OS, Base & Bounds, Virtual Memory Intro. Instructor: Steven Ho

Transcription

1 OS, Base & Bounds, Virtual Memory Intro Instructor: Steven Ho

2 Review Warehouse Scale Computers Supports many of the applications we have come to depend on Software must cope with failure, load variation, and latency/bandwidth limitations Hardware sensitive to cost and energy efficiency Request Level Parallelism High request volume, each largely independent Replication for better throughput, availability MapReduce Convenient data-level parallelism on large dataset across large number of machines Spark is a framework for executing MapReduce algorithms 2

3 Word Count in Spark s Python API // RDD: primary abstraction of a distributed collection of items file = sc.textfile( hdfs:// ) // Two kinds of operations: // Actions: RDD Value // Transformations: RDD RDD // e.g. flatmap, Map, reducebykey file.flatmap(lambda line: line.split()).map(lambda word: (word, 1)).reduceByKey(lambda a, b: a + b) 3

4 MapReduce Word Count Example Map phase: (doc name, doc contents) list(word, count) // I do I learn [( I,1),( do,1),( I,1),( learn,1)] map(key, value): for each word w in value: emit(w, 1) Reduce phase: (word, list(count)) (word, count_sum) // ( I, [1,1]) ( I,2) reduce(key, values): result = 0 for each v in values: result += v emit(key, result) 4

5 Word Count in Spark s Python API // RDD: primary abstraction of a distributed collection of items file = sc.textfile( hdfs:// ) // Two kinds of operations: // Actions: RDD Value // Transformations: RDD RDD // e.g. flatmap, Map, reducebykey file.flatmap(lambda line: line.split()).map(lambda word: (word, 1)).reduceByKey(lambda a, b: a + b) 5

6 MapReduce Word Count Example Map phase: (doc name, doc contents) list(word, count) // I do I learn [( I,1),( do,1),( I,1),( learn,1)] map(key, value): for each word w in value: emit(w, 1) Reduce phase: (word, list(count)) (word, count_sum) // ( I, [1,1]) ( I,2) reduce(key, values): result = 0 for each v in values: result += v emit(key, result) 6

7 Word Count in Spark s Python API // RDD: primary abstraction of a distributed collection of items file = sc.textfile( hdfs:// ) // Two kinds of operations: // Actions: RDD Value // Transformations: RDD RDD // e.g. flatmap, Map, reducebykey file.flatmap(lambda line: line.split()).map(lambda word: (word, 1)).reduceByKey(lambda a, b: a + b) 7

8 MapReduce Word Count Example Map phase: (doc name, doc contents) list(word, count) // I do I learn [( I,1),( do,1),( I,1),( learn,1)] map(key, value): for each word w in value: emit(w, 1) Reduce phase: (word, list(count)) (word, count_sum) // ( I, [1,1]) ( I,2) reduce(key, values): result = 0 for each v in values: result += v a emit(key, result) b 8

9 Agenda OS Intro Administrivia OS Boot Sequence and Operation Multiprogramming/time-sharing Introduction to Virtual Memory 9

10 CS61C so far C Programs #include <stdlib.h> Project 3 RISC-V Assembly.foo lw t0, 4(s0) addi t1, t0, 3 beq t1, t2, foo nop CPU Project 2 Labs int fib(int n) { return fib(n-1) + fib(n-2); } Project 1 Caches Memory 10

11 So how is this any different? Screen Keyboard Storage 11

12 Adding I/O C Programs #include <stdlib.h> Project 3 RISC-V Assembly CPU Project 2.foo lw t0, 4(s0) addi t1, t0, 3 beq t1, t2, foo nop Screen Caches int fib(int n) { return fib(n-1) + fib(n-2); } Project 1 Keyboard Storage I/O (Input/Output) Memory 12

13 Raspberry Pi ($40 on Amazon) Storage I/O (Micro SD Card) CPU+$s, etc. Memory Screen I/O (HDMI) Serial I/O (USB) Network I/O (Ethernet) 13

14 It s a real computer! 14

15 But wait That s not the same! When we run Venus, it only executes one program and then stops. When I switch on my computer, I get this: Yes, but that s just software! The Operating System (OS) 15

16 Well, just software The biggest piece of software on your machine? How many lines of code? These are 84 m guesstimates: ill ion s of cod line e! Codebases (in millions of lines of code). CC BY-NC 3.0 David McCandless

17 What does the OS do? One of the first things that runs when your computer starts (right after firmware/bootloader) Loads, runs and manages programs: Multiple programs at the same time (time-sharing) Isolate programs from each other (isolation) Multiplex resources between applications (e.g., devices) Services: File System, Network stack, etc. Finds and controls all the devices in the machine in a general way (using device drivers ) 17

