SEO

We have the need to compile code on a development PA-RISC box and push it to
a
locked down production box (no compilers allowed) running Itanium. I'm
pretty
sure you can cross compile with gcc but I've had no luck finding the info.

Any pointers?

TIA

Dan Mercer

On 28 Feb., 07:46, "Dan Mercer" <[email protected]> wrote:
> We have the need to compile code on a development PA-RISC box and push it to
> a
> locked down production box (no compilers allowed) running Itanium. I'm
> pretty
> sure you can cross compile with gcc but I've had no luck finding the info.
>
> Any pointers?

No need for cross compilation. Itanium can run PA-RISC code pretty
well. I did a test
once with my old C3750 (got it for 100 Euro from EBay :-).
But the gcc parisc code generator isn't the best try to get HP's c
compiler or use the latest
gcc which seem to have some improved code generator (sponsored from
HP).

"llothar" <[email protected]> wrote in message
news:48d0869b-0629-48aa-bb4a-33411ca7e682@e10g2000prf.googlegroups.com...
> On 28 Feb., 07:46, "Dan Mercer" <[email protected]> wrote:
>> We have the need to compile code on a development PA-RISC box and push it
>> to
>> a
>> locked down production box (no compilers allowed) running Itanium. I'm
>> pretty
>> sure you can cross compile with gcc but I've had no luck finding the
>> info.
>>
>> Any pointers?
>
> No need for cross compilation. Itanium can run PA-RISC code pretty
> well. I did a test
> once with my old C3750 (got it for 100 Euro from EBay :-).
> But the gcc parisc code generator isn't the best try to get HP's c
> compiler or use the latest
> gcc which seem to have some improved code generator (sponsored from
> HP).

1. We're dealing with gigabytes of data here, any emulation penalty negates
the value of moving
to faster hardware. Benchmarks indicate we'd be better off with native
code.
2. It all runs under gcc now - we've had some problems converting from gcc
to aCC.
3. We don't as yet have access to the development box that has itanium aCC
and we'd like
to get a head start.

Anyone else?

Dan Mercer

> 1. We're dealing with gigabytes of data here, any emulation penalty
> negates the value of moving to faster hardware. Benchmarks
> indicate we'd be better off with native code.

> 2. It all runs under gcc now - we've had some problems converting
> from gcc to aCC.

Is that how you got native code with which to compare in 1?

> 3. We don't as yet have access to the development box that has
> itanium aCC and we'd like to get a head start.

You might consider http://www.testdrive.hp.com/

rick jones
--
portable adj, code that compiles under more than one compiler
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

Dan Mercer wrote:
> 1. We're dealing with gigabytes of data here, ...
> Benchmarks indicate we'd be better off with native code.

Then you better be using HP's aC++ compiler. It does have an -Agcc option.

> 2. It all runs under gcc now - we've had some problems converting from
> gcc to aCC.

What kind?

On 1 Mrz., 05:09, "Dan Mercer" <[email protected]> wrote:
> "llothar" <[email protected]> wrote in message
>
> news:48d0869b-0629-48aa-bb4a-33411ca7e682@e10g2000prf.googlegroups.com...
>
>
>
> > On 28 Feb., 07:46, "Dan Mercer" <[email protected]> wrote:
> >> We have the need to compile code on a development PA-RISC box and push it
> >> to
> >> a
> >> locked down production box (no compilers allowed) running Itanium. I'm
> >> pretty
> >> sure you can cross compile with gcc but I've had no luck finding the
> >> info.
>
> >> Any pointers?
>
> > No need for cross compilation. Itanium can run PA-RISC code pretty
> > well. I did a test
> > once with my old C3750 (got it for 100 Euro from EBay :-).
> > But the gcc parisc code generator isn't the best try to get HP's c
> > compiler or use the latest
> > gcc which seem to have some improved code generator (sponsored from
> > HP).
>
> 1. We're dealing with gigabytes of data here, any emulation penalty negates
> the value of moving
> to faster hardware. Benchmarks indicate we'd be better off with native
> code.

Are you bullshitting me?

If you want use gcc, don't talk about fucking performance. This sucks
more then
then gcc generated code sucks (which is about 50% of the performace of
optimized
code from the Intel compiler). Emulation with optimized PA-RISC is
better then non
optimized gcc EPIC code.

What are your benchmarks?