TY - GEN
T1 - Test suite for evaluating performance of MPI implementations that support MPI_THREAD_MULTIPLE
AU - Thakur, Rajeev
AU - Gropp, William
PY - 2007
Y1 - 2007
N2 - MPI implementations that support the highest level of thread safety for user programs, MPI_THREAD_MLTIPLE, are becoming widely available. Users often expect that different threads can execute independently and that the MPI implementation can provide the necessary level of thread safety with only a small overhead. The MPI Standard, however, requires only that no MPI call in one thread block MPI calls in other threads; it makes no performance guarantees. Therefore, some way of measuring an implementation's performance is needed. In this paper, we propose a number of performance tests that are motivated by typical application scenarios. These tests cover the overhead of providing the MPI_THREAD_MLTIPLE level of thread safety for user programs, the amount of concurrency in different threads making MPI calls, the ability to overlap communication with computation, and other features. We present performance results with this test suite on several platforms (Linux cluster, Sun and IBM SMPs) and MPI implementations (MPICH2, Open MPI, IBM, and Sun).
AB - MPI implementations that support the highest level of thread safety for user programs, MPI_THREAD_MLTIPLE, are becoming widely available. Users often expect that different threads can execute independently and that the MPI implementation can provide the necessary level of thread safety with only a small overhead. The MPI Standard, however, requires only that no MPI call in one thread block MPI calls in other threads; it makes no performance guarantees. Therefore, some way of measuring an implementation's performance is needed. In this paper, we propose a number of performance tests that are motivated by typical application scenarios. These tests cover the overhead of providing the MPI_THREAD_MLTIPLE level of thread safety for user programs, the amount of concurrency in different threads making MPI calls, the ability to overlap communication with computation, and other features. We present performance results with this test suite on several platforms (Linux cluster, Sun and IBM SMPs) and MPI implementations (MPICH2, Open MPI, IBM, and Sun).
UR - http://www.scopus.com/inward/record.url?scp=38449109012&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=38449109012&partnerID=8YFLogxK
U2 - 10.1007/978-3-540-75416-9_13
DO - 10.1007/978-3-540-75416-9_13
M3 - Conference contribution
AN - SCOPUS:38449109012
SN - 9783540754152
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 46
EP - 55
BT - Recent Advances in Parallel Virtual Machine and Message Passing Interface - 14th European PVM/MPI Users' Group Meeting, Proceedings
PB - Springer
T2 - 14th European PVM/MPI Users' Group Meeting on Parallel Virtual Machine and Message Passing Interface
Y2 - 30 September 2007 through 3 October 2007
ER -