🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.powerpc
Article #3210

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

#3210
From: Lawson English
Date: Fri, 23 Feb 1996 22:22
69 lines
4180 bytes
Joe Ragosta <jragosta@dca.net> wrote:
: In article <4gl353$d36@news-f.iadfw.net>, clyde@iadfw.net (Clyde Coffey) wrote:
[snipt]
: > 
: > The document you cite distinguishes between applications and server
: > processes. Copland starts a cooperative process for each application.
: > All cooperative processes share memory space.

: No. You're misreading it. Older applications which are not Copland savvy
: are cooperatively multitasked. New apps have a separate memory space and
: are shared preemptively.


Nope to both (sorta). *ALL* applications that have a (an?) human interface 
must sit in 
the "BLue Box" and effectively, they cooperatively multitaslk WITH EACH 
OTHER (but see below about what "effectively" means).

There is no "protected memory" qirh Blue Box tasks, so if one goes 
down, they all may. However, there may be a kind of "no man's land" 
memory protection 
between each BB task that will help reduce the likelihood of a Blue Box 
crash, even if a specific BB task goes boom.


Getting back to multi-tasking, the way things work on Copland is that ALL 
tasks are pre-emptive, even the Blue Box ones. HOWEVER, BLue Box tasks 
call WaitNextEvent, just as they do under System 7.x. When the currently 
active Blue Box task (System 7-style application) calls WNE, the PRocess 
Manager sets a flag that blocks that task from receiving any  
CPU time and resets that flag for the next BB task in the queue, effectively 
allowing only cooperative multi-tasking between BB tasks, just like under 
System 7.x.

However, the currently active BB task is in preemptive multi-tasking mode 
with the *rest* of the system, including any secondary tasks spawned by 
Copland-aware BB applications.

In theory, a well written COpland-savvy BB task could keep only the human 
interface stuff in the Blue Box and spawn secondary (fully pre-emptive 
with everthing and fully protected to boot) tasks to do all the work.

This could mean that even if your 3D VR game crashes the entire 
Blue Box, a downloaded Java applet running on a faceless Java interpreter 
could be grinding away, doing its thing, waiting for your web  browser 
(Cyberdog 
anyone?) to relaunch and re-establish human interface communications with 
its secondary tasks.

Technically, everything is pre-emptive under Copland, but only one BLue 
Box (System 7 or Copland app) task is allowed to receive CPU
time and all others are blocked until that active task calls WNE, 
effectively making all *applications* cooperatively multi-tasked, even 
ones written specifically for Copland.

Presumably, Gershwin (System 9) will simply remove the restriction 
about the WNE call and everything will be multi-tasking with everything.

BTW, I understand that there are 32 priority levels for CPU time under 
Copland, but I'm not sure how they apply to BB applications as opposed to 
secondary tasks.


--
-------------------------------------------------------------------------------
Lawson English                            __  __     ____  ___       ___ ____
english@primenet.com                     /__)/__) / / / / /_  /\  / /_    /
                                        /   / \  / / / / /__ /  \/ /___  /
-------------------------------------------------------------------------------
Technical Details
Message-ID: <4glem9$li5@nnrp1.news.primenet.com>
Path: arc1-ca.pugleaf.net!news.primenet.com!english
References: <nevin-0602962253080001@150.135.1.79> <danpop.824304112@rscernix> <4g00uj$gaq@nuke.csu.net> <danpop.824477998@rscernix> <4g3t0l$b71@nuke.csu.net> <danpop.824606547@rscernix> <mem-1902961857250001@durin.pha.jhu.edu> <4gckcj$de3@cville-srv.wam.u

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!news.primenet.com!english
From: Lawson English <english@primenet.com>
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....???
Followup-To: comp.sys.powerpc,comp.os.linux.advocacy,comp.sys.mac.advocacy,comp.os.os2.advocacy,comp.sys.amiga.advocacy,comp.sys.intel
Date: 23 Feb 1996 15:22:01 -0700
Organization: Primenet (602)395-1010
Sender: root@primenet.com
Message-ID: <4glem9$li5@nnrp1.news.primenet.com>
References: <nevin-0602962253080001@150.135.1.79> <danpop.824304112@rscernix> <4g00uj$gaq@nuke.csu.net> <danpop.824477998@rscernix> <4g3t0l$b71@nuke.csu.net> <danpop.824606547@rscernix> <mem-1902961857250001@durin.pha.jhu.edu> <4gckcj$de3@cville-srv.wam.u
X-Posted-By: english@usr4.primenet.com