macOS Sequoia系统下Open-MPI与MPICH运行简单测试脚本均失败的问题求助

各位好，我是刚接触MPI的新手，最近在我的macOS Sequoia 15.5（Darwin 24.5.0，Intel x86_64处理器）上运行Open-MPI和MPICH时遇到了棘手的问题——不管用哪种方式安装，运行最简单的MPI测试程序都会崩溃，我怀疑是编译器或系统库出了大问题，试过各种方法都没解决，实在没办法了来求助大家。

我的环境信息 #

• 系统：macOS Sequoia 15.5（Darwin 24.5.0）
• 处理器：Intel x86_64架构
• MPI版本：试过Open-MPI和MPICH的多个版本，均通过以下方式安装

尝试过的安装方式 #

我已经把能想到的MPI安装方法都试了一遍，但结果完全一致：

• Homebrew常规二进制安装：brew install open-mpi / brew install mpich
• Homebrew源码编译安装：brew install --build-from-source open-mpi / brew install --build-from-source mpich
• 指定Homebrew自带的编译器进行Homebrew安装
• 手动从官方源码包编译安装
• 通过PETSc自动下载编译安装（这个方法我之前了解到很可靠，使用--download-openmpi=yes或--download-mpich=yes参数）

我的测试代码 #

为了排除业务逻辑问题，我写了一段最基础的MPI测试代码：

#include <mpi.h>
#include <stdio.h>

int main(int argc, char** argv) {
    int rank, ierr;
    ierr = MPI_Init(&argc, &argv);
    printf("ierr %d\n", ierr);
    MPI_Comm_rank(MPI_COMM_WORLD, &rank);
    printf("Hello from rank %d\n", rank);
    MPI_Finalize();
    return 0;
}

执行步骤与错误输出 #

编译过程完全正常，但执行多进程任务时就会触发段错误，以下是完整的终端输出：

ccroberts@lumos Desktop % mpicc test_mpi.c -o test_mpi_c
ccroberts@lumos Desktop % mpiexec -n 2 test_mpi_c
ierr 0
Hello from rank 0
[lumos:45729] *** Process received signal ***
[lumos:45729] Signal: Segmentation fault: 11 (11)
[lumos:45729] Signal code: Address not mapped (1)
[lumos:45729] Failing at address: 0x0
[lumos:45729] [ 0] 0 libsystem_platform.dylib 0x00007ff81b08b25d _sigtramp + 29
[lumos:45729] [ 1] 0 ??? 0x0000000000000000 0x0 + 0
[lumos:45729] [ 2] 0 libmpi.40.dylib 0x000000010be896df ompi_rte_finalize + 415
[lumos:45729] [ 3] 0 libmpi.40.dylib 0x000000010be8c535 ompi_mpi_instance_finalize_common + 1260
[lumos:45729] [ 4] 0 libmpi.40.dylib 0x000000010be8c8b8 ompi_mpi_instance_finalize + 128
[lumos:45729] [ 5] 0 libmpi.40.dylib 0x000000010be86139 ompi_mpi_finalize + 728
[lumos:45729] [ 6] 0 test_mpi_c 0x000000010bbe051d main + 93
[lumos:45729] [ 7] 0 dyld 0x00007ff81acad530 start + 3056
[lumos:45729] *** End of error message ***
--------------------------------------------------------------------------
It looks like MPI_INIT failed for some reason; your parallel process is likely to abort.
There are many reasons that a parallel process can fail during MPI_INIT;
some of which are due to configuration or environment problems.
This failure appears to be an internal failure; here's some additional
information (which may only be relevant to an Open MPI developer):
PML add procs failed
--> Returned "Not found" (-13) instead of "Success" (0)
--------------------------------------------------------------------------
--------------------------------------------------------------------------
It looks like MPI_INIT failed for some reason; your parallel process is likely to abort.
There are many reasons that a parallel process can fail during MPI_INIT;
some of which are due to configuration or environment problems.
This failure appears to be an internal failure; here's some additional
information (which may only be relevant to an Open MPI developer):
ompi_mpi_init: ompi_mpi_instance_init failed
--> Returned "Not found" (-13) instead of "Success" (0)
--------------------------------------------------------------------------
[lumos:00000] *** An error occurred in MPI_Init
[lumos:00000] *** reported by process [3963092993,1]
[lumos:00000] *** on a NULL communicator
[lumos:00000] *** Unknown error
[lumos:00000] *** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
[lumos:00000] *** and MPI will try to terminate your MPI job as well)
--------------------------------------------------------------------------
prterun noticed that process rank 0 with PID 45729 on node lumos exited on signal 11 (Segmentation fault: 11).
--------------------------------------------------------------------------

我已经把自己能想到的排查方向都试过了，但完全找不到问题根源，作为MPI新手实在束手无策，希望有经验的朋友能帮我分析一下这到底是哪里出了问题，非常感谢！

内容来源于stack exchange