18 Agenda OS Intro Administrivia OS Boot Sequence and Operation Multiprogramming/time-sharing Introduction to Virtual Memory 18

19 Administrivia Proj4 due on Friday (8/03) Project Party tonight, Soda 405/411, 4-6pm HW6 due tonight, HW7 released Guerilla Session on 540AB The final will be 8/ /2060! We re almost there! :D 19

20 Agenda OS Intro Administrivia OS Boot Sequence and Operation Multiprogramming/time-sharing Introduction to Virtual Memory 20

21 What happens at boot? When the computer switches on, it does the same as Venus: the CPU executes instructions from some start address (stored in Flash ROM) CPU Memory mapped 0x2000: addi t0, x0, 0x1000 lw t0, 4(s0) (Code to copy firmware into regular memory and jump into it) PC = 0x2000 (some default value) Address Space 21

22 What happens at boot? When the computer switches on, it does the same as Venus: the CPU executes instructions from some start address (stored in Flash ROM) 4. Init: Launch an application that waits for input in loop (e.g., Terminal/Desktop/ BIOS: Find a storage device and load first sector (block of data) 2. Bootloader (stored on, e.g., disk): Load the OS kernel from disk into a location in memory and jump into it. 3. OS Boot: Initialize services, drivers, etc. 22

23 Launching Applications Applications are called processes in most OSs. Created by another process calling into an OS routine (using a syscall, more details later). Depends on OS, but Linux uses fork (see OpenMP threads) to create a new process, and execve to load application. Loads executable file from disk (using the file system service) and puts instructions & data into memory (.text,.data sections), prepare stack and heap. Set argc and argv, jump into the main function. 23

24 Supervisor Mode If something goes wrong in an application, it can crash the entire machine. What about malware, etc.? The OS may need to enforce resource constraints to applications (e.g., access to devices). To protect the OS from the application, CPUs have a supervisor mode bit (also need isolation, more later). You can only access a subset of instructions and (physical) memory when not in supervisor mode (user mode). You can change out of supervisor mode using a special instruction, but not into it (unless there is an interrupt). 24

25 Syscalls How to switch back to OS? OS sets timer interrupt, when interrupts trigger, drop into supervisor mode. What if we want to call into an OS routine? (e.g., to read a file, launch a new process, send data, etc.) Need to perform a syscall: set up function arguments in registers, and then raise software interrupt OS will perform the operation and return to user mode This way, the OS can mediate access to all resources, including devices, the CPU itself, etc. 25

26 Syscalls in Venus Venus provides many simple syscalls using the ecall RISC-V instruction How to issue a syscall? Place the syscall number in a0 Place arguments to the syscall in the a1 register Issue the ecall instruction This is how your RISC-V code has been able to produce output all along ecall details depend on the ABI (Application Binary Interface) 26

27 Example Syscall Let s say we want to print an integer stored in s3: Print integer is syscall #1 li a0, 1 add a1, s3, x0 ecall 27

28 Venus s Environmental Calls 28

29 Agenda OS Intro Administrivia OS Boot Sequence and Operation Multiprogramming/time-sharing Introduction to Virtual Memory 29

30 Multiprogramming OS runs multiple applications at the same time. But not really (unless have a core per process) Switches between processes very quickly. This is called a context switch. When jumping into process, set timer interrupt. When it expires, store PC, registers, etc. (process state). Pick a different process to run and load its state. Set timer, change to user mode, jump to the new PC. Deciding what process to run is called scheduling. 30

31 Protection, Translation, Paging Supervisor mode does not fully isolate applications from each other or from the OS. Application could overwrite another application s memory. Remember the linker in CALL: application assumes that code is in certain location. How to prevent overlaps? May want to address more memory than we actually have (e.g., for sparse data structures). Solution: Virtual Memory. Give each process the illusion of a full memory address space that it has completely to itself. 31

32 Meet the Staff Emaan Trash TV Judge Judy Show Favorite Breaking Bad Plot Twist Best Bathroom 7th Floor Soda in Cal Best Study 1st Floor Moffitt Spot 6/21/2018 Sruthi Sean Bachelor in Paradise Dr. Phil Arrival Memento Stanley Hall Bottom Floor of Cory The VLSB Library Upper Floor of East Asian CS61C Su18 - Lecture 4 32

33 Agenda OS Intro Administrivia OS Boot Sequence and Operation Multiprogramming/time-sharing Introduction to Virtual Memory 33

