[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-16 Thread William Stein
On Tue, Jun 16, 2009 at 5:33 PM, Dr. David Kirkby wrote: > > William Stein wrote: >> On Sun, Jun 14, 2009 at 10:21 AM, Dr. David >> Kirkby wrote: >>> The fix suggested by Clint to allow ATLAS to build (namely making >>> GuessSmallNB return 28) does indeed allow ATLAS to build. >>> >>> I could crat

[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-16 Thread Dr. David Kirkby
William Stein wrote: > On Sun, Jun 14, 2009 at 10:21 AM, Dr. David > Kirkby wrote: >> The fix suggested by Clint to allow ATLAS to build (namely making >> GuessSmallNB return 28) does indeed allow ATLAS to build. >> >> I could crate a patch file for that, which is only applied on Solaris, >> but c

[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-14 Thread William Stein
On Sun, Jun 14, 2009 at 10:21 AM, Dr. David Kirkby wrote: > > The fix suggested by Clint to allow ATLAS to build (namely making > GuessSmallNB return 28) does indeed allow ATLAS to build. > > I could crate a patch file for that, which is only applied on Solaris, > but clearly this is not a long-te

[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-14 Thread Dr. David Kirkby
The fix suggested by Clint to allow ATLAS to build (namely making GuessSmallNB return 28) does indeed allow ATLAS to build. I could crate a patch file for that, which is only applied on Solaris, but clearly this is not a long-term solution, as I assume it will degrade performance on Solaris.

[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-13 Thread Dr. David Kirkby
William Stein wrote: > 2009/6/13 Clint Whaley : >> The actual error you guys are showing below looks more like something >> really odd happened during an install: It is trying to choose a small >> block factor, and has chosen 0. Looking at the place the assertion >> failed, its GuessSmallNB() i

[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-13 Thread William Stein
2009/6/13 Clint Whaley : >>> After getting sqlite 3.5.3.p3 building on t2 with gcc-4.4.0, Sage is >>> falling over on atlas-3.8.3.p2. Before I waste any time on this, is a >>> fix known, or anyone have any useful information about the problem? >>> >>> I'd rather not waste time on it if the reason

[sage-devel] Re: atlas-3.8.3.p2 failing on Solaris 10 with gcc-4.4.0

2009-06-13 Thread William Stein
2009/6/12 Dr. David Kirkby : > > After getting sqlite 3.5.3.p3 building on t2 with gcc-4.4.0, Sage is > falling over on atlas-3.8.3.p2. Before I waste any time on this, is a > fix known, or anyone have any useful information about the problem? > > I'd rather not waste time on it if the reason for