[Date Prev][Date Next][Thread Prev][Thread Next] [Search] [Date Index] [Thread Index]

Re: [MacPerl] acgi vs cgi (fwd)



According to Nat Irons:
> >>But I'm not sure how this relates to MacPerl, as MacPerl can only execute
> >>one at a time anyway, right?
> >
> >True, but it has provisions to handle more than one connection at once.
> >That's what an ACGI does- it doesn't do multiprocessing. =)
> 
> Multiprocessing, to me, means use of multiple processors.
> 
> My confusion is the difference, in Macperl's case, between .cgi (which I
> understand queues incoming requests and executes them sequentially) and
> .acgi (which from your description appears to do the same).  Am I missing
> something, or does acgi simply not mean much to MacPerl circa 1997?

Ummmmmmmm.....no.  Multiprocessing does not equal multiple
processors.  Multiprocessing means Multiple Processes.  The
Mac kind-of does multiprocessing.  You can check this out
by first bringing up BBEdit, MacPerl, QPQ, and Netscape.
Check under the "?" area.  You will see all of these things
are running on the system.  They are (if you want to
stretch the actual means of the word) multiprocessing.  In
my humble opinion though - what they are actually doing is
a form of timesharing.  But others will probably disagree.

Anyway, see my previous posts.  And BTW, you can have
multiple processors in a multiprocessing system but that is
sometimes referred to as parallel processing because you
can have multiple processors executing the same (or
different) instructions but the cpus work in parallel with
each other.  This is similar to DEC's method of figuring
out an address in memory.  Years ago (back in the early
80's) DEC created the Vax 11/780s.  In order to compute one
memory address there was something like a ten step process
which ran in parallel mode by splitting up the incoming
memory addres into various parts, adding, subtracting,
etc... to this number in different areas and finally
recombining the information into a single 64bit (or was it
128 bit?) location.  Learned this in their systems course.
Anyway, the whole process was done in a parallel mode using
multiple processors to perform the address manipulation.
Cray did the same kind of thing for both address location
as well as just processing information.  Not that you had
to use the multiple processors (you could just use one) but
their compilers actually would take your code and split it
up so it ran more efficiently with multiple processors.

***** Want to unsubscribe from this list?
***** Send mail with body "unsubscribe" to mac-perl-request@iis.ee.ethz.ch