Date: Sat, 15 Sep 2007 19:07:26 -0700 From: Steve Kargl <sgk@troutmask.apl.washington.edu> To: "Thomas D. Dean" <tomdean@speakeasy.org> Cc: freebsd-current@freebsd.org Subject: Re: Compiler Problems? Message-ID: <20070916020726.GB6447@troutmask.apl.washington.edu> In-Reply-To: <200709160102.l8G12bAl001481@dv6000.tddhome> References: <200709152159.l8FLxfha008167@dv6000.tddhome> <20070915222422.GA4487@troutmask.apl.washington.edu> <200709152255.l8FMtMSD008381@dv6000.tddhome> <20070915231526.GA4788@troutmask.apl.washington.edu> <200709152339.l8FNdVK1008558@dv6000.tddhome> <20070915235852.GB4998@troutmask.apl.washington.edu> <200709160102.l8G12bAl001481@dv6000.tddhome>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, Sep 15, 2007 at 06:02:37PM -0700, Thomas D. Dean wrote: > (gdb) bt > #0 0x082c6050 in slamc3_ () > #1 0x082c6b08 in slamc1_ () > #2 0x082c6375 in slamc2_ () > #3 0x082c6bb9 in slamch_ () Ugh. Make sure that slamch.f (which is part of BLAS/LAPACK) is compiled without optimization. This routine tries to use clever tricks to determine machine parameters (e.g., EPSILON), and it can get stuck in infinite loops with optimization. This is known issue with LAPACK. See http://www.netlib.org/lapack/faq.html#1.25 -- Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20070916020726.GB6447>