Merlin Moncure <mmoncure@gmail.com> writes:
> On Sun, Dec 4, 2011 at 4:55 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Pavel Stehule <pavel.stehule@gmail.com> writes:
> it looks like gcc bug - gcc 4.5.1 20100924 (Red Hat 4.5.1) It was
> configured just with --enable-debug and --enable-cassert
>>
>> Is this x86? �I can't reproduce it on x86_64.
> reading all the comments in the gcc bug report, this is because x86
> targets the x87 fpu by default which is where the bug is -- it's a
> hardware problem. x86_64 targets sse which has stricter standards for
> rounding. most x86 processors support sse -- is there a reason why we
> don't target sse?
Well, older machines won't have sse, and in any case I think x86 is not
the only architecture with the issue, just the most popular one.
Floating-point registers that are wider than standard double are hardly
an unusual idea.
regards, tom lane