1
CMSC 412 - S96 (lect 11) copyri
g
ht 1996 Jeffrey K. Hollin
g
sworth
Announcements
●Reading 8 (8.1-8.2, 8.5-8.6)
●Project #3 was handed out in section
– proc2.c is now available
– will need to produce a short paper writeup for this
assignment
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
OS Lecture Notes: Process Scheduling & Memory Mgmt in UNIX & Windows NT, Study notes of Operating Systems
A portion of lecture notes from a computer science course (cmsc 412) in the fall semester of 1996. The notes cover the topics of process scheduling and memory management in both unix and windows nt operating systems. Information on priority algorithms, feedback scheduling, unix system priorities, and windows nt scheduling. It also discusses the advantages of virtual addressing and paging.
Typology: Study notes
Pre 2010
1 / 12
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download OS Lecture Notes: Process Scheduling & Memory Mgmt in UNIX & Windows NT and more Study notes Operating Systems in PDF only on Docsity!
copyright 1996 Jeffrey K. Hollingsworth
Announcements
l
Reading 8 (8.1-8.2, 8.5-8.6)
l
Project #3 was handed out in section
- proc2.c is now available– will need to produce a short paper writeup for this
assignment
copyright 1996 Jeffrey K. Hollingsworth
Priority Algorithms
l
Fixed Queues
- processes are statically assigned to a queue– sample queues: system, foreground, background
l
Multilevel Feedback
- processes are dynamically assigned to queues– penalize jobs that have been running longer– preemptive, with dynamic priority– have
N
ready queues (RQ0-RQ
N
- start process in RQ0• if quantum expires, moved to i + 1 queue
copyright 1996 Jeffrey K. Hollingsworth
UNIX System V
l
Multilevel feedback, with
- RR within each priority queue– 10ms second preemption– priority based on process type and execution history, lower
value is higher priority
l
priority recomputed once per second, and schedulerselects new process to run
l
For process j, P(i) = Base + CPU(i-1)/2 + nice
– P(
i) is priority of process
j at interval
i
- Base is base priority of process
j
– CPU(
i) = U(
i)/2 + CPU(
i-1)/
• U(
i) is CPU use of process
j in interval
i
- exponentially weighted average CPU use of process
j
through interval
i
- nice is user-controllable adjustment factor
UNIX (cont.) copyright 1996 Jeffrey K. Hollingsworth
l
Base priority divides all processes into (non-overlapping) fixed bands of decreasing priority levels
- swapper, block I/O device control, file manipulation,
character I/O device control, user processes
l
bands optimize access to block devices (disk), allowOS to respond quickly to system calls
l
penalizes CPU-bound processes w.r.t. I/O bound
l
targets general-purpose time sharing environment
copyright 1996 Jeffrey K. Hollingsworth
Windows NT (cont.)
l
In real-time class, all threads have fixed priority thatnever changes
l
In variable class, priority begins at an initial value,and can change, up or down
- FIFO queue at each level, but thread can switch queues
l
Dynamic priority for a thread can be from 2 to 15
- if thread interrupted because time slice is up, priority lowered– if interrupted to wait on I/O event, priority raised– favors I/O-bound over CPU-bound threads– for I/O bound threads, priority raised more for interactive
waits (e.g. keyboard, display) than for other I/O (e.g. disk)
copyright 1996 Jeffrey K. Hollingsworth
Managing Memory
l
Main memory is big, but what if we run out
- use virtual memory– keep part of memory on disk
- bigger than main memory• slower than main memory
l
Want to have several program in memory at once
- keeps processor busy while one process waits for I/O– need to protect processes from each other– have several tasks running at once
- compiler, editor, debugger• word processing, spreadsheet, drawing program
l
Use
virtual addresses
- look like normal addresses– hardware translates them to
physical addresses
copyright 1996 Jeffrey K. Hollingsworth
Paging
l
Divide physical memory into fixed sized chunkscalled
pages
- typical pages are 512 bytes to 64k bytes– When a process is to be executed, load the pages that
are
actually used into memory
l
Have a table to map virtual pages to physical pages
l
Consider a 32 bit addresses
- 4096 byte pages (12 bits for the page)– 20 bits for the page number
PageTable
Main Memory
Virtual Address
Location
Present
Rd/Write
20 bits
12 bits
copyright 1996 Jeffrey K. Hollingsworth
Problems with Page Tables
l
One page table can get very big
20
entries (for most programs, most items are empty)
l
solution1: use a hierarchy of page tables
Page Table
Main Memory
Virtual Address
10 bits
12 bits
Page Directory
10 bits
Pg Tbl Ptr
Physical Page #