RE: gcc 2.95.2 on hpux 11

[Will Menninger <gcc at enterprise dot to dot bss dot boeing dot com>]

Mark,

I saw your e-mail on the gcc mail list.  I believe I'm using those same
utah binaries on our HPUX 11 (HP N4000 w/PA-8500 processors) system,
though I forget exactly where I snagged gcc 2.95.2 from.  I don't know
how they were derived, and I haven't been able to use them to generate
pa-risc 2.0 code--I don't think they will do this.  The gcc generated
pa-risc 1.1 code is very reliable and works just fine, but I've found
that HP's aCC compiler generates pa-risc 2.0 code which runs quite a bit
faster (approx. 50% or so on our codes, which are a mix of floating
point and integer) than the gcc generated pa-risc 1.1 code.

Here is an example of performance with a finite-element field solving
code which has a problem size of about 50 MB, but this was with a
slightly older version of gcc (I'm thinking gcc 2.7.x):

Platform           Compiler         Run Time   Run date
--------------------------------------------------------
PA-8500 440 MHz    aCC pa2.0        1:18.2     8-07-99
PA-8500 440 MHz    aCC pa1.1        1:27.5     8-06-99
PA-8500 440 MHz    gcc pa1.1        1:54.5     8-06-99

Good luck.

-Will


"V. Mark Lehky" writes:

>Dear Sirs,
>
>I have gotten your e-mail addresses from hpux.cs.utah.edu.  I have a
>question in regards to the binaries found at
><http://hpux.cs.utah.edu/hppd/hpux/Gnu/gcc-2.95.2/>.
>
>I myself have been trying to compile gcc 2.95.2 on our hpux 11.00 machine
>without success.  I eventually found that this is not supported; see
><http://gcc.gnu.org/install/specific.html#hppa*-hp-hpux11>.  Please note
>that we are running our machine in 32b mode.  My question is: how were the
>above binaries derived, and how reliable are they?
>
>Thank you, Mark L.
>
>-- 
>##
> # V. Mark Lehky
> # Software Test Engineer
> # Advanced Simulation Division
> #
> # PowerSuite: verification solution of tomorrow, available today
> ##
>