Synchronization, Hakim Weatherspoon CS 3410, Spring 2013 Computer Science Cornell University

Download this Presentation

xiaoxin4564 PPT making software
0

Presentation Transcript

  • 1.Synchronization Hakim Weatherspoon CS 3410, Spring 2013 Computer Science Cornell University P&H Chapter 2.11 and 5.8
  • 2.Big Picture: Parallelism and Synchronization How do I take advantage of multiple processors; parallelism? How do I write (correct) parallel programs, cache cohrency and synchronization? What primitives do I need to implement correct parallel programs?
  • 3.Goals for Today Understand Cache Coherency Problem Define Cache coherency problem Synchronizing parallel programs Atomic Instructions HW support for synchronization How to write parallel programs Threads and processes Critical sections, race conditions, and mutexes
  • 4.Big Picture: Parallelism and Synchronization
  • 5.Next Goal: Parallelism and Synchronization Cache Coherency Problem: What happens when to two or more processors cache shared data?
  • 6.Next Goal: Parallelism and Synchronization Cache Coherency Problem: What happens when to two or more processors cache shared data? i.e. the view of memory held by two different processors is through their individual caches. As a result, processors can see different (incoherent) values to the same memory location.
  • 7.Big Picture: Parallelism and Synchronization Each processor core has its own L1 cache
  • 8.Big Picture: Parallelism and Synchronization Each processor core has its own L1 cache
  • 9.Big Picture: Parallelism and Synchronization Each processor core has its own L1 cache Core0 Cache Memory I/O Interconnect Core1 Cache Core3 Cache Core2 Cache
  • 10.Shared Memory Multiprocessors Shared Memory Multiprocessor (SMP) Typical (today): 2 – 4 processor dies, 2 – 8 cores each HW provides single physical address space for all processors Assume physical addresses (ignore virtual memory) Assume uniform memory access (ignore NUMA) Core0 Cache Memory I/O Interconnect Core1 Cache Core3 Cache Core2 Cache
  • 11.Shared Memory Multiprocessors Shared Memory Multiprocessor (SMP) Typical (today): 2 – 4 processor dies, 2 – 8 cores each HW provides single physical address space for all processors Assume physical addresses (ignore virtual memory) Assume uniform memory access (ignore NUMA) ... Core0 Cache Memory I/O Interconnect Core1 Cache CoreN Cache ... ...
  • 12.Cache Coherency Problem ... Core0 Cache Memory I/O Interconnect Core1 Cache CoreN Cache ... ... Thread A (on Core0) Thread B (on Core1)for(int i = 0, i < 5; i++) { for(int j = 0; j < 5; j++) { x = x + 1; x = x + 1;} } What will the value of x be after both loops finish?
  • 13.Cache Coherency Problem Thread A (on Core0) Thread B (on Core1)for(int i = 0, i < 5; i++) { for(int j = 0; j < 5; j++) { x = x + 1; x = x + 1;} } What will the value of x be after both loops finish? 6 8 10 All of the above None of the above
  • 14.Cache Coherency Problem ... Core0 Cache Memory I/O Interconnect Core1 Cache CoreN Cache ... ... Thread A (on Core0) Thread B (on Core1)for(int i = 0, i < 5; i++) { for(int j = 0; j < 5; j++) { LW $t0, addr(x) LW $t0, addr(x) ADDIU $t0, $t0, 1 ADDIU $t0, $t0, 1 SW $t0, addr(x) SW $t0, addr(x)} } $t0=0 $t0=1 x=1 $t0=0 $t0=1 x=1 x should be greater than 1 after both threads loop at least once!
  • 15.Cache Coherence Problem Suppose two CPU cores share a physical address space Write-through caches ... Core0 Cache Memory I/O Interconnect Core1 Cache CoreN Cache ... ...
  • 16.Coherence Defined Cache coherence defined... Informal: Reads return most recently written value Formal: For concurrent processes P1 and P2 P writes X before P reads X (with no intervening writes) read returns written value P1 writes X before P2 reads X  read returns written value P1 writes X and P2 writes X all processors see writes in the same order all see the same final value for X Aka write serialization
  • 17.Cache Coherence Protocols Operations performed by caches in multiprocessors to ensure coherence Migration of data to local caches Reduces bandwidth for shared memory Replication of read-shared data Reduces contention for access Snooping protocols Each cache monitors bus reads/writes
