Big count support for datatypes #13460

devreal · 2025-10-23T00:51:09Z

Utilize the count and disp arrays for type-punning of int and MPI_Count arguments to datatype creation functions. Store and pack integer or size_t depending on what is needed. Adjust places where bigcount support for datatypes was missing.

This is work in progress and not yet properly tested, other than compiling through it. I wanted to put it out to get some eyes on it to see whether people agree with the design and that it's the right direction, before putting more effort into testing.

Utilize the count and disp arrays for type-punning of int and MPI_Count arguments to datatype creation functions. Pack integer or MPI_Count depending on what is needed when packing a datatype. Adjust places where bigcount support for datatpyes was missing. Signed-off-by: Joseph Schuchart <[email protected]>

Copilot

Pull Request Overview

This PR implements bigcount support for MPI datatypes by introducing type-punned count and displacement arrays that can store either 32-bit integers or 64-bit size_t/ptrdiff_t values. The implementation uses the least significant bit of pointers as a flag to distinguish between the two representations, allowing the datatype framework to handle both standard MPI and bigcount MPI_Count variants transparently.

Key changes:

Introduced opal_count_array_t and opal_disp_array_t wrapper types with helper functions for type-punned storage
Updated datatype creation functions to accept these wrapper types instead of raw int/ptrdiff_t arrays
Modified datatype argument storage to track both 32-bit and 64-bit count arrays separately
Removed conditional bigcount compilation blocks and temporary array allocations throughout the codebase

Reviewed Changes

Copilot reviewed 54 out of 54 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
opal/util/count_disp_array.h	New header defining type-punned array wrapper types and manipulation functions
opal/util/Makefile.am	Added new header to build system
ompi/util/count_disp_array.h	Refactored to delegate to opal implementation
ompi/datatype/ompi_datatype.h	Updated function signatures to use count_array_t types and size_t for counts
ompi/datatype/ompi_datatype_args.c	Rewrote argument packing/unpacking to handle both int and size_t arrays
ompi/datatype/ompi_datatype_create*.c	Updated to use count_array_t wrappers and size_t
ompi/mpi/c/type_*.c.in	Removed bigcount overflow checks and temp arrays, use wrapper types
ompi/mca//*.c	Updated datatype creation calls to use wrapper macros
opal/datatype/opal_datatype_*.c	Changed parameter types from int32_t to ssize_t/size_t

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-10-23T00:52:42Z

ompi/mpi/c/type_get_envelope.c.in


-   rc = ompi_datatype_get_args( type, 0, num_integers, NULL, num_addresses, NULL,
-                           num_datatypes, NULL, combiner );
+   size_t ci, size_t cl, ca, cd;


Syntax error: missing type specifier for variables ca and cd. Should be size_t ci, cl, ca, cd;

Suggested change

size_t ci, size_t cl, ca, cd;

size_t ci, cl, ca, cd;

Copilot · 2025-10-23T00:52:42Z

ompi/mpi/c/type_get_envelope.c.in

-   rc = ompi_datatype_get_args( type, 0, num_integers, NULL, num_addresses, NULL,
-                           num_datatypes, NULL, combiner );
+   size_t ci, size_t cl, ca, cd;
+   rc = ompi_datatype_get_args( type, 0, num_integers, NULL, &cl, NULL &ca, NULL,


Missing comma between NULL and &ca in function call arguments.

Suggested change

rc = ompi_datatype_get_args( type, 0, num_integers, NULL, &cl, NULL &ca, NULL,

rc = ompi_datatype_get_args( type, 0, num_integers, NULL, &cl, NULL, &ca, NULL,

Copilot · 2025-10-23T00:52:42Z

ompi/datatype/ompi_datatype_args.c

-        for (i = 0 ; i < number_of_length ; ++i) {
-            array_of_length[i] = opal_swap_bytes4(array_of_length[i]);
+        if (i_size == 8) {
+            size_t* arr = (size_t*)opal_count_array_ptr(array_of_length);


Variable array_of_length is undefined in this scope. This appears to be using the wrong variable name - should likely be array_of_counts or similar based on the context.

Copilot · 2025-10-23T00:52:43Z

ompi/datatype/ompi_datatype_args.c

+                arr[i] = opal_swap_bytes8(arr[i]);
+            }
+        } else {
+            int* arr = (int*)opal_count_array_ptr(array_of_length);


Variable array_of_length is undefined in this scope. Should be array_of_counts or the appropriate variable name.

Copilot · 2025-10-23T00:52:43Z

ompi/datatype/ompi_datatype_args.c

+            }
+        } else {
+            int* arr = (int*)opal_count_array_ptr(array_of_length);
+            for (i = 0 ; i < number_of_length ; ++i) {


Variable number_of_length is undefined in this scope. Should be number_of_counts or number_of_ints based on the updated variable naming.

Copilot · 2025-10-23T00:52:43Z

ompi/datatype/ompi_datatype_args.c

+                cl = count+1;
+            }
+            ompi_datatype_create_hindexed( count, a_i[1], disp_array, d[0], &datatype );
+            ompi_datatype_set_args( datatype, cl, ci, a_i, count, disp_array, 1, d, MPI_COMBINER_HINDEXED );


Arguments cl and ci are passed in reverse order. Based on the function signature ompi_datatype_set_args(..., size_t ci, size_t cl, ...), they should be ci, cl not cl, ci.

