SHMEM\_DOUBLE\_G(3) Open MPI SHMEM\_DOUBLE\_G(3)

shmem\_char\_g &lt;#shmem-char-g&gt;(3), shmem\_float\_g &lt;#shmem-float-g&gt;(3), shmem\_int\_g &lt;#shmem-int-g&gt;(3), shmem\_long\_g &lt;#shmem-long-g&gt;(3), shmem\_short\_g &lt;#shmem-short-g&gt;(3), shmem\_longlong\_g &lt;#shmem-longlong-g&gt;(3), shmem\_longdouble\_g &lt;#shmem-longdouble-g&gt;(3) - These routines provide a low latency mechanism to read basic types (char, short, int, float, double, long, long long, long double) from symmetric data objects on remote PEs.

# [SYNOPSIS](#SYNOPSIS)

C or C++:

```
#include <mpp/shmem.h>
char shmem_char_g(const char *addr, int pe)
short shmem_short_g(const short *addr, int pe)
int shmem_int_g(const int *addr, int pe)
long shmem_long_g(const long *addr, int pe)
long shmem_longlong_g(const long long *addr, int pe)
float shmem_float_g(const float *addr, int pe)
double shmem_double_g(const double *addr, int pe)
long shmem_longdouble_g(const long double *addr, int pe)
```

# [DESCRIPTION](#DESCRIPTION)

These routines provide a very low latency get capability for single elements of most basic types.

The arguments are as follows:

[**addr**](#addr)

The remotely accessible array element or scalar data object which will receive the data on the remote PE.

[**pe**](#pe)

The number of the remote PE.

**See also:**

[*intro\_shmem*(3)](/man/enterprise-packages/openmpi-cuda-12.8/intro_shmem.3) [*shmem\_get*(3)](/man/enterprise-packages/openmpi-cuda-12.8/shmem_get.3)

# [Copyright](#Copyright)

2003-2026, The Open MPI Community

September 16, 2026

---

Document: shmem_double_g(3) from openmpi-cuda-12.8 5.0.11-r1 (Chainguard OS)
Source: https://github.com/open-mpi/ompi