Date: Wed, 30 Dec 1998 15:17:35 +0000 (GMT) From: Doug Rabson <dfr@nlsystems.com> To: Poul-Henning Kamp <phk@critter.freebsd.dk> Cc: Hidetoshi Shimokawa <simokawa@sat.t.u-tokyo.ac.jp>, dfr@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: src/sys/alpha/alpha vm_machdep.c machdep.c fp_emulate.c trap.c src/sys/alpha/include vmparam.h src/sys/kern init_main.c kern_exec.c src/sys/sys imgact.h Message-ID: <Pine.BSF.4.01.9812301515510.1813-100000@herring.nlsystems.com> In-Reply-To: <87569.915030208@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 30 Dec 1998, Poul-Henning Kamp wrote: > In message <Pine.BSF.4.01.9812301501080.1813-100000@herring.nlsystems.com>, Dou > g Rabson writes: > >On Wed, 30 Dec 1998, Hidetoshi Shimokawa wrote: > > > >> dfr> * Disable all floating point exceptions by default. > >> > >> My understanding is that we still get SIGFPE without > >> "-mtrap-precision=i -mfp-trap-mode=su" option of cc and default > >> cc doesn't have this option. Am I right? > > > >That is true (I think). I'm not quite sure what happens if exceptions are > >disabled but an operation generates an underflow. > > Don't you just get denormalized numbers, NaNs and similar monsters ? On the alpha, denorms, INFs and NaNs are not generated by the hardware. If a corner-case expression is evaluated, it can optionally trap to a kernel floating point emulator which can handle all the gory details of overflows and underflows. -- Doug Rabson Mail: dfr@nlsystems.com Nonlinear Systems Ltd. Phone: +44 181 442 9037 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.01.9812301515510.1813-100000>