There is one issue that came up during the "C++ on
MPE" discussion that really needs to be extended to
compilers in general on MPE/iX:
On HP-UX at least some compilers (C, C++, and
I'm not sure how many others) currently provide the
following five choices for compiler optimization:
(level 0 = none)
level 1 = local-only (minimal)
level 2 = global (same as MPE current maximum)
level 3 = all above + across procedures (better)
level 4 = all above + involve loader at load time (best)
Note not all MPE/iX native mode compilers can even
provide level 2 optimization; some only do level 1.
So why is MPE/iX a poor stepchild at the HP language
labs ??....... Of course if level 2 code runs only half as
fast as level 4 (I have no idea what the actual spread is)
you need to buy a processor that is twice as fast to do
the same work, and your HP sales rep will be happy......
How about it HP: When can HP 3000 users expect
their HP-supported NM compilers to support level 3
and level 4 optimization ????
Ken Sletten