  • 18.Snooping Snooping for Hardware Cache Coherence All caches monitor bus and all other caches Bus read: respond if you have dirty data Bus write: update/invalidate your copy of data Snoop Snoop Snoop ... Core0 Cache Memory I/O Interconnect ... ... Snoop Core1 Cache Snoop CoreN Cache Snoop
  • 19.Invalidating Snooping Protocols Cache gets exclusive access to a block when it is to be written Broadcasts an invalidate message on the bus Subsequent read in another cache misses Owning cache supplies updated value
  • 20.Writing Write-back policies for bandwidth Write-invalidate coherence policy First invalidate all other copies of data Then write it in cache line Anybody else can read it Permits one writer, multiple readers In reality: many coherence protocols Snooping doesn’t scale Directory-based protocols Caches and memory record sharing status of blocks in a directory
  • 21.Takeaway Informally, Cache Coherency requires that reads return most recently written value. With multiprocessors maintaining cache coherency can be difficult and requires cache coherency protocols like Snooping cache coherency protocols.
  • 22.Next Goal: Synchronization Is cache coherency sufficient? i.e. Is cache coherency (what values are read) sufficient to maintain consistency (when a written value will be returned to a read). Both coherency and consistency are required to maintain consistency in shared memory programs.
  • 23.Is Cache Coherency Sufficient? ... Core0 Cache Memory I/O Interconnect Core1 Cache CoreN Cache ... ... Thread A (on Core0) Thread B (on Core1)for(int i = 0, i < 5; i++) { for(int j = 0; j < 5; j++) { LW $t0, addr(x) LW $t0, addr(x) ADDIU $t0, $t0, 1 ADDIU $t0, $t0, 1 SW $t0, addr(x) SW $t0, addr(x)} } Very expensive and difficult to maintain consistency
  • 24.Synchronization Two processors sharing an area of memory P1 writes, then P2 reads Data race if P1 and P2 don’t synchronize Result depends of order of accesses Hardware support required Atomic read/write memory operation No other access to the location allowed between the read and write Could be a single instruction E.g., atomic swap of register ↔ memory (e.g. ATS, BTS; x86) Or an atomic pair of instructions (e.g. LL and SC; MIPS)
  • 25.Synchronization in MIPS Load linked: LL rt, offset(rs) Store conditional: SC rt, offset(rs) Succeeds if location not changed since the LL Returns 1 in rt Fails if location is changed Returns 0 in rt Example: atomic swap (to test/set lock variable) try: MOVE $t0,$s4 ;copy exchange value LL $t1,0($s1) ;load linked SC $t0,0($s1) ;store conditional BEQZ $t0,try ;branch store fails MOVE $s4,$t1 ;put load value in $s4 Any time a processor intervenes and modifies the value in memory between the LL and SC instruction, the SC returns 0 in $t0, causing the code to try again.
  • 26.Mutex from LL and SC Linked load / Store Conditional m = 0; // m=0 means lock is free; otherwise, if m=1, then lock locked mutex_lock(int *m) { while(test_and_test(m)){} } int test_and_set(int *m) { old = *m; *m = 1; return old; } LL Atomic SC
  • 27.Mutex from LL and SC Linked load / Store Conditional m =0; mutex_lock(int *m) { while(test_and_test(m)){} } int test_and_set(int *m) { LI $t0, 1 LL $t1, 0($a0) SC $t0, 0($a0) MOVE $v0, $t1 } BEQZ $t0, try try:
  • 28.Mutex from LL and SC Linked load / Store Conditional mutex_lock(int *m) { while(test_and_test(m)){} } int test_and_set(int *m) { try: LI $t0, 1 LL $t1, 0($a0) SC $t0, 0($a0) BEQZ $t0, try MOVE $v0, $t1 }
  • 29.Mutex from LL and SC Linked load / Store Conditional mutex_lock(int *m) { test_and_set: LI $t0, 1 LL $t1, 0($a0) BNEZ $t1, test_and_set SC $t0, 0($a0) BEQZ $t0, test_and_set } mutex_unlock(int *m) { *m = 0; }
  • 30.Mutex from LL and SC Linked load / Store Conditional mutex_lock(int *m) { test_and_set: LI $t0, 1 LL $t1, 0($a0) BNEZ $t1, test_and_set SC $t0, 0($a0) BEQZ $t0, test_and_set } mutex_unlock(int *m) { SW $zero, 0($a0) } This is called a Spin lock Aka spin waiting
  • 31.Alternative Atomic Instructions Other atomic hardware primitives - test and set (x86) - atomic increment (x86) - bus lock prefix (x86) - compare and exchange (x86, ARM deprecated) - linked load / store conditional (MIPS, ARM, PowerPC, DEC Alpha, …)
