Intel(R) MPI Benchmarks User Guide

Download PDF
ID 766171
Date 8/21/2024
Public
Document Table of Contents
Document Table of Contents x
Intel(R) MPI Benchmarks User Guide
Intel(R) MPI Benchmarks User Guide x
Introducing Intel(R) MPI Benchmarks Getting Started MPI-1 Benchmarks MPI-2 Benchmarks MPI-3 Benchmarks Multithreaded MPI-1 Benchmarks MPI-1 Benchmarks with GPU support Benchmark Methodology Notices and Disclaimers
Introducing Intel(R) MPI Benchmarks x
About this Document Getting Help and Support Related Information
Getting Started x
Memory and Disk Space Requirements Building Intel(R) MPI Benchmarks Running Intel(R) MPI Benchmarks Running Benchmarks in Multiple Mode
MPI-1 Benchmarks x
Classification of MPI-1 Benchmarks Single Transfer Benchmarks Parallel Transfer Collective Benchmarks IMB-P2P Benchmarks
Single Transfer Benchmarks x
PingPong, PingPongSpecificSource, PingPongAnySource PingPing, PingPingSpecificSource, PingPingAnySource
Parallel Transfer x
Sendrecv Exchange Uniband Biband
Collective Benchmarks x
Reduce Reduce_scatter Reduce_scatter_block Allreduce Allgather Allgatherv Scatter Scatterv Gather Gatherv Alltoall Bcast Barrier
IMB-P2P Benchmarks x
PingPong PingPing Unirandom Birandom Corandom
MPI-2 Benchmarks x
Naming Conventions IMB-MPI-2 Benchmark Classification MPI-2 Benchmark Modes IMB-EXT Benchmarks IMB-IO Blocking Benchmarks IMB-IO Non-blocking Benchmarks
IMB-EXT Benchmarks x
Unidir_Put Unidir_Get Bidir_Put Bidir_Get Accumulate Window
IMB-IO Blocking Benchmarks x
S_[ACTION]_indv S_[ACTION]_expl P_[ACTION]_indv P_[ACTION]_expl P_[ACTION]_shared P_[ACTION]_priv C_[ACTION]_indv C_[ACTION]_expl C_[ACTION]_shared Open_Close
MPI-3 Benchmarks x
IMB-NBC Benchmarks IMB-RMA Benchmarks
IMB-NBC Benchmarks x
Measuring Communication and Computation Overlap Measuring Pure Communication Time Iallgather Iallgather_pure Iallgatherv Iallgatherv_pure Iallreduce Iallreduce_pure Ialltoall Ialltoall_pure Ialltoallv Ialltoallv_pure Ibarrier Ibarrier_pure Ibcast Ibcast_pure Igather Igather_pure Igatherv Igatherv_pure Ireduce Ireduce_pure Ireduce_scatter Ireduce_scatter_pure Iscatter Iscatter_pure Iscatterv Iscatterv_pure
IMB-RMA Benchmarks x
Classification of IMB-RMA Benchmarks Accumulate All_get_all All_put_all Bidir_get Bidir_put Compare_and_swap Exchange_Get Exchange_Put Fetch_and_op Get_accumulate Get_all_local Get_local One_put_all One_get_all Put_all_local Put_local Truly_passive_put Unidir_get Unidir_put
Benchmark Methodology x
Control Flow Command-line Control Command-Line Control for IMB-P2P Benchmarks Command-Line Control for IMB-MT Benchmarks -thread_level {single|funneled|serialized|multiple|nompinit} Option -stride Option -warmup Option -repeat Option -barrier {on|off|special} Option -count Option -malloc_align Option -malloc_algo {serial|continuous|parallel} Option -check {on|off} Option -datatype {int|char} Option Parameters Controlling Intel(R) MPI Benchmarks Hard-Coded Settings Checking Results Output
Hard-Coded Settings x
Communicators, Active Processes Message/I-O Buffer Lengths Buffer Initialization Other Preparations for Benchmarking Warm-Up Phase (IMB-MPI1, IMB-EXT, IMB-NBC, and IMB-RMA) Synchronization Actual Benchmarking
Output x
Sample 1 - IMB-MPI1 PingPong Allreduce Sample 2 - IMB-MPI1 PingPing Allreduce Sample 3 - IMB-IO p_write_indv Sample 4 - IMB-EXT.exe
Intel(R) MPI Benchmarks User Guide

Command-Line Control for IMB-MT Benchmarks

Level Two Title