pa-risc1.1 pa-risc2.0

[Stan Sieler]

Code compiled for PA-RISC 1.0 (or 1.1)
will run fine on 2.0 in and of itself.

However, you'll usually find that 1.0 or 1.1 code was compiled on an older version of
HP-UX, and HP-UX doesn't have a perfect
record of backward compatibility for
old code running on newer releases.

In other words, try running it :)

We often compile code for 1.1 on HP-UX 11i v2 and ship it to clients with 2.0 systems ... because we also have clients with 1.1 systems on 11i v2 (the 1.1 will run on both, the 2.0 won't run on 1.1 systems).

The performance difference between the 1.1 and 2.0 code is generally negligble. (And, in some cases, 1.1 32-bit is faster than 2.0 64-bit ... your mileage may vary).

Stan
sieler@allegro.com

[Sandman!]

Matter of fact the stepwise procedure can be collapsed into a single pipelined command as follows:

# getconf CPU_VERSION | xargs -i echo 0d{}=x | adb | xargs -i grep 0x{} /usr/include/sys/unistd.h

...and the above will work for both PA-RISC and IPF machines.

[Dennis Handly]

The last supported OS that would produce 1.0 code was 10.10.
Where are you getting your PA 1.0 code?
(Of course you originally said 1.1.)

>Sandman: I would say porting from PA-RISC 1.0 to PA-RISC 2.0 should not be a problem

Yes, this should be forward compatible, unless kernel intrusive.

>Stan: The performance difference between the 1.1 and 2.0 code is generally negligible.

This is not true if doing a lot of shlib calls to small functions.