On 6/23/06, Carl Lowenstein <[EMAIL PROTECTED]> wrote:
On 6/22/06, Carl Lowenstein <[EMAIL PROTECTED]> wrote:
>
> It's pretty well known that floating point on the Alpha is not quite
> compliant with IEEE754 with regard to exception handling. I have
> learned that there is a .configure option in with the Python sources:
> "--with-fpectl" meaning floating-point exception control.
'>
"./configure --with-fpectl" doesn't seem to help.
For the one or two people who might still be reading this thread, I
think I have finally discovered the secret sauce. It is well hidden
in "info gcc" and in fact is not documented in the info file on the
Alpha (gcc 2.95.4). The info file on my FC3 system (gcc 3.4.4) says:
'-mieee'
The Alpha architecture implements floating-point hardware
optimized for maximum performance. It is mostly compliant with
the IEEE floating point standard. However, for full compliance,
software assistance is required. . . .
So adding -mieee to CFLAGS in the Makefile seems to actually cure the
problem of detecting Floating Point Exceptions. Both the expected
presence of FPE and the expected absence of FPE are now handled in the
Python test suite Note -- I didn't have to use a newer gcc, just use
the semi-undocumented flag.
carl
--
carl lowenstein marine physical lab u.c. san diego
[EMAIL PROTECTED]
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
