote: Its just a little annoying that a package like atlas, which is a dependency of a dependency drags in a big build like gcc when I have no other reason for it... I understand that some packages may need a newer compiler than the standard OSX ones, but it would be nice if they could agree on which one ;) I was wondering why anything depends on atlas at all. Is atlas noticeably better than the lapack and blas routines in the Accelerate framework? I couldn't find any comparisons on-line. Yep, that's something I suggested some weeks ago, too [1]. I am sure there are valid reasons to use our own atlas, but I don't know them. Rainer [1] http://lists.macosforge.org/pipermail/macports-dev/2010-April/011682.html ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On Jun 2, 2010, at 14:35, Chris Jones wrote: Its just a little annoying that a package like atlas, which is a dependency of a dependency drags in a big build like gcc when I have no other reason for it... I understand that some packages may need a newer compiler than the standard OSX ones, but it would be nice if they could agree on which one ;) AFAIK there *was* a consensus to use gcc43 by default for science ports. I had not heard any update on that decision, and pointed this discrepancy out to Marcus who updated atlas and qrupdate to use gcc44 instead. See the mailing list link I sent you in a previous reply in this thread. ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On Wed, Jun 2, 2010 at 17:40, Ryan Schmidt ryandes...@macports.org wrote: AFAIK there *was* a consensus to use gcc43 by default for science ports. I had not heard any update on that decision, and pointed this discrepancy out to Marcus who updated atlas and qrupdate to use gcc44 instead. See the mailing list link I sent you in a previous reply in this thread. That was a while ago so maybe it's time we revisit this? With my py25-numpy and py2{5,6}-scipy maintainers hat on I'd vote for a switch to gcc44 as the default compiler for scientific ports. Cheers Adam ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On Jun 2, 2010, at 18:05, Adam Mercer wrote: On Wed, Jun 2, 2010 at 17:40, Ryan Schmidt wrote: AFAIK there *was* a consensus to use gcc43 by default for science ports. I had not heard any update on that decision, and pointed this discrepancy out to Marcus who updated atlas and qrupdate to use gcc44 instead. See the mailing list link I sent you in a previous reply in this thread. That was a while ago so maybe it's time we revisit this? With my py25-numpy and py2{5,6}-scipy maintainers hat on I'd vote for a switch to gcc44 as the default compiler for scientific ports. Great! Now we're getting somewhere. :) Before we go updating everything to gcc44, let's make sure gcc44 works everywhere we need it to. Do we have any thoughts on the ticket where gcc44 can't be built on Tiger Intel? Anybody else see that problem or is it only on my machine? Do we care about Tiger anymore? http://trac.macports.org/ticket/20753 What about this ticket where gcc44 can't be built on (some?) Leopard machines? Anybody else see this problem? http://trac.macports.org/ticket/20635 ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On Wed, Jun 2, 2010 at 18:37, Ryan Schmidt ryandes...@macports.org wrote: That was a while ago so maybe it's time we revisit this? With my py25-numpy and py2{5,6}-scipy maintainers hat on I'd vote for a switch to gcc44 as the default compiler for scientific ports. Great! Now we're getting somewhere. :) Before we go updating everything to gcc44, let's make sure gcc44 works everywhere we need it to. Do we have any thoughts on the ticket where gcc44 can't be built on Tiger Intel? Anybody else see that problem or is it only on my machine? Do we care about Tiger anymore? http://trac.macports.org/ticket/20753 I haven't had access to a Tiger machine for a while, plus it's no longer supported by Apple... What about this ticket where gcc44 can't be built on (some?) Leopard machines? Anybody else see this problem? http://trac.macports.org/ticket/20635 I originally reported this ticket, but I successfully built gcc44 on both an Intel and PPC Leopard machines earlier this week. Can't recall when this started working for me, this one fell of my radar (probably when I updated by main machine to Snow Leopard). Cheers Adam ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
On Jun 2, 2010, Stephen Langer wrote: I was wondering why anything depends on atlas at all. Is atlas noticeably better than the lapack and blas routines in the Accelerate framework? I couldn't find any comparisons on-line. I was wondering the same thing as well. Since I am only installing atlas to get gimp to work anyway I would prefer not to have to build it for some miniscule speed increase in some obscure filter (if gimp even actually ends up using any parts of blas or lapack, but I guess it's possible it's important). Scott ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users
Re: atlas depends on gcc43 *and* gcc44 ???
Le 2 juin 2010 à 23:13, Stephen Langer a écrit : I was wondering why anything depends on atlas at all. Is atlas noticeably better than the lapack and blas routines in the Accelerate framework? I couldn't find any comparisons on-line. I am unsure. But, as long as Apple does not state that its Blas and Lapack libraries are OpenCL based, in which case they may be hundred times quicker that CPU-thread-based Atlas, I am lead to believe that Atlas compiled with the latest gcc45 should be more efficient than Apple blas or lapack, most probably generated with gcc42 (the last GPL2 version) and backward-compatible with MacOS 10.4 or 10.5; Maybe they can get better efficiency with Clang/LLVM. However, I faintly remember comparing the Atlas built-in tests, and there was a clear gap between gcc43 and gcc44. It should not be very difficult to build up a test, like finding the eigenvalues of a random but symmetric definite positive matrix of great size (say : 50,000 x 50,000 or more) in Xcode, and link either with the built-in framework or our Atlas. Cheers Vincent PS : Does Apple provide scipy in its Python distribution? ___ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo.cgi/macports-users