400 likes | 694 Views
Parallel Computing—Introduction to Message Passing Interface (MPI). Two Important Concepts. Two fundamental concepts of parallel programming are: Domain decomposition Functional decomposition. Domain Decomposition. Functional Decomposition. Message Passing Interface (MPI).
E N D
Parallel Computing—Introduction to Message Passing Interface (MPI)
Two Important Concepts • Two fundamental concepts of parallel programming are: • Domain decomposition • Functional decomposition
Message Passing Interface (MPI) • MPI is a standard (an interface or an API): • It defines a set of methods that are used by application developers to write their applications • MPI library implement these methods • MPI itself is not a library—it is a specification document that is followed! • MPI-1.2 is the most popular specification version • Reasons for popularity: • Software and hardware vendors were involved • Significant contribution from academia • MPICH served as an early reference implementation • MPI compilers are simply wrappers to widely used C and Fortran compilers • History: • The first draft specification was produced in 1993 • MPI-2.0, introduced in 1999, adds many new features to MPI • Bindings available to C, C++, and Fortran • MPI is a success story: • It is the mostly adopted programming paradigm of IBM Blue Gene systems • At least two production-quality MPI libraries: • MPICH2 (http://www-unix.mcs.anl.gov/mpi/mpich2/) • OpenMPI (http://open-mpi.org) • There’s even a Java library: • MPJ Express (http://mpj-express.org)
Message Passing Model • Message passing model allows processors to communicate by passing messages: • Processors do not share memory • Data transfer between processors required cooperative operations to be performed by each processor: • One processor sends the message while other receives the message
Memory CPU Distributed Memory Cluster Proc 1 Proc 2 Proc 0 message LAN Ethernet Myrinet Infiniband etc Proc 3 Proc 7 Proc 6 Proc 4 Proc 5
Writing “Hello World” MPI Program • MPI is very simple: • Initialize MPI environment: • MPI_Init(&argc,&argv); // C Code • MPI.Init(args); // Java Code • Send or receive message: • MPI_Send(..); // C Code • MPI.COMM_WORLD.Send(); // Java Code • Finalize MPI environment • MPI_Finalize(); // C Code • MPI.Finalize(); // Java Code
Hello World in C #include <stdio.h> #include <string.h> #include “mpi.h” .. // Initialize MPI MPI_Init(&argsc,&&argsv); // Find out the `id’ or `rank’ of current process MPI_Comm_Rank(MPI_COMM_WORLD,&my_rank); //get the rank // Get total number of processes MPI_Comm_Size(MPI_COMM_WORLD,&p); //get total processor // Print the rank of the process printf(“Hello World from process no %d”,my_rank); MPI_Finalize(); ..
Hello World in Java import java.util.*; import mpi.*; .. // Initialize MPI MPI.Init(args); // start up MPI // Get total number of processes and rank size = MPI.COMM_WORLD.Size(); rank = MPI.COMM_WORLD.Rank(); System.out.println(“Hello World <”+rank+”>”); MPI_Finalize(); ..
After Initialization import java.util.*; import mpi.*; .. // Initialize MPI MPI.Init(args); // start up MPI // Get total number of processes and rank size = MPI.COMM_WORLD.Size(); rank = MPI.COMM_WORLD.Rank(); ..
What is size? import java.util.*; import mpi.*; .. // Get total number of processes size = MPI.COMM_WORLD.Size(); .. • Total number of processes in a communicator: • The size of MPI.COMM_WORLD is 6
What is rank? import java.util.*; import mpi.*; .. // Get total number of processes rank = MPI.COMM_WORLD.Rank(); .. • The “unique” identify (id) of a process in a communicator: • Each of the six processes in MPI.COMM_WORLD has a distinct rank or id
Running “HelloWorld” in C • Write parallel code • Start MPICH2 daemon • Write machines file • Start the parallel job
Running “Hello World” in Java • The code is executed on a cluster called “Starbug”: • One head-node “holly” and eight compute-nodes • Steps: • Write machines files • Bootstrap MPJ Express (or any MPI library) runtime • Write parallel application • Compile and execute
Single Program Multiple Data (SPMD) Model import java.util.*; import mpi.*; public class HelloWorld { MPI.Init(args); // start up MPI size = MPI.COMM_WORLD.Size(); rank = MPI.COMM_WORLD.Rank(); if (rank == 0) { System.out.println(“I am Process 0”); } else if (rank == 1) { System.out.println(“I am Process 1”); } MPI.Finalize(); }
Single Program Multiple Data (SPMD) Model import java.util.*; import mpi.*; public class HelloWorld { MPI.Init(args); // start up MPI size = MPI.COMM_WORLD.Size(); rank = MPI.COMM_WORLD.Rank(); if (rank%2 == 0) { System.out.println(“I am an even process”); } else if (rank%2 == 1) { System.out.println(“I am an odd process”); } MPI.Finalize(); }
Point to Point Communication • The most fundamental facility provided by MPI • Basically “exchange messages between two processes”: • One process (source) sends message • The other process (destination) receives message
Point to Point Communication • It is possible to send message for each basic datatype: • Floats, Integers, Doubles … • Each message contains a “tag”—an identifier Tag1 Tag2
Integers Process 4 Tag COMM_WORLD Point to Point Communication Process 1 Process 2 Process 0 message Process 3 Process 7 Process 6 Process 4 Process 5
Blocking and Non-blocking • There are blocking and non-blocking version of send and receive methods • Blocking versions: • A process calls send() or recv(), these methods return when the message has been physically sent or received • Non-blocking versions: • A process calls isend() or irecv(), these methods return immediately • The user can check the status of message by calling test() or wait() • Note the “i” in isend() and irecv() • Non-blocking versions provide overlapping of computation and communication: • It also depends on the “quality” of the implementation
“Blocking” Sender Receiver send() recv() CPU waits CPU waits time “Non Blocking” Sender Receiver isend() irecv() CPU perform task CPU perform task time iwait() iwait() CPU waits CPU waits
Modes of Send • The MPI standard defines four modes of send: • Standard • Synchronous • Buffered • Ready
Synchronous Mode (Rendezvous Protocol used for large messages)
Performance Evaluation of Point to Point Communication • Normally ping pong benchmarks are used to calculate: • Latency: How long it takes to send N bytes from sender to receiver? • Throughput: How much bandwidth is achieved? • Latency is a useful measure for studying the performance of “small” messages • Throughput is a useful measure for studying the performance of “large” messages