AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Remove MPI compile-time dependency (#264)

Browse Source 
* use dlsym for MPI functions

to allow compiling without MPI support, convert the usage of MPI functions and symbols to be based on a dlopen/dlsym based mechanism. Turns out this cannot be done entirely vendor neutral, slightly different solutions might be required for Open MPI, MPICH and the new MPI ABI.

* checkpoint

more work to be done.

* checkpoint 2

* checkpoint 3

* checkpoint 4

examples compile and link correctly

* checkpoitn 5 (I think)

* Checkpoitn 6

* dyld-mpi: adapt GDA

* dyldmpi: tests that depend on MPI need to link with it themselves

* do not ../mpi_instance.h

* dyldmpi: make the symetricHeapTestFixture compile

* dyldmpi: Change cmakery, compiles and run gda w/o external MPI

* Make it also compile in external MPI mode

* dyldmpi: ipc unit tests compile but do not link

* dyldmpi: new approach, if external mpi required, link with mpi,
otherwise use ompi5 abi

* C-style comments in cmakelist..

* dyldmpi: examples: do not fail compiling if MPI not found at build time,
instead do not compile the MPI required examples

* more updates to CMake logic

* convert RO backend

and a few other cleanups

* update some unit tests

to work with the dlopen MPI environment correctly.

---------

Co-authored-by: Aurelien Bouteiller <abouteil@amd.com>
This commit is contained in:
Edgar Gabriel
2025-10-01 08:06:56 -05:00
committed by GitHub
parent 6bb46887e8
commit e4c427a736
50 changed files with 712 additions and 294 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/backend_bc.cpp
+5 -4
View File
@@ -59,6 +59,7 @@ Backend::Backend(MPI_Comm comm) : heap(comm, nullptr) {
Backend::Backend(TcpBootstrap* bootstrap) : heap(MPI_COMM_NULL, bootstrap) {
init();
backend_bootstr = bootstrap;
backend_comm = MPI_COMM_NULL;
my_pe = bootstrap->getRank();
num_pes = bootstrap->getNranks();
@@ -106,9 +107,9 @@ void Backend::init(void) {
void Backend::init_mpi_once(MPI_Comm comm) {
if (comm == MPI_COMM_NULL) comm = MPI_COMM_WORLD;
NET_CHECK(MPI_Comm_dup(comm, &backend_comm));
NET_CHECK(MPI_Comm_size(backend_comm, &num_pes));
NET_CHECK(MPI_Comm_rank(backend_comm, &my_pe));
NET_CHECK(mpilib_ftable_.Comm_dup(comm, &backend_comm));
NET_CHECK(mpilib_ftable_.Comm_size(backend_comm, &num_pes));
NET_CHECK(mpilib_ftable_.Comm_rank(backend_comm, &my_pe));
}
void Backend::track_ctx(Context* ctx) {
@@ -140,7 +141,7 @@ void Backend::destroy_remaining_ctxs() {
Backend::~Backend() {
CHECK_HIP(hipFree(print_lock));
if (backend_comm != MPI_COMM_NULL)
NET_CHECK(MPI_Comm_free(&backend_comm));
NET_CHECK(mpilib_ftable_.Comm_free(&backend_comm));
}
void Backend::dump_stats() {