National Center for Supercomputing Applications MPI for better scalability & application performance Byoung-Do Kim, Ph.D. National Center for Supercomputing Applications University of Illinois at Urbana-Champaign Seungdo Hong Dept. Of Mechanical Engineering Pusan National University, Pusan, Korea

National Center for Supercomputing Applications Outline MPI basic MPI collective communication MPI datatype Data parallelism: domain decomposition Algorithm Implementation Examples Conclusion

National Center for Supercomputing Applications MPI Basics MPI_Init starts up the MPI runtime environment at the beginning of a run. MPI_Finalize shuts down the MPI runtime environment at the end of a run. MPI_Comm_size gets the number of processes in a run, N p (typically called just after MPI_Init ). MPI_Comm_rank gets the process ID that the current process uses, which is between 0 and N p -1 inclusive (typically called just after MPI_Init ).

National Center for Supercomputing Applications PROGRAM my_mpi_program IMPLICIT NONE INCLUDE "mpif.h" [other includes] INTEGER :: my_rank, num_procs, mpi_error_code [other declarations] CALL MPI_Init(mpi_error_code) !! Start up MPI CALL MPI_Comm_Rank(my_rank, mpi_error_code) CALL MPI_Comm_size(num_procs, mpi_error_code) [actual work goes here] CALL MPI_Finalize(mpi_error_code) !! Shut down MPI END PROGRAM my_mpi_program MPI example code in Fortran

National Center for Supercomputing Applications MPI example code in C #include #include "mpi.h" [other includes] int main (int argc, char* argv[]) { /* main */ int my_rank, num_procs, mpi_error; [other declarations] MPI_Init(&argc, &argv); /* Start up MPI */ MPI_Comm_rank(MPI_COMM_WORLD, &my_rank); MPI_Comm_size(MPI_COMM_WORLD, &num_procs); [actual work goes here] MPI_Finalize(); /* Shut down MPI */ } /* main */

National Center for Supercomputing Applications How an MPI Run Works Every process gets a copy of the executable: Single Program, Multiple Data (SPMD). They all start executing it. Each looks at its own rank to determine which part of the problem to work on. Each process works completely independently of the other processes, except when communicating.

National Center for Supercomputing Applications Send & Receive MPI_SEND(buf,count,datatype,dest,tag,comm) MPI_SEND(buf,count,datatype,source,tag,comm,status) When MPI sends a message, it doesn’t just send the contents; it also sends an “envelope” describing the contents: Buf: initial address of send buffer Count: number of entries to send Data type: datatype of each entry Source: rank of sending process Dest: rank of process to receive Tag (message ID) Comm: communicator (e.g., MPI_COMM_WORLD )

National Center for Supercomputing Applications MPI_SENDRECV MPI_SendRecv(sendbuf,sendcount,send type,dest,sendtag,recvbuf,recvcoun t,recvtype,source,recvtag,comm,sta tus) Useful for communications patterns where each node both sends and receives messages. Executes a blocking send & receive operation Both function use the same communicator, but have distinct tag argument

National Center for Supercomputing Applications Collective Communication Broadcast ( MPI_Bcast ) –A single proc sends the same data to every proc Reduction ( MPI_Reduce ) –All the procs contribute data that is combined using a binary operation (min, max, sum, etc.): One proc obtains the final answer Allreduce ( MPI_Allreduce ) –Same as MPI_Reduce, but every proc contains the final answer Gather ( MPI_Gather ) –Collect the data from every proc and store the data on proc root Scatter ( MPI_Scatter ) –Split the data on proc root into np segment

National Center for Supercomputing Applications

MPI Datatype CFortran 90 charMPI_CHARCHARACTERMPI_CHARACTER intMPI_INTINTEGERMPI_INTEGER floatMPI_FLOATREALMPI_REAL doubleMPI_DOUBLEDOUBLE PRECISION MPI_DOUBLE_PRECISION MPI supports several other data types, but most are variations of these, and probably these are all you’ll use.

National Center for Supercomputing Applications Data packaging Use MPI derived datatype constructor if data to be transmitted consists of a subset of the entries in an array MPI_type_contiguous : builds a derived type whose elements are contiguous entries in an array MPI_Type_vector : for equally spaced entries MPI_Type_indexed : for binary entries of an array

National Center for Supercomputing Applications MPI_Type_Vector MPI_TYPE_VECTOR(count,blocklength,stride, oldtype, newtype) IN countnumber of blocks (int) IN blocklengthnumber of elements in each block (int) INstride spacing between start of each block, measured as number of elements (int) IN oldtypeold datatype (handle) OUT newtype new datatype (handle) oldtypeblocklength stride = = count

National Center for Supercomputing Applications Virtual Topology MPI_cart_creat(comm_old,ndims,dims,peri od,reorder,comm,cart) –Describe Cartesian structure of arbitrary dimension –Create a new communicator, contains information on the structure of the Cartesian topology. –Returns a handle to a new communicator with the topology information. MPI_cart_rank(comm,coords,rank) MPI_cart_coords(comm,rank,maxdims,coord s) Mpi_cart_shift(comm,direction,disp,rank _source,rank_dest)

National Center for Supercomputing Applications Application: 3-D Heat Conduction Problem Solving heat conduction equation by TDMA (Tri-Diagonal Matrix Algorithm)

National Center for Supercomputing Applications Domain Decomposition Data parallelization: Extensibility, Portability Divide computational domain into many sub-domains based on number of processors Solves the same problem on the sub-domians, need to transfer the b.c. information of overlapping boundary area Requires communication between the subdomains in every time step Major parallelization method in CFD applications In order to get a good scalability, need to implement algorithms carefully.

