🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.powerpc
Article #3119

Re: Debate: PowerMac "trounces" Pentium....???

#3119
From: Charles Bouldin
Date: Sat, 10 Feb 1996 02:53
48 lines
2727 bytes
In article <713.6612T338T2196@und.ida.liu.se>, c92manen@und.ida.liu.se
(Mans Engman) wrote:

> [snip]
> 
> >I already posted that the Language Systems and Absoft fortran compilers
> >will do exactly what you want. I run many computationally intensive codes
> >on a PowerMac with perfectly smooth background execution "via a simple
> >compilation". The compilers put in the proper code to support the Mac OS
> >multitasking and it works just fine. The source code remains absolutely
> >unchanged portable ANSI 77 fortran.
> 
> "Proper code"?

Let's try different phrasing: Code that works correctly and causes the
program to multitask.
> 
> Ok, I'm rather ignorant about McI****h programming, but I just wonder
under what
> circumstances the compilers put in that code? Does it put taskswitching code
> every XXX loops? And how can I be sure it doesn't put that code there
when it's
> not necessary?

It puts it where you tell it to, through compiler options, like this:

-bkg=0         never yield time to other processes
-bkg=1         yield time when the program stops at an i/o statement
-bkg=2         yield time when you hit a subroutine entry point
-bkg=3         yield time at the top of every loop
-bkg=4         yield time whenever you hit a statement with a line #

Each "level" includes the context switches implied by all the lower
levels. In addition you can tell the compiler how many ticks (1/60th
second) must elapse between times when your code is swapped out. I
typically set this to some number comparable to my reaction time, so then
I can let my fortran codes run in the background, then just click on
something and "instantly" I have control over the computer running my word
processor or whatever.

I have no idea why it is only the fortran compilers that support this on
the Mac. It is a wonderful feature and works flawlessly.

I've even ported "real time" data collection codes and gotten them to work
in the background while I run data analysis codes in the foreground. The
big lie is that the Mac can't do that, but it does. And note, this is done
on a Mac IIci, not the latest
PowerMac.
Technical Details
Message-ID: <bouldin-0902962053250001@async098.nist.gov>
Path: arc1-ca.pugleaf.net!async098.nist.gov!user
References: <bouldin-0602961057070001@dafs.nist.gov> <713.6612T338T2196@und.ida.liu.se>

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!async098.nist.gov!user
From: bouldin@enh.nist.gov (Charles Bouldin)
Newsgroups: comp.sys.powerpc,comp.os.linux.advocacy,comp.sys.mac.advocacy,comp.os.os2.advocacy,comp.sys.amiga.advocacy,comp.sys.intel
Subject: Re: Debate: PowerMac "trounces" Pentium....???
Date: Fri, 09 Feb 1996 20:53:25 -0600
Organization: NIST
Message-ID: <bouldin-0902962053250001@async098.nist.gov>
References: <bouldin-0602961057070001@dafs.nist.gov> <713.6612T338T2196@und.ida.liu.se>
NNTP-Posting-Host: async098.nist.gov
X-Newsreader: Value-Added NewsWatcher 2.0b24.0+