SHMEM_LONGLONG_MIN_TO_ALL(3)

📦
version 5.0.10 · 4 versions
, in ubuntu stonking

💖 NAME 💖

shmem_longlong_min_to_all and all its besties — the whole minimum-reduction squad. They do a minimum function reduction across a set of processing elements (PEs). 🫶

Here's the full crew, all of them (3) manpages, all of them iconic:

✨ SYNOPSIS ✨

C or C++:

#include <mpp/shmem.h>

void shmem_double_min_to_all(double *target, const double *source,
  int nreduce, int PE_start, int logPE_stride, int PE_size,
  double *pWrk, long *pSync)

void shmem_float_min_to_all(float *target, const float *source,
  int nreduce, int PE_start, int logPE_stride, int PE_size,
  float *pWrk, long *pSync)

void shmem_int_min_to_all(int *target, const int *source,
  int nreduce, int PE_start, int logPE_stride, int PE_size,
  int *pWrk, long *pSync)

void shmem_long_min_to_all(long *target, const long *source,
  int nreduce, int PE_start, int logPE_stride, int PE_size,
  long *pWrk, long *pSync)

void shmem_longdouble_min_to_all(long double *target,
  const long double *source, int nreduce, int PE_start,
  int logPE_stride, int PE_size, long double *pWrk,
  long *pSync)

void shmem_longlong_min_to_all(long long *target,
  const long long *source, int nreduce, int PE_start, int logPE_stride,
  int PE_size, long long *pWrk, long *pSync)

void shmem_short_min_to_all(short *target, const short *source,
  int nreduce, int PE_start, int logPE_stride, int PE_size,
  short *pWrk, long *pSync)

Fortran:

INCLUDE "mpp/shmem.fh"

INTEGER pSync(SHMEM_REDUCE_SYNC_SIZE)
INTEGER nreduce, PE_start, logPE_stride, PE_size

CALL SHMEM_INT4_MIN_TO_ALL(target, source, nreduce, PE_start,
& logPE_stride, PE_size, pWrk, pSync)

CALL SHMEM_INT8_MIN_TO_ALL(target, source, nreduce, PE_start,
& logPE_stride, PE_size, pWrk, pSync)

CALL SHMEM_REAL4_MIN_TO_ALL(target, source, nreduce, PE_start,
& logPE_stride, PE_size, pWrk, pSync)

CALL SHMEM_REAL8_MIN_TO_ALL(target, source, nreduce, PE_start,
& logPE_stride, PE_size, pWrk, pSync)

CALL SHMEM_REAL16_MIN_TO_ALL(target, source, nreduce, PE_start,
& logPE_stride, PE_size, pWrk, pSync)

🌸 DESCRIPTION 🌸

Okay so bestie, here's the tea. The shared memory (SHMEM) reduction routines compute one or more reductions across symmetric arrays on multiple virtual PEs. And what is a reduction, you ask? It's when you take an associative binary operation and run it across a whole set of values. That's it. That's the whole vibe. 💅

Want the rest of the reduction lineup? The full roster lives in intro_shmem(3) — go look, iykyk.

Now, like all SHMEM collective routines, each of these assumes that only the PEs in the active set call the routine. And if a PE that is not in the active set goes and calls a SHMEM collective routine?? Undefined behavior. Not cute. Don't do it. 😭

The nreduce argument is the number of separate reductions you're doing. The source array on every PE in the active set hands over one element for each reduction. Then the results land in the target array on every PE in the active set. The active set itself is defined by the PE_start, logPE_stride, PE_size triplet — that trio is running the whole show, main character energy fr fr.

Oh and — the source and target arrays may be the same array. But they may not be overlapping arrays. Those are two very different things, bestie, do not conflate them. 🙈

🎀 The arguments, one by one 🎀

  • target — A symmetric array of length nreduce elements that receives the results of the reduction operations. The data type of target varies with the version of the reduction routine you're calling. Calling from C/C++? Peep the SYNOPSIS for the data type info. Calling from Fortran? Here's the target data type tea:
    • shmem_int4_min_to_all: Integer, with an element size of 4 bytes
    • shmem_int8_min_to_all: Integer, with an element size of 8 bytes
    • shmem_real4_min_to_all: Real, with an element size of 4 bytes
    • shmem_real8_min_to_all: Real, with an element size of 8 bytes
    • shmem_real16_min_to_all: Real, with an element size of 16 bytes
  • source — A symmetric array of length nreduce elements, containing one element for each separate reduction operation. The source argument must have the same data type as target. No mixing, no matching, periodt.
  • nreduce — The number of elements in the target and source arrays. Has to be of type integer. If you're in Fortran, it must be a default integer value.
  • PE_start — The lowest virtual PE number of the active set of PEs. Must be of type integer. In Fortran? Default integer value, babe.
  • logPE_stride — The log (base 2) of the stride between consecutive virtual PE numbers in the active set. Must be of type integer. Fortran: default integer value.
  • PE_size — The number of PEs in the active set. Must be of type integer. And yep — if you're using Fortran, default integer value.
  • pWrk — A symmetric work array. It must have the same data type as target. In C/C++, this contains max(nreduce/2 + 1, _SHMEM_REDUCE_MIN_WRKDATA_SIZE) elements. In Fortran, this contains max(nreduce/2 + 1, SHMEM_REDUCE_MIN_WRKDATA_SIZE) elements. 👑
  • pSync — A symmetric work array. In C/C++, pSync is of type long and size _SHMEM_REDUCE_SYNC_SIZE. In Fortran, pSync is of type integer and size SHMEM_REDUCE_SYNC_SIZE — and if you're in Fortran, it must be a default integer value. Every element of this array must be initialized with the value _SHMEM_SYNC_VALUE (C/C++) or SHMEM_SYNC_VALUE (Fortran) before any of the PEs in the active set enter the reduction routine. That's not a suggestion, that's the law. ⭐
The values of the arguments nreduce, PE_start, logPE_stride, and PE_size must be equal on all PEs in the active set. The same target and source arrays, and the same pWrk and pSync work arrays, must be passed to all PEs in the active set. Everyone matching, everyone synced, no solo acts. 🫶

