Processes and Threads - Graduate Computer Science Systems

Download Report

Transcript Processes and Threads - Graduate Computer Science Systems

Recitation 6

Virtual Memory Sample Questions Project 3 – Build Pthread Lib Extra Office Hour: Wed 4pm-5pm HILL 367 1

Registers

Memory Hierarchy

Memory Virtual Memory Cache Answer: Pretend we had something bigger => Virtual Memory 2

Memory Access Cost

3

Virtual Memory

Paging (1) The position and function of the MMU 4

Paging (2)

The relation between virtual addresses and physical memory addres ses given by page table 5

Paging: Address Translation

CPU virtual address p d p f page table f physical address f d d memory 6

Page Tables (1)

Internal operation of MMU with 16 4 KB pages 7

Two-Level Page-Table Scheme

8

Translation Lookaside Buffers

• Translation on every memory access  fast must be • What to do? Caching, of course … – – Why does caching work? Temporal locality!

Same as normal memory cache – cache is smaller so can spend more $$ to make it faster • Cache for page table entries is called the Translation Lookaside Buffer (TLB) – – Typically fully associative Relatively small number of entries (e.g., 64 or 128 entries) • On every memory access, we look for the page  frame mapping in the TLB 9

Paging: Address Translation with TLB  TLB hit/ TLB miss CPU virtual address p d p/f TLB f f physical address f d d memory 10

• Page frames • (Virtual) Pages • Page fault • Hit /Miss • Page table • TLB

Terms

11

Page Replacement

• Highly motivated to find a good replacement policy – That is, when evicting a page, how do we choose the best victim in order to minimize the page fault rate?

• Is there an optimal replacement algorithm? If yes, what is it?

• Let’s look at an example: – Suppose we have 3 memory frames and are running a program that has the following reference pattern 7, 0, 1, 2, 0, 3, 0, 4, 2, 3 • Algorithms: Optimal, Not Recently Used(NRU), Second Chance, FIFO, LRU(Least Recently Used), Working Set & Clock, Nth Chance … 12

Optimal Page Replacement

• Suppose we know the access pattern in advance 7, 0, 1, 2, 0, 3, 0, 4, 2, 3 • Optimal algorithm is to replace the page that will not be used for the longest period of time • What’s the problem with this algorithm?

• Realistic policies try to predict future behavior on the basis of past behavior – Works because of locality 13

Not Recently Used Page Replacement Algorithm • Each page has Reference bit, Modified bit – bits are set when page is referenced, modified • Pages are classified 1.

not referenced, not modified 2.

3.

not referenced, modified referenced, not modified 4.

referenced, modified • NRU removes page at random – from lowest numbered non empty class 14

FIFO Page Replacement Algorithm

• Maintain a linked list of all pages – in order they came into memory • Page at beginning of list replaced (remove the last cell) • Disadvantage – Last cells may be often used 15

Least Recently Used (LRU)

• Assume pages used recently will used again soon – throw out page that has been unused for longest time • Must keep a linked list of pages – – most recently used at front, least at rear update this list every memory reference !!

• Alternatively keep counter in each page table entry – choose page with lowest value counter – periodically zero the counter 16

Segmentation (1)

• One-dimensional address space with growing tables • One table may bump into another 17

Segmentation (2)

Allows each table to grow or shrink, independently 18

Segmentation (3)

Comparison of paging and segmentation 19

Implementation of Pure Segmentation

(a)-(d) Development of checkerboarding (e) Removal of the checkerboarding by compaction 20

Segmentation with Paging: MULTICS (1) • Descriptor segment points to page tables • Segment descriptor – numbers are field lengths 21

Segmentation with Paging: MULTICS (2) A 34-bit MULTICS virtual address 22

Example Question1

• 4 page frames, page requests=0172327103 • Page Fault Rate??? For FIFO, LRU, (OPT)

FIFO

MMMMMH H HMH 0 1 7 2 3 2 7 1 0 3 0 1 7 2 3 3 3 3 0 0 0 1 7 2 2 2 2 3 3 0 1 7 7 7 7 2 2 0 1 1 1 1 7 7

LRU

MMMMMH H HMM 0 1 7 2 3 2 7 1 0 3 0 1 7 2 3 2 7 1 0 3 0 1 7 2 3 2 7 1 0 0 1 7 7 3 2 7 1 0 1 1 1 3 2 7 23

Example Question 2

Assume a process is divided into 4 equal-sized segments, and that the system builds a 16-entry page descriptor for each segment. Thus, the system has a combination of segmentation & paging. Assume also that the page size is 4KB.

a) b) What is the maximum size of each segment?

Assume that an element in physical location 0x21ABC (in hex) is accessed by a thread of this process. What segment, page & offset is the thread trying to access?

ANSWER: a) 16 * 4KB b) 4K = 2^12, therefore, ABC is for offset.

16=2^4,  page 1.

and “2” is for segment 24

Example Question 3

A computer has 32-bit virtual addresses & 4-KB pages. The program and data together fit in the lowest page (0-4095). The stack fits in the highest page. How many entries are needed in the page table if one-level paging is used? How many for 2-level page, with 10 bits in each part?

ANSWER: Twenty bits are used for the virtual page numbers, leaving 12 over for the offset.

This yields a 4-KB page. Twenty bits for the virtual page implies 2^20 pages.

25

Example Question 4

A computer whose processes have 1024 pages in their address spaces keeps its page tables in memory. The overhead required for reading a word from the page table is 5 ns. To reduce this overhead, the computer has a TLB, which hold 32 (virtual page->physical page frame) pairs, and can do a look up in 1 ns. What hit rate is needed to reduce the mean overhead to 2 ns?

ANSWER: The effective instruction time is 1

h + 5(1 − h), where h is the hit rate.

If we equate this formula with 2 and solve for h, we find that h must be at least 0.75.

26

Project 3

• • Nov 7, 2008 5:00 pm • Implement basic functions of Pthread Lib.

• Use

makecontext(), getcontext(), setcontext(), and swapcontext().

Useful Link:

http://www.opengroup.org/onlinepubs/00969 5399/functions/makecontext.html

27