This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Architecture floating-point underflow information wanted
- From: Richard Henderson <rth at twiddle dot net>
- To: Uros Bizjak <ubizjak at gmail dot com>
- Cc: libc-alpha at sourceware dot org, "Joseph S. Myers" <joseph at codesourcery dot com>
- Date: Tue, 25 Sep 2012 10:06:46 -0700
- Subject: Re: Architecture floating-point underflow information wanted
- References: <CAFULd4awrK4xxbYd7Ln65pWFAvUBagOe0FNO_M5df+9XUH+efA@mail.gmail.com>
On 09/25/2012 08:29 AM, Uros Bizjak wrote:
>> > alpha
> cpu : Alpha
> cpu model : EV68AL
> cpu variation : 7
> cpu revision : 0
>
> The test returns "after rounding".
FWIW, you're testing the kernel implementation.
# The 21264 will not produce a Denormal result for the underflow
# exception. Instead, a true zero (+0) is written to the destination
# register. In the 21264 the FPCR underflow to zero (UNDZ) bit must
# be set if underflow disable (UNFD) bit is set. If desired, trapping
# on underflow can be enabled by the instruction and the FPCR, and
# software may compute the Denormal value as defined in the IEEE Standard.
r~