🦋 Before you call it 🦋

Before any PE calls a reduction routine, you have to make sure these conditions exist (and yes, synchronization via a barrier or some other method is often needed to guarantee that, so don't skip it):

  • The pWrk and pSync arrays on all PEs in the active set are not still in use from a prior call to a collective SHMEM routine.
  • The target array on all PEs in the active set is ready to accept the results of the reduction.

🌈 Upon return 🌈

Once the reduction routine comes back, here's what's true for the local PE:

  • The target array is updated.
  • The values in the pSync array are restored to the original values. (She resets herself, we love a self-cleaning queen. 💅)

📝 NOTES 📝

The terms collective, symmetric, and cache aligned are defined over in intro_shmem(3) — go get your vocabulary fixed if you need to.

All SHMEM reduction routines reset the values in pSync before they return, so a particular pSync buffer only needs to be initialized the first time it's used. One and done. Slay.

But — and this is the important part — you must ensure that the pSync array is not being updated on any PE in the active set while any of the PEs participate in processing of a SHMEM reduction routine. Be careful of these situations:

  • If the pSync array is initialized at run time, some type of synchronization is needed to make sure that all PEs in the working set have initialized pSync before any of them enter a SHMEM routine called with the pSync synchronization array.
  • A pSync or pWrk array can be reused in a subsequent reduction routine call only if none of the PEs in the active set are still processing a prior reduction routine call that used the same pSync or pWrk arrays.

In general, you can only be sure of that by doing some type of synchronization. However — special case alert 🚨 — if you're calling reduction routines with the same active set, you can allocate two pSync and pWrk arrays and alternate between them on successive calls. Alternate, alternate, alternate. It's giving double-buffer. 💖

🔥 EXAMPLES 🔥

Example 1: This Fortran example statically initializes the pSync array and finds the minimum value of real variable FOO across all the even PEs.

INCLUDE "mpp/shmem.fh"

INTEGER PSYNC(SHMEM_REDUCE_SYNC_SIZE)
DATA PSYNC /SHMEM_REDUCE_SYNC_SIZE*SHMEM_SYNC_VALUE/
PARAMETER (NR=1)
REAL FOO, FOOMIN, PWRK(MAX(NR/2+1,SHMEM_REDUCE_MIN_WRKDATA_SIZE))
COMMON /COM/ FOO, FOOMIN, PWRK
INTRINSIC MY_PE

IF ( MOD(MY_PE(),2) .EQ. 0) THEN
  CALL SHMEM_REAL8_MIN_TO_ALL(FOOMIN, FOO, NR, 0, 1, N$PES/2,
  & PWRK, PSYNC)
  PRINT *, 'Result on PE ', MY_PE(), ' is ', FOOMIN
ENDIF

Example 2: Consider the following C/C++ call:

shmem_int_min_to_all( target, source, 3, 0, 0, 8, pwrk, psync );

The preceding call is more efficient, but semantically equivalent to, the combination of the following calls:

shmem_int_min_to_all(&(target[0]), &(source[0]), 1, 0, 0, 8,
  pwrk1, psync1);
shmem_int_min_to_all(&(target[1]), &(source[1]), 1, 0, 0, 8,
  pwrk2, psync2);
shmem_int_min_to_all(&(target[2]), &(source[2]), 1, 0, 0, 8,
  pwrk1, psync1);

Note that two sets of pWrk and pSync arrays are used alternately because no synchronization is done between calls. See? We told you. Alternating. Icons only. ✨

👀 SEE ALSO 👀

2003-2026, The Open MPI Community

And that's the whole page, bestie! This one came from the Open MPI crew, dated Sep 10, 2026, and the copyright belongs to The Open MPI Community, 2003-2026. Go forth and reduce your minimums across the active set. Min to all, min to all. 💜🦋✨

📦 where this page comes from