Parallel Programming
Message Passing (MPI)
Explicit Parallelism
●
Similar thing as multithreading for shared memory.
●
Explicit parallelism is more common with message
passing.
–
User has explicit control over processes.
–
Good: control can be used to performance benefit.
–
Bad: user has to deal with it.
Distributed Memory - Message Passing
mem1 mem2 mem3 memN
proc1 proc2 proc3 procN
network
Distributed Memory - Message Passing
●
A variable x, a pointer p, or an array a[] refer to
different memory locations, depending on the
processor.
●
We discuss message passing as a programming
model (can be on any hardware)
What does the user have to do?
●
This is what we said for shared memory:
–
Decide how to decompose the computation into
parallel parts.
–
Create (and destroy) processes to support that
decomposition.
–
Add synchronization to make sure
dependencies are covered.
●
Is the same true for message passing?
SOR Example
for some number of timesteps/iterations {
for (i=0; i<n; i++ )
for( j=0; j<n, j++ )
temp[i][j] = 0.25 *
( grid[i-1][j] + grid[i+1][j] +
grid[i][j-1] + grid[i][j+1] );
for( i=0; i<n; i++ )
for( j=0; j<n; j++ )
grid[i][j] = temp[i][j];
}
Shared Memory
grid 1 temp 1
2 2
3 3
4 4
proc1 proc2 proc3 procN
Message-Passing Data Distribution (only
middle processes)
grid grid
2 3
temp temp
2 3
proc2 proc3
Is this going to work?
Same code as we used for shared memory
for( i=from; i<to; i++ )
for( j=0; j<n; j++ )
temp[i][j] = 0.25*( grid[i-1][j] + grid[i+1][j]
+ grid[i][j-1] + grid[i][j+1]);
No, we need extra boundary elements for grid.
Data Distribution (only middle processes)
grid grid
2 3
temp temp
2 3
proc2 proc3
Is this going to work?
Same code as we used for shared memory
for( i=from; i<to; i++)
for( j=0; j<n; j++ )
temp[i][j] = 0.25*( grid[i-1][j] + grid[i+1][j]
+ grid[i][j-1] + grid[i][j+1]);
No, on the next iteration we need boundary
elements from our neighbors.
Data Communication (only middle processes)
grid grid
proc2 proc3
Is this now going to work?
Same code as we used for shared memory
for( i=from; i<to; i++ )
for( j=0; j<n; j++ )
temp[i][j] = 0.25*( grid[i-1][j] + grid[i+1][j]
+ grid[i][j-1] + grid[i][j+1]);
No, we need to translate the indices.
Index Translation
for( i=0; i<n/p; i++)
for( j=0; j<n; j++ )
temp[i][j] = 0.25*( grid[i-1][j] + grid[i+1][j]
+ grid[i][j-1] + grid[i][j+1]);
Remember, all variables are local.
Index Translation is Optional
●
Allocate the full arrays on each processor.
–
Leave indices alone.
●
Higher memory use.
–
Sometimes necessary.
What does the user need to do?
●
Divide up program in parallel parts.
●
Create and destroy processes to do above.
• Partition and distribute the data.
• Communicate data at the right time.
• (Sometimes) perform index translation.
●
Still need to do synchronization?
–
Sometimes, but many times goes hand in hand
with data communication.
Message Passing Systems
●
Provide process creation and destruction.
●
Provide message passing facilities (send
and receive, in various flavors) to distribute
and communicate data.
●
Provide additional synchronization facilities.
MPI (Message Passing Interface)
●
Is the defacto message passing standard.
●
Available on virtually all platforms.
●
Grew out of an earlier message passing
system, PVM, now outdated.
MPI Process Creation/Destruction
MPI_Init( int *argc, char ***argv )
Initiates a computation.
MPI_Finalize()
Terminates a computation.
MPI Process Identification
MPI_Comm_size( comm, &size )
Determines the number of processes.
MPI_Comm_rank( comm, &pid )
Pid is the process identifier of the caller.
MPI Basic Send
MPI_Send(buf, count, datatype, dest, tag, comm)
buf: address of send buffer
count: number of elements
datatype: data type of send buffer elements
dest: process id of destination process
tag: message tag (ignore for now)
comm: communicator (ignore for now)
MPI Basic Receive
MPI_Recv(buf, count, datatype, source, tag, comm, &status)
buf: address of receive buffer
count: size of receive buffer in elements
datatype: data type of receive buffer elements
source: source process id or MPI_ANY_SOURCE
tag and comm: ignore for now
status: status object