  • 32.Now we can write parallel and correct programs Thread A                                  Thread Bfor(int i = 0, i < 5; i++) {      for(int j = 0; j < 5; j++) {   x = x + 1;                                 x = x + 1; } } mutex_lock(m); mutex_lock(m); mutex_unlock(m); mutex_unlock(m);
  • 33.Takeaway Informally, Cache Coherency requires that reads return most recently written value. With multiprocessors maintaining cache coherency can be difficult and requires cache coherency protocols like Snooping cache coherency protocols. Cache coherency controls what values are read, but may be insufficient or very expensive to maintain consistency (when a written value will be returned to a read). We need synchronization primitives to more efficiently implement parallel and correct programs.
  • 34.Next Goal How do we write parallel programs?
  • 35.Processes How do we cope with lots of activity? Simplicity? Separation into processes Reliability? Isolation Speed? Program-level parallelism gcc emacs nfsd lpr ls www emacs nfsd lpr ls www OS OS
  • 36.Process and Program Process OS abstraction of a running computation The unit of execution The unit of scheduling Execution state+ address space From process perspective a virtual CPU some virtual memory a virtual keyboard, screen, … Program “Blueprint” for a process Passive entity (bits on disk) Code + static data Header Code Initialized data BSS Symbol table Line numbers Ext. refs
  • 37.Code Initialized data BSS Heap Stack DLL’s mapped segments Process and Program Header Code Initialized data BSS Symbol table Line numbers Ext. refs Program Process
  • 38.Role of the OS Role of the OS Context Switching Provides illusion that every process owns a CPU Virtual Memory Provides illusion that process owns some memory Device drivers & system calls Provides illusion that process owns a keyboard, … To do: How to start a process? How do processes communicate / coordinate?
  • 39.How to create a process? Q: How to create a process? A: Double click After boot, OS starts the first process …which in turn creates other processes parent / child  the process tree
  • 40.pstree example $ pstree | view - init-+-NetworkManager-+-dhclient |-apache2 |-chrome-+-chrome | `-chrome |-chrome---chrome |-clementine |-clock-applet |-cron |-cupsd |-firefox---run-mozilla.sh---firefox-bin-+-plugin-cont |-gnome-screensaver |-grep |-in.tftpd |-ntpd `-sshd---sshd---sshd---bash-+-gcc---gcc---cc1 |-pstree |-vim `-view
  • 41.Processes Under UNIX Init is a special case. For others… Q: How does parent process create child process? A: fork() system call Wait. what? int fork() returns TWICE!
  • 42.Example main(int ac, char **av) { int x = getpid(); // get current process ID from OS char *hi = av[1]; // get greeting from command line printf(“I’m process %d\n”, x); int id = fork(); if (id == 0) printf(“%s from %d\n”, hi, getpid()); else printf(“%s from %d, child is %d\n”, hi, getpid(), id); } $ gcc -o strange strange.c $ ./strange “Hey” I’m process 23511 Hey from 23512 Hey from 23511, child is 23512
  • 43.Inter-process Communication Parent can pass information to child In fact, all parent data is passed to child But isolated after (C-O-W ensures changes are invisible) Q: How to continue communicating? A: Invent OS “IPC channels” : send(msg), recv(), …
  • 44.Inter-process Communication Parent can pass information to child In fact, all parent data is passed to child But isolated after (C-O-W ensures changes are invisible) Q: How to continue communicating? A: Shared (Virtual) Memory!
  • 45.Processes and Threads
  • 46.Processes are heavyweight Parallel programming with processes: They share almost everything code, shared mem, open files, filesystem privileges, … Pagetables will be almost identical Differences: PC, registers, stack Recall: process = execution context + address space
  • 47.Processes and Threads Process OS abstraction of a running computation The unit of execution The unit of scheduling Execution state+ address space From process perspective a virtual CPU some virtual memory a virtual keyboard, screen, … Thread OS abstraction of a single thread of control The unit of scheduling Lives in one single process From thread perspective one virtual CPU core on a virtual multi-core machine
  • 48.Multithreaded Processes
