Changelog:
===========
- Fix initialization in GPU-aware builds when no devices are present
- Fix internal pmix.h header conflict when building with an external
PMIx library.
- Fix build issue with Slurm by removing dependency on libslurm and
always using internal logic for parsing the Slurm hostfile.
- Fix potential stale GPU IPC handle usage resulting in data corruption
or crashes
- Update XPMEM thresholds to avoid excessive buffer mapping overhead
- Fix potential hang in ROMIO when setting info hints on certain files
- Improved detection of incompatible PMI[x] client/server configuration
- Fix use of PMIX_PREFIX attribute for certain versions of OpenPMIx
- Fix Intel GPU output with MPIR_CVAR_DEBUG_SUMMARY
- Fix F08 binding compilation with nvfortran
- Fix line continuation Hydra's --configfile option
- Fix valgrind uninitialized read warnings in ch3
- Fix missing mpixxx_opts.conf file with help text for mpicc and friends
- Fixes for several compiler warnings
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This reverts commit 24ff52ba3b.
The original patch was my bad. The patches for oe-core were re-worked,
but I forgot the recall this patch.
In fact, inheriting qemu is needed because it sets a clear barriar
for people to use qemu user mode. And the QEMU_OPTIONS settings
are also in qemu.bbclass.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Avoid inheriting qemu.bbclass and use oe.qemu.xxx instead.
Also, the 'qemu-native' dep is replaced by 'qemuwrapper-cross' for
PACKAGE_WRITE_DEPS. qemuwrapper-cross is the one that is actually
used by postints and it has 'qemu-native' in DEPENDS.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
===========
- Update embedded libfabric to fix a build issue on FreeBSD
- Fix HIP support for use with AMD GPUs
- Fix potential invalid context issue in CUDA memory hooks
- Fix GPU fallback path in ch4/ofi for Intel GPU buffers
- Fix IPC handle destruction with Level Zero API (Intel GPU)
- Fix potential crash in MPI_ISENDRECV with derived datatypes
- Fix bug in persistent MPI_GATHER that incorrectly cached buffer
contents at init time
- Fix memory allocation bug in ROMIO GPFS driver
- Fix missing error names in ch4/ofi netmod
- Fix potential hang in multi-VCI active message RMA
- Fix bug in ch3 large count support with derived datatypes
- Fix manpage generation to provide aliases for large-count versions
- Fix potential crash in Hydra with long PMI command messages
- Fix bug in exit status capture in Hydra when there are multiple
processes with non-zero exit
- Fix implementation of C/Fortran status conversion functions
- Fix implementation of MPI_Type_create_f90_xxx functions
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
============
# Update embedded libfabric to v1.20.1 and fix compilation with GCC 14.
# Fix dataloop support for MPIX_Type_iov routines
# Fix crash in Hydra when system has many local ip addresses
# Fix RMA fallback check in ch4:ofi netmod
# Fix MPI_UNDEFINED handling in mpi_f08 module
# Fix Slurm environment variable inheritance in Hydra
# Fix multiple issues with large count support in ROMIO
# Fix potential hang in init using PMIx client and nonstandard keys
# Fix crash if PMIx client cannot get appnum from server during init
# Fix other build errors and warnings
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
=============
-Disable flat_namespace to build libmpifort on macOS by default
-Prefix warning messages with "MPICH"
-Add --disable-doc configure option
-Fix support for building MPICH Fortran support with Xcode 15 on macOS
-Fix bug in MPI_WIN_SHARED_QUERY when window is not shared
-Fix bug in ch4/ofi gpu pipelining
-Fixes for Intel GPU support
-Fix memory leak in ch4/shm collectives
-Fix bug in MPI_COMM_SPLIT with intercommunicators and non-zero root
-Fix bug in DAOS ROMIO driver
-Fix bug in cycling error code array
-Return an error if there is failure to create a datatype in mpi_f08
module for noncontiguous user data
-Return an error when shared memory segment creation fails
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog:
==========
with recent versions of LLVM/Clang
different floating point values
attribute on the same object
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
mpich 4.1+ has its own official way to allow cross-compiling,
use it.
Remove the old PACKAGECONFIG[x11], the configuration option
does not exist anymore.
Add PACKAGECONFIG knobs to allow building the C++ and FORTRAN
parts. Both are needed by intel-oneapi-mkl in meta-intel. Don't
enable either by default to keep the current minimalistic build.
Remove configuration options that don't exist in the new version.
Fix TMPDIR being emitted into binaries and scripts like mpicc and mpicxx
Signed-off-by: Zoltán Böszörményi <zboszor@gmail.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changes in 4.0.3
=================
Fix message queue dumping interface support
Fix multinic usage in ch4:ofi
Fix bug in MPI_WIN_CREATE in ch4:ucx when UCX >= 1.13.0
Fix MPIR_pmi_barrier when PMI2 is used
Fix ROMIO lazy mutex initialization
Fix build with HIP support
Fix potential dynamic process message mixups in ch3
Add missing const to MPI_Pready_list array_of_partitions argument
Add support for C++ datatypes even when the C++ binding is disabled
Add support for Intel OneAPI compilers
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Changelog from 3.4.3 to 4.0.2:
https://www.mpich.org/category/news/
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <raj.khem@gmail.com>
This is the result of automated script (0.9.1) conversion:
oe-core/scripts/contrib/convert-overrides.py .
converting the metadata to use ":" as the override character instead of "_".
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
The build needs them in some subcomponents, it does not come to front
when using libgcc as default but when compiler-rt is used then it does
not link because its expecting functions from libgcc
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Explicitly link with libgcc for __addtf3, __multf3 etc 128bit builtins
on x86_64 as they are not available with compiler-rt and when using
compiler-rt the link would fail
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Add fixes to work with/with-out X
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Jose Quaresma <quaresma.jose@gmail.com>
Cc: Andreas Müller <schnitzeltony@gmail.com>
Fixes
QA Issue: /usr/bin/parkill contained in package mpich requires /mnt/a/oe/build/tmp/
hosttools/perl, but no providers found in RDEPENDS_mpich
also move BASH_SHELL in same realm
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Remove the reference of S from .la and DEBUG_PREFIX_MAP
from .pc file to avoid QA issues.
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* remove tabs which sneaked in since last cleanup
* meta-oe layers are using consistent indentation with 4 spaces, see
http://www.openembedded.org/wiki/Styleguide
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
On Fedora 17, bash has moved to /usr/bin/bash and the configure process finds
it on the host machine there, this ensures that it is set correctly for the
target.
Signed-off-by: Victor Rodriguez <victor.rodriguez.bahena@intel.com>
Set --disable-rpath to extra opts in order to avoid issues
in package_qa also change default pm to gforker because
hydra has problems with RPATH.
TODO: Make a patch for solve hydra rpath issues.
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
This patch add mpich support to yocto in order to run MPI framework on embedded
systems.
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Victor Rodriguez <victor.rodriguez.bahena@intel.com>
Signed-off-by: Alexandru.Vaduva <Alexandru.Vaduva@enea.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
