does anyone know how to turn off the 'IEEE gradual underflow' mode
on Intel Pentium processors with Microsoft's C/C++ compiler?
In floating-point environments built on the architecture of some
of the SPARC processors (TI TMS390Z5 and TI TMS390S10), and
the Intel family, gradually underflowed results are almost always
calculated by the floating-point unit in the CPU or coprocessor.
Therefore, gradual underflow is much less likely to cause a
performance degradation than it is when implemented in software.
Although gradual underflow is handled by the Intel-FPU and not by
software, turning off gradual underflow can show a remarkable
I couldn't find any function in Microsoft's clib which allows
to switch to non-standard arithmetic mode, thus turning off
gradual underflow. There exists a function called '_control87',
with which you can toggle bits of the floating point exception mask
and therefore control which fp-exceptions are handled internally
[default] or by your own software. But it seems these _control87-related
functions don't allow me to switch to non-standard arithmetic mode.
I'm sure there exists a possibility to turn off standard underflow
mode with the Digital Fortan compiler for Windows/Intel platforms;
Any pointer to this matter is highly appreciated!
Phone: +49-234-x Office:x=700-7988 Secretary:x=700-7997 FAX:x=7094-210