mirror/freebsd-ports
1
0
Fork 0
mirror of https://git.FreeBSD.org/ports.git synced 2024-11-30 01:15:52 +00:00
Code Issues Releases Activity
208,242 Commits 45 Branches 100 Tags 3.3 GiB
a7540350a8
Commit Graph

66 Commits

Author SHA1 Message Date
Maho Nakata
0c9e2954a8 Update to 3.8.2. 
Submitted by:	 Kengo Ichiki <ichiki@ualberta.ca>
Slightly modified so that his patches fit ports policy.
 2008-06-12 07:28:33 +00:00
Maho Nakata
deced0b16f * Fix bug listed in http://math-atlas.sourceforge.net/errata.html#RMAAT . 
* Bump portrevision.
* Fix pkg-descr. Still do not pass portlint though.

Submitted by:	bf <bf2006a@yahoo.com>
 2008-02-07 07:47:19 +00:00
Maho Nakata
c2b23d6c81 Use MANUAL_PACKAGE_BUILD instead. 2008-01-07 00:24:25 +00:00
Maho Nakata
311264a24f Update to 3.8.0. 
Only verified on i386 and amd64 FreeBSD 7BETA.

Discussions with: bf <bf2006a@yahoo.com>, truckman
and Tijl Coosemans <tijl@ulyssis.org>.

PR: 118270 and 118271
 2007-12-19 22:47:57 +00:00
Maho Nakata
6befbb3a3b Using new LAPACK 3.1.1 port (threfore BUMPPORTREVISION) 
and unbreak.
 2007-09-30 04:02:10 +00:00
Maho Nakata
f14bc5b144 Temporary marked as broken. Creating ATLAS'ed LAPACK part is now broken. 2007-09-29 12:57:39 +00:00
Maho Nakata
b4be9b354e In ATLAS/INSTALL.txt it is said that gcc42 is superior to both gcc3 and gcc41 
- and FreeBSD 7 current imported gcc42.

Spotted by:	Li-Lun \"Leland\" Wang" <llwang@infor.org>
 2007-07-30 08:04:04 +00:00
Maho Nakata
ec963b99bd Now we use USE_FORTRAN=yes to use FORTRAN compiler. 2007-07-12 08:13:25 +00:00
Maho Nakata
c33f0d8ede Fix errata listed at http://math-atlas.sourceforge.net/errata.html and 
bump portrevision.

Submitted by:	bf <bf2006a@yahoo.com>
 2007-03-08 08:26:19 +00:00
Maho Nakata
66a388ece5 For test target, now we use GMAKE to MAKE. In some environment 
GMAKE fails.
 2007-02-19 01:49:51 +00:00
Maho Nakata
968c3e16c6 Remove CONFLICT line as elmer-mathlibs are deprecated by migration 
to gfortran.
 2007-01-16 02:39:08 +00:00
Maho Nakata
2a8a26e219 Atlas should not depend on GNU F77. Now 
it doesn't expect existence of /usr/include/g2c.h
 2007-01-15 05:45:25 +00:00
Thierry Thomas
6ae81701d8 gmake is needed as a build dependency. 
Approved by:	maho
 2007-01-14 13:26:55 +00:00
Maho Nakata
31bae9c2b3 Fix build with WITH_OPTIMIZED_FLAGS. 2007-01-11 21:42:56 +00:00
Maho Nakata
ac0188f25e * Migrate to gfortran. but stock GCC is used. 
* lapack/blas are compiled with gforntran, thus we build TEST section as well.
 2007-01-11 07:03:13 +00:00
Maho Nakata
2a8418244e * Shared lib number has been bumped 1 to 2 
* Bump port revision
 2007-01-09 22:39:17 +00:00
Maho Nakata
6d395117ef * Migrate to gfortran. 
* portlint.
* (option) Installing benchmark tests for real FLOPS.
* Bump portrevision.
 2007-01-06 22:18:51 +00:00
Maho Nakata
9a93aff2da Fix error when WITH_OPTIMIZED_FLAGS is specified. 
PR:		100991
Submitted by:	Masayuki Takemura <ICG04984@nifty.com>
 2006-08-07 01:40:52 +00:00
Maho Nakata
21c6b5f817 A better WITH_OPTIMIZED_FLAGS. 2006-07-09 01:35:28 +00:00
Edwin Groothuis
efe7d5f7f4 Remove USE_REINPLACE from ports starting with M 2006-05-10 22:37:39 +00:00
Maho Nakata
cd2fb50f10 Register conflicts with math/elmer-mathlibs. 
Submitted by:	thierry and
		Pedro F. Giffuni <giffunip@asme.org>
 2006-04-26 05:13:41 +00:00
Maho Nakata
28171a1dc8 o cleanups 
o faster build time; 4h45m19.23s for my Pentium M 1.6G.
  now static libs are build via knob, timing torerance patch is applied via knob.

Submitted by:	many
 2006-02-14 02:30:16 +00:00
Edwin Groothuis
53354e7e9c SHA256ify 
Approved by: krion@
 2006-01-22 12:59:47 +00:00
Maho Nakata
a01ed8fdc4 Currently with "cpu I686_CPU" which is default, 
kernel automatically enables SSE if it is available.
so remove messages which are not necessary.

Submitted by:	 Darrien <darrien.l@gmail.com>
 2005-11-19 22:45:34 +00:00
Maho Nakata
818f3b359e Add cputype of sempron; athlon type CPU 
Submitted by: Pedro A M Vazquez <vazquez@iqm.unicamp.br>
 2005-10-30 08:00:58 +00:00
Kris Kennaway
5d5c1d4ad3 Set NO_PACKAGE: atlas optimizes for the local machine, so a package is not 
generally useful.