National Center for Supercomputing Applications 1-D decomposition ! ! MPI Cartesian Coordinate Communicator ! CALL MPI_CART_CREATE (MPI_COMM_WORLD, NDIMS, DIMS, PERIODIC,REORDER,CommZ,ierr) CALL MPI_COMM_RANK (CommZ,myPE,ierr) CALL MPI_CART_COORDS (CommZ,myPE, NDIMS,CRDS,ierr) CALL MPI_CART_SHIFT (CommZ,0,1,PEb,PEt,ierr) ! ! MPI Datatype creation ! CALL MPI_TYPE_CONTIGUOUS (Nx*Ny,MPI_DOUBLE_PRECISION,XY_p,ierr) CALL MPI_TYPE_COMMIT(XY_p,ierr)

National Center for Supercomputing Applications 2-D decomposition CALL MPI_CART_CREATE (MPI_COMM_WORLD, NDIMS, DIMS, PERIODIC,REORDER,CommXY,ierr) CALL MPI_COMM_RANK (CommXY,myPE,ierr) CALL MPI_CART_COORDS (CommXY,myPE,NDIMS,CRDS,ierr) CALL MPI_CART_SHIFT (CommXY,1,1,PEw,PEe,ierr) CALL MPI_CART_SHIFT (CommXY,0,1,PEs,PEn,ierr) ! ! MPI Datatype creation ! CALL MPI_TYPE_VECTOR (cnt_yz,block_yz,strd_yz,MPI_DOUBLE_PRECISION, YZ_p,ierr) CALL MPI_TYPE_COMMIT (YZ_p,ierr) CALL MPI_TYPE_VECTOR (cnt_xz,block_xz,strd_xz,MPI_DOUBLE_PRECISION, XZ_p,ierr) CALL MPI_TYE_COMMIT (XZ_p,ierr)

National Center for Supercomputing Applications 3-D decomposition CALL MPI_CART_CREATE (MPI_COMM_WORLD,…,commXYZ,ierr) CALL MPI_COMM_RANK (CommXYZ,myPE,ierr) CALL MPI_CART_COORDS (CommXYZ,myPE,NDIMS,CRDS,ierr) CALL MPI_CART_SHIFT (CommXYZ,2,1,PEw,PEe,ierr) CALL MPI_CART_SHIFT (CommXYZ,1,1,PEs,PEn,ierr) CALL MPI_CART_SHIFT (CommXYZ,0,1,PEb,PEt,ierr) ! CALL MPI_TYPE_VECTOR (cnt_yz,block_yz,strd_yz, MPI_DOUBLE_PRECISION,YZ_p,ierr) CALL_MPI_TYPE_COMMIT (YZ_p,ierr) CALL MPI_TYPE_VECTOR (cnt_xz,block_xz,strd_xz, MPI_DOUBLE_PRECISION,XZ_p,ierr) CALL MPI_TYEP_COMMIT (XZ_p,ierr) CALL MPI_TYPE_CONTIGUOUS (cnt_xy, MPI_DOUBLE_PRECISION,XY_p,ierr) CALL MPI_TYPE_COMMIT (XY_p,ierr)

National Center for Supercomputing Applications Scalability : 1-D Good Scalability up to small number of processors (16) After choke point, communication overhead becomes dominant. Performance degrade with large number of processors

National Center for Supercomputing Applications Scalability: 2-D Strong Scalability up to large number of processors Actual runtime larger than 1- D case in the case of small number of processors Sweep direction of TDMA solver affects the parallel performance due to communication overhead

National Center for Supercomputing Applications Scalability: 3-D Superior scalability behavior over the other two cases No choke point observed up to 512 processors Communication overhead ignorable compared to total runtime.

National Center for Supercomputing Applications SpeedUps

National Center for Supercomputing Applications Superlinear Speedup of 3-D Parallel Case Benefit from Intel Itanium chip architecture (Large L3 cache, bypassing L1 for floating point calculation) Small message size per communication due to good scalability

National Center for Supercomputing Applications Conclusion 1-D decomposition is OK for small application size, but has communication overhead problem when the size increases 2-D shows strong scaling behavior, but need to be careful when apply due to influences from numerical solvers’ characteristics. 3-D demonstrates superior scalability over the other two, have superlinear problem due to hardware architecture. There is no one-size-fit-all magic solution. In order to get the best scalability & application performance, the MPI algorithm, application characteristics, and hardware architectures are in harmony for the best possible solution.