  • 49.Threads #include int counter = 0; void PrintHello(int arg) { printf(“I’m thread %d, counter is %d\n”, arg, counter++); ... do some work ... pthread_exit(NULL); } int main () { for (t = 0; t < 4; t++) { printf(“in main: creating thread %d\n", t); pthread_create(NULL, NULL, PrintHello, t); } pthread_exit(NULL); }
  • 50.Threads versus Fork in main: creating thread 0 I’m thread 0, counter is 0 in main: creating thread 1 I’m thread 1, counter is 1 in main: creating thread 2 in main: creating thread 3 I’m thread 3, counter is 2 I’m thread 2, counter is 3 If processes?
  • 51.Example Multi-Threaded Program Example: Apache web server void main() {setup(); while (c = accept_connection()) { req = read_request(c); hits[req]++; send_response(c, req); } cleanup(); }
  • 52.Race Conditions Example: Apache web server Each client request handled by a separate thread (in parallel) Some shared state: hit counter, ... (look familiar?) Timing-dependent failure  race condition hard to reproduce  hard to debug Thread 52 ... hits = hits + 1; ... Thread 205 ... hits = hits + 1; ... Thread 52 read hits addi write hits Thread 205 read hits addi write hits
  • 53.Programming with threads Within a thread: execution is sequential Between threads? No ordering or timing guarantees Might even run on different cores at the same time Problem: hard to program, hard to reason about Behavior can depend on subtle timing differences Bugs may be impossible to reproduce Cache coherency isn’t sufficient… Need explicit synchronization to make sense of concurrency!
  • 54.Race conditions Race Condition Timing-dependent error when accessing shared state Depends on scheduling happenstance… e.g. who wins “race” to the store instruction? Concurrent Program Correctness =all possible schedules are safe Must consider every possible permutation In other words… … the scheduler is your adversary
  • 55.Critical sections What if we can designate parts of the execution as critical sections Rule: only one thread can be “inside” Thread 52 read hits addi write hits Thread 205 read hits addi write hits
  • 56.Mutexes Q: How to implement critical section in code? A: Lots of approaches…. Mutual Exclusion Lock (mutex) acquire(m): wait till it becomes free, then lock it release(m): unlock it apache_got_hit() { pthread_mutex_lock(m); hits = hits + 1; pthread_mutex_unlock(m) }
  • 57.Takeaway Informally, Cache Coherency requires that reads return most recently written value. With multiprocessors maintaining cache coherency can be difficult and requires cache coherency protocols like Snooping cache coherency protocols. Cache coherency controls what values are read, but may be insufficient or very expensive to maintain consistency (when a written value will be returned to a read). We need synchronization primitives to more efficiently implement parallel and correct programs. Processes and Threads are the abstraction that we use to write parallel programs? Fork and Joint and Interprocesses communication (IPC) can be used to coordinate processes. Threads are used to coordinate use of shared memory within a process.
  • 58.Next time Next time. Higher level synchronization primitives (other abstractions to implement a critical section beyond mutexes)?
  • 59.Administrivia Project3 due next week, Monday, April 22nd Design Doc due yesterday, Monday, April 15th Games night Friday, April 26th, 5-7pm. Location: B17 Upson Homework4 is due tomorrow, Wednesday, April 17th Work alone Question1 on Virtual Memory is pre-lab question for Lab4 Prelim3 is next week, Thursday, April 25th Time and Location: 7:30pm in Phillips 101 and Upson B17 Old prelims are online in CMS
  • 60.Administrivia Next three weeks Week 12 (Apr 15): Project3 design doc due and HW4 due Week 13 (Apr 22): Project3 due and Prelim3 Week 14 (Apr 29): Project4 handout Final Project for class Week 15 (May 6): Project4 design doc due Week 16 (May 13): Project4 due
Related
Latest Release
SlideHTML5 Home Explore About Us
Support Contact Us FAQ Blog
Legal Terms DMCA Cookies Policy Privacy Terms of Service
SlideHtml5

Free online presentation maker for an exceptional way of presenting. Make interactive online presentations that impress, inspire, and connect.

© 2021 SlideHTML5. All rights reserved

© 2021 SlideHTML5.

All rights reserved