Skip to content

MPI 调优

通信算法

MPI 实现一般都会为集合通信提供不同的算法，以适应不同的通信模式。这些算法在特定场景下可能会有不同的性能表现，因此需要根据实际情况进行调优。

一篇经典的例子是：GitHub: OpenMPI Bcast and Allreduce much slower than Intel-MPI (unrelated to EFA)。这位作者在 AWS 集群上测试了 OpenMPI 和 IntelMPI 各种算法的性能，给出了详细的测试结果。

Bcast algorithm comparison

可以看出性能差距还是比较明显的，因此算法调优是非常重要的。

  • 列出可用的集合通信算法：
ompi_info --param coll tuned --level 9
MCA coll tuned: parameter "coll_tuned_allreduce_algorithm" (current
                          value: "ignore", data source: default, level: 5
                          tuner/detail, type: int)
                          Which allreduce algorithm is used. Can be locked
                          down to any of: 0 ignore, 1 basic linear, 2
                          nonoverlapping (tuned reduce + tuned bcast), 3
                          recursive doubling, 4 ring, 5 segmented ring. Only
                          relevant if coll_tuned_use_dynamic_rules is true.
                          Valid values: 0:"ignore", 1:"basic_linear",
                          2:"nonoverlapping", 3:"recursive_doubling",
                          4:"ring", 5:"segmented_ring", 6:"rabenseifner"
  • 通过 MCA 参数设置算法，以 recursive doubling 为例：
--mca coll_tuned_use_dynamic_rules 1 \
--mca coll_tuned_allreduce_algorithm 3 \
--mca coll_tuned_allreduce_algorithm_segmentsize 4096 \
--mca coll_tuned_allreduce_algorithm_tree_fanout 4
  • 设置 Bcast 算法：
orterun \
--mca coll_tuned_use_dynamic_rules 1 \
--mca coll_tuned_bcast_algorithm $algo \