Reviewed by:	maho (maintainer)
 2005-09-07 01:52:21 +00:00
Maho Nakata
852063c6cd Add two patches for stable builds. 
o L1 cache size is not correctly detected having a bad clock
o in some case small number of tries in tuning doesn't meet convergence
  see FAQ http://math-atlas.sourceforge.net/errata.html#tol
 2005-06-28 04:45:59 +00:00
David E. O'Brien
c3d8037aeb At Kris's request, back out the MACHINE_ARCH spelling correction until 
after 5.4-RELEASE.
 2005-04-12 03:26:56 +00:00
David E. O'Brien
f2fc2d60ae Assist getting more ports working on AMD64 by obeying the 
Ports Collection documentation and use 'ARCH' rather than 'MACHINE_ARCH'.
 2005-04-11 08:04:41 +00:00
Maho Nakata
7759e1bf9d Add a warning message 
Suggested by: Rob D. <162144@gmail.com>
 2005-01-25 05:47:56 +00:00
Maho Nakata
e1a067b7f1 Reduce optimization level so that not get trapped 
at infinite loop
 2005-01-25 05:40:14 +00:00
Maho Nakata
df33287cc1 Avoid the use of != statements in ports Makefile 
PR:		71095
Submitted by:	eik
 2004-10-24 03:37:02 +00:00
Maho Nakata
26bd231cac Adding newline at the end of files 
(lapack.h and blas.h)

Submitted by:	thierry@
 2004-10-19 01:10:44 +00:00
Kris Kennaway
84e8d56ed5 Increase USE_GCC to 3.4 for those ports which compile with it. 
Approved by:	portmgr
 2004-09-30 05:32:00 +00:00
Maho Nakata
162ea9b647 type -> arch 
Pointy hat to: maho
 2004-09-02 00:25:45 +00:00
Maho Nakata
8b77f79ccd ad hoc fix for 5-STABLE, having CPU 
other than Pentium4 (Opteron, Pentium III,
Athlon... etc) this was due to recent change
in make(1)
 2004-09-01 14:26:13 +00:00
Maho Nakata
0cb5e86302 Build fix for 5-STABLE 2004-09-01 09:59:25 +00:00
Maho Nakata
4559055c68 o Significantly performance loss by overriding several 
optimization flags...
o update to 3.7.8
 2004-08-18 05:46:25 +00:00
Maho Nakata
94bcc3f41a 1) correct the names of shared libs for threaded version 
2) Some Pentium 4 series have more caches (not 512kB, but 1M ~ 2M)
 2004-07-14 04:49:52 +00:00
Maho Nakata
c72ea155ce Make portlint happy by replacing space to tab 
PR:		65409
Submitted by:	trevor
 2004-05-06 19:49:54 +00:00
Maho Nakata
db4e7c3aca make portlint happy 2004-03-30 14:49:59 +00:00
Maho Nakata
7aa685c787 real fix for PR:64692 
PR:		64692
Submitted by:	Kay Lehmann <kay_lehmann@web.de>
 2004-03-30 14:48:49 +00:00
Maho Nakata
410118a607 unbreak for 4-STABLE i386 and alpha 
PR:	64692 (loosely depended on)
Submitted by:	Kay Lehmann <kay_lehmann@web.de> and kris via bento
 2004-03-27 05:55:45 +00:00
Maho Nakata
308503d550 Do not hardcord PICFLAG 
Submitted by:	ru
 2004-03-20 05:20:46 +00:00
Maho Nakata
10e9ee4f10 Backout previous change of PICFLAG, since 
behaviour of -fPIC and -fpic are different.
Here is the comment form obrien:

--
"-fpic" is a [minor?] optimization for machines that can handle it:

  -fpic
    Generate position-independent code (PIC) suitable for use in a shared
    library, if supported for the target machine. Such code accesses all
    constant addresses through a global offset table (GOT). The dynamic
    loader resolves the GOT entries when the program starts (the dynamic
    loader is not part of GCC; it is part of the operating system). If
    the GOT size for the linked executable exceeds a machine-specific
    maximum size, you get an error message from the linker indicating
    that -fpic does not work; in that case, recompile with -fPIC instead.
    (These maximums are 16k on the m88k, 8k on the SPARC, and 32k on the
    m68k and RS/6000. The 386 has no such limit.)

  -fPIC
    If supported for the target machine, emit position-independent code,
    suitable for dynamic linking and avoiding any limit on the size of
    the global offset table. This option makes a difference on the m68k,
    m88k, and the SPARC.

Thanks to: obrien
 2004-03-19 04:20:12 +00:00
Maho Nakata
fb23a1b086 -fPIC is appropreate on all architectures. 
Submitted by: kris
 2004-03-18 23:06:11 +00:00
Maho Nakata
3718781425 Fix build on alpha 4-stable 
Submitted by:	kris via bento
 2004-03-17 09:46:36 +00:00
Maho Nakata
5e41ece017 Update to 3.6.0 2004-02-28 09:45:06 +00:00
Maho Nakata
f73c1421f5 math/atlas should not LIB_DEPENDS on math/lapack. 
lapack is required only for building.

PR:		61963
Submitted by:	AMAKAWA Shuhei <sa264@cam.ac.uk>
 2004-02-01 02:04:48 +00:00
Maho Nakata
f97bb94e5f Backout to 3.5.5 
a) newer version build fails like:
/usr/local/lib/libatlas.so: undefined reference to `ATL_dpKBmm_b1'
/usr/local/lib/libatlas.so: undefined reference to `ATL_dpKBmm_b0'

b) Define PORTEPOCH=1

Reported by:	Pedro F. Giffuni <giffunip@yahoo.com>
		kris
		Thierry Thomas <thierry@pompo.net>
		Jonathan <j.e.drews@worldnet.att.net>
 2003-08-08 10:42:58 +00:00