David Gaudine scripsit:
|I should add that, in the unlikely event that -fno_automatic solves
|your problem, you should send the bug report to the developers of your
|source code. -fno_automatic solves a problem that should not come up
|with "correct" code.
Ahem, it didn't work. So far no solution
Douglas Bates scripsit:
|Are the differences in floating point results relative to other
|machines based on Intel processors? Remember that the floating point
|registers on Intel processors use an 80 bit IEEE extended
|representation. Other architectures such as SPARC use the 64 bit IEEE
|represe
On Thu, 7 Nov 1996, Arrigo Triulzi wrote:
> I would greatly appreciate any information or patches. A bug
> report is in the process of being prepared for both the g77 developers
> and the f2c developers. Incidentally, should I send a bug report to
> the Debian developers too?
I should add
On Thu, 7 Nov 1996, Arrigo Triulzi wrote:
> people here are having very serious problems with both g77 and
> f2c: the numbers which come out at the end are different from any
> other architecture we can lay our hands on, i.e. they have declared
> the boxes unfit for numerical calculations.
which version of f2c are you using? the last debian version 960717-0
is kind of old, and there is at least one bug in the 960717-0 version
which makes ishft fail miserably when dealing with signed integers.
this kind of messed up our random number generators big time. the f2c
maintainer was very re
Hello all,
people here are having very serious problems with both g77 and
f2c: the numbers which come out at the end are different from any
other architecture we can lay our hands on, i.e. they have declared
the boxes unfit for numerical calculations.
This is a pretty major proble
6 matches
Mail list logo