34 Adding Disks to Hierarchy Use VM as a mechanism to connect memory and disk in the memory hierarchy 34

35 Memory Hierarchy Regs Instr Operands Earlier: Caches Upper Level Faster L1 Cache Blocks L2 Cache Blocks Now: Virtual Memory Memory Pages Disk Files Tape Larger Lower Level 35

36 Virtual Memory Goals Allow multiple processes to simultaneously occupy memory and provide protection: Don t let programs read/write each other s memory Give each program the illusion that it has its own private address space Suppose code starts at address 0x , then different processes each think their code resides at that same address! Each program must have a different view of memory 36

37 Virtual Memory Goals Next level in the memory hierarchy: Provides program with illusion of a very large main memory: Working set of pages reside in main memory - others reside on disk. Also allows OS to share memory, protect programs from each other Today, more important for protection vs. just another level of memory hierarchy Each process thinks it has all the memory to itself (Historically, it predates caches) 37

38 Virtual to Physical Address Translation Each program operates in its own virtual address space; ~only program running Each is protected from the other OS can decide where each goes in memory Hardware gives virtual physical mapping 38

39 Dynamic Address Translation Location-independent programs Protection prog1 Independent programs should not affect each other inadvertently need for a bound register Multiprogramming drives requirement for resident supervisor (OS) software to manage context switches between multiple programs prog2 Physical Memory Programming and storage management ease need for a base register OS 39

40 Modern Virtual Memory Systems Illusion of a large, private, uniform store Protection OS several programs, each with their private address space and one or more shared address spaces progi Demand Paging Provides the ability to run programs larger than the primary memory Primary Memory Swapping Store Hides differences in machine configurations The price is address translation on each memory reference VA mapping TLB PA 40

41 Analogy: Shopping at IKEA Furniture name like virtual address Actual warehouse location like physical address IKEA shopping list like page table you write mappings from furniture name to location On shopping list, whether the item is sold out like valid bit indicating in main memory vs. on disk Whether you can flop on a bed, draw on chalk board, turn lights off/onlike access rights 41

42 Simple Example: Base and Bound Reg 42

43 Simple Base and Bound Translation Segment Length Load X Logical Address Bounds Violation? + Physical Address Physical Memory Bound Register current segment Base Register Base Physical Address Program Address Space Base and bounds registers are visible/accessible only when processor is running in supervisor mode 43

44 Separate Areas for Program and Data (Scheme used on all Cray vector supercomputers prior to X1, 2002) + Program Bound Register Program Counter Program Base Register data segment Logical Address Data Base Register Program Address Space Bounds Violation? Physical Address Bounds Violation? Logical Address Main Memory Load X Data Bound Register Mem. Address Register program segment + Physical Address What is an advantage of this separation? 44

45 Bare 5-Stage Pipeline Physical P Address C Inst. Cache Physical Address D Decode E + M Physical Address Memory Controller Physical Address Data Cache W Physical Address Main Memory (DRAM) In a bare machine, the only kind of address is a physical address 45

46 Base and Bound Machine Prog. Bound Register Logical Address P C + Data Bound Register Bounds Violation? Inst. Cache D Decode E + M Physical Address Program Base Register Logical Address Bounds Violation? Data Cache + W Physical Address Physical Address Data Base Register Memory Controller Physical Address Physical Address Main Memory (DRAM) 46

47 Memory Fragmentation OS Space prog 1 16K prog 2 24K 24K prog 3 32K 24K Programs 4 & 5 start OS Space 16K BIG IDEA: 24K prog 2 USE RAM AS $$ prog 4 16K FOR 8K DISK prog 1 prog 3 32K prog 5 24K free Programs 2 & 5 end OS Space prog 1 16K 24K prog 4 16K 8K prog 3 32K 24K What As programs if we want start toand run end, process memory 6, and becomes we need 32K of fragmented. Therefore, if we ever need a large chunk space?! of memory, programs will need to be moved. 47

48 Question: What kind of cache should memory be? (A) Direct Mapped, write back/write allocate (B) Direct Mapped, write through/no write allocate (C) Fully Associative, write back/write allocate (D) Fully Associative, write thorugh/no write allocate 48

49 Question: What kind of cache should memory be? (A) Direct Mapped, write back/write allocate (B) Direct Mapped, write through/no write allocate (C) Fully Associative, write back/write allocate (D) Fully Associative, write through/no write allocate 49

50 Summary The role of the Operating System Booting a computer: BIOS, bootloader, OS boot, initialization Base and bounds for multiple processes Simple, but doesn t give us everything we want Virtual memory bridges memory and disk Provides illusion of independent address spaces to processes and protects them from each other 50