Suggested change

ompi_datatype_set_args( datatype, cl, ci, a_i, count, disp_array, 1, d, MPI_COMBINER_HINDEXED );

ompi_datatype_set_args( datatype, ci, cl, a_i, count, disp_array, 1, d, MPI_COMBINER_HINDEXED );

Copilot · 2025-10-23T00:52:43Z

ompi/datatype/ompi_datatype_args.c

    ompi_datatype_args_t* args = (ompi_datatype_args_t*)datatype->args;
    char* next_packed = (char*)*packed_buffer;

    if( ompi_datatype_is_predefined(datatype) ) {


Variable iposition is uninitialized at this point. It's declared as int *iposition = NULL; on line 446 but never assigned before use. Should be initialized from next_packed like the later code does.

Suggested change

if( ompi_datatype_is_predefined(datatype) ) {

if( ompi_datatype_is_predefined(datatype) ) {

iposition = (int*)next_packed;

Copilot · 2025-10-23T00:52:44Z

ompi/mpi/c/type_contiguous.c.in

+        ompi_datatype_set_args( *newtype,
+                                (sizeof(count) == sizeof(size_t)) ? 0 : 1,
+                                (sizeof(count) == sizeof(size_t)) ? 1 : 0,
+                                a_i, 0, OMPI_COUNT_ARRAY_NULL, 1, &oldtype, MPI_COMBINER_CONTIGUOUS );


Using OMPI_COUNT_ARRAY_NULL where OMPI_DISP_ARRAY_NULL is expected based on the function signature. The parameter at this position should be of type opal_disp_array_t.

Suggested change

a_i, 0, OMPI_COUNT_ARRAY_NULL, 1, &oldtype, MPI_COMBINER_CONTIGUOUS );

a_i, 0, OMPI_DISP_ARRAY_NULL, 1, &oldtype, MPI_COMBINER_CONTIGUOUS );

Copilot · 2025-10-23T00:52:44Z

ompi/datatype/ompi_datatype_args.c

-    array_of_length = (int*)next_buffer;
-    next_buffer    += (number_of_length * sizeof(int));
+    if (number_of_ints > 0) {
+        array_of_ints = (int*)next_buffer;


Missing increment of next_buffer after assigning array_of_ints. The pointer should be advanced by number_of_ints * sizeof(int) to maintain proper buffer alignment, similar to how other arrays are handled.

Suggested change

array_of_ints = (int*)next_buffer;

array_of_ints = (int*)next_buffer;

next_buffer += number_of_ints * sizeof(int);

hppritcha · 2025-10-23T17:27:47Z

okay this is pretty big! i will probably not be able to get to most of this till Monday.

The signatures for MPI_Type_get_envelope and MPI_Type_get_envelope_c are different so we cannot generate them. This file seems to be unused. Signed-off-by: Joseph Schuchart <[email protected]>

Signed-off-by: Joseph Schuchart <[email protected]>

devreal requested review from Copilot and hppritcha October 23, 2025 00:51

devreal marked this pull request as draft October 23, 2025 00:51

github-actions bot added the Target: main label Oct 23, 2025

Copilot AI reviewed Oct 23, 2025

View reviewed changes

devreal requested a review from bosilca October 23, 2025 15:28

devreal added 2 commits October 24, 2025 13:41

Remove type_get_envelope.c.in

1184358

The signatures for MPI_Type_get_envelope and MPI_Type_get_envelope_c are different so we cannot generate them. This file seems to be unused. Signed-off-by: Joseph Schuchart <[email protected]>

Fix copying of values darray in ompi_datatype_set_args

775746f

Signed-off-by: Joseph Schuchart <[email protected]>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Big count support for datatypes #13460

Big count support for datatypes #13460

devreal commented Oct 23, 2025

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

Copilot AI Oct 23, 2025

Uh oh!

hppritcha commented Oct 23, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	rc = ompi_datatype_get_args( type, 0, num_integers, NULL, &cl, NULL &ca, NULL,
	rc = ompi_datatype_get_args( type, 0, num_integers, NULL, &cl, NULL, &ca, NULL,

	ompi_datatype_set_args( datatype, cl, ci, a_i, count, disp_array, 1, d, MPI_COMBINER_HINDEXED );
	ompi_datatype_set_args( datatype, ci, cl, a_i, count, disp_array, 1, d, MPI_COMBINER_HINDEXED );

	if( ompi_datatype_is_predefined(datatype) ) {
	if( ompi_datatype_is_predefined(datatype) ) {
	iposition = (int*)next_packed;

	a_i, 0, OMPI_COUNT_ARRAY_NULL, 1, &oldtype, MPI_COMBINER_CONTIGUOUS );
	a_i, 0, OMPI_DISP_ARRAY_NULL, 1, &oldtype, MPI_COMBINER_CONTIGUOUS );

	array_of_ints = (int*)next_buffer;
	array_of_ints = (int*)next_buffer;
	next_buffer += number_of_ints * sizeof(int);

Big count support for datatypes #13460

Are you sure you want to change the base?

Big count support for datatypes #13460

Conversation

devreal commented Oct 23, 2025

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Oct 23, 2025

Choose a reason for hiding this comment

Uh oh!

hppritcha commented Oct 23, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants