Skip to content

Latest commit

 

History

History
224 lines (167 loc) · 9.98 KB

README.md

File metadata and controls

224 lines (167 loc) · 9.98 KB
Raw

Intel® SHMEM

Installation   |   Usage   |   Release Notes   |   Documentation   |   How to Contribute   |   License

Intel® SHMEM provides an efficient implementation of GPU-initiated communication on systems with Intel GPUs.

Table of Contents

Prerequisites

  • Linux OS
  • Intel® oneAPI DPC++/C++ Compiler 2024.0 or higher.

SYCL support

Intel® oneAPI DPC++/C++ Compiler with Level Zero support.

Installation

Building Level Zero

For detailed information on Level Zero, refer to the Intel® Graphics Compute Runtime for oneAPI Level Zero and OpenCL™ Driver repository or to the installation guide for oneAPI users.

To install, download the oneAPI Level Zero from the repository.

git clone https://github.com/oneapi-src/level-zero.git

Build Level Zero following instructions below.

cd level-zero
mkdir build
cd build
cmake -DCMAKE_INSTALL_PREFIX=<level_zero_dir> ..
make -j
make install

The Host Back-End Library

Intel® SHMEM requires a host OpenSHMEM or MPI back-end to be used for host-sided operations support. In particular, the OpenSHMEM back-end relies on a collection of extension APIs (shmemx_heap_create, shmemx_heap_preinit, and shmemx_heap_postinit) to coordinate the Intel® SHMEM and OpenSHMEM heaps. We recommend Sandia OpenSHMEM v1.5.3rc1 or newer for this purpose. A work-in-progress branch of OSHMPI is also supported but is currently considered experimental. See the Building OSHMPI section before for more details.

We recommend the Intel® MPI Library as the MPI back-end option for the current version of Intel® SHMEM. See the Building Intel® SHMEM section below for more details.

Building Sandia OpenSHMEM (SOS)

Download the SOS repo to be configured as a back-end for Intel® SHMEM.

git clone --recurse-submodules https://github.com/Sandia-OpenSHMEM/SOS.git SOS

Build SOS following instructions below. FI_HMEM support in the provider is required for use with Intel® SHMEM. To enable FI_HMEM with a supported provider, we recommend a specific set of config flags. Below are two examples for configuring and building SOS with two providers supporting FI_HMEM. To configure SOS with the verbs;ofi_rxm provider, use the following instructions:

cd SOS
./autogen.sh
CC=icx CXX=icpx ./configure --prefix=<shmem_dir> --with-ofi=<ofi_installation> --enable-pmi-simple --enable-ofi-mr=basic --disable-ofi-inject --enable-ofi-hmem --disable-bounce-buffers --enable-hard-polling
make -j
make install

To configure SOS with the HPE Slingshot provider cxi, please use the following instructions:

cd SOS
./autogen.sh
CC=icx CXX=icpx ./configure --prefix=<shmem_dir> --with-ofi=<ofi_installation> --enable-pmi-simple --enable-ofi-mr=basic --disable-ofi-inject --enable-ofi-hmem --disable-bounce-buffers --enable-ofi-manual-progress --enable-mr-endpoint --disable-nonfetch-amo --enable-manual-progress
make -j
make install

To configure SOS with the psm3 provider, please use the following instructions:

cd SOS
./autogen.sh
CC=icx CXX=icpx ./configure --prefix=<shmem_dir> --with-ofi=<ofi_installation> --enable-pmi-simple --enable-manual-progress --enable-ofi-hmem --disable-bounce-buffers --enable-ofi-mr=basic --enable-mr-endpoint
make -j
make install

Please choose an appropriate PMI configure flag based on the available PMI client library in the system. Please check for further instructions on SOS Wiki pages. Optionally, users may also choose to add --disable-fortran since fortran interfaces will not be used.

Building OSHMPI (Optional and experimental)

Intel® SHMEM has experimental support for OSHMPI when built using the Intel® MPI Library. Here is information on how to Get Started with Intel® MPI Library on Linux.

To download the OSHMPI repository:

git clone -b wip/ishmem --recurse-submodules https://github.com/davidozog/oshmpi.git oshmpi

After ensuring Intel® MPI Library is enabled (for example, by sourcing the /opt/intel/oneapi/setvars.sh script), please build OSHMPI following the instructions below.

cd oshmpi
./autogen.sh
CC=mpiicx CXX=mpiicpx ./configure --prefix=<shmem_dir> --disable-fortran --enable-rma=direct --enable-amo=direct --enable-async-thread=yes
make -j
make install

Building Intel® SHMEM

Check that the SOS build process has successfully created a <shmem_dir> directory with include and lib as subdirectories. Please find shmem.h and shmemx.h in include.

Build Intel® SHMEM with an OpenSHMEM back-end using the following instructions:

cd ishmem
mkdir build
cd build
CC=icx CXX=icpx cmake .. -DENABLE_OPENSHMEM=ON -DSHMEM_DIR=<shmem_dir> -DCMAKE_INSTALL_PREFIX=<ishmem_install_dir>
make -j

Alternatively, Intel® SHMEM can be built by enabling an Intel® MPI Library back-end. Here is information on how to Get Started with Intel® MPI Library on Linux.

CC=icx CXX=icpx cmake .. -DENABLE_OPENSHMEM=OFF -DENABLE_MPI=ON -DMPI_DIR=<impi_dir> -DCMAKE_INSTALL_PREFIX=<ishmem_install_dir>

where <impi_dir> is the path to the Intel® MPI Library installation.

Enabling both the OpenSHMEM and MPI back-ends is also supported. In this case, the desired backend can be selected via the environment variable, ISHMEM_RUNTIME, which can be set to either "OpenSHMEM" or "MPI". The default value for ISHMEM_RUNTIME is "OpenSHMEM".

Usage

Launching Example Application

Validate that Intel® SHMEM was built correctly by running an example program.

  1. Add the path for the back-end library to the environment, for example:
export LD_LIBRARY_PATH=<shmem_dir>/lib:$LD_LIBRARY_PATH

When enabling only the Intel® MPI Library back-end, simply source the appropriate setvars.sh script. When enabling both OpenSHMEM and MPI back-ends, first source the setvars.sh script, then configure the dynamic linker to load the OpenSHMEM library (for example by prepending <shmem_dir>/lib to LD_LIBRARY_PATH).

  1. Run the example program or test on an allocated node using a process launcher:
ISHMEM_RUNTIME=<back-end> mpiexec.hydra -n 2 -hosts <allocated_node_id> ./scripts/ishmrun ./test/unit/int_get_device

where <back-end> is the selected host back-end library.

  • Note: Current supported launchers include: MPI process launchers (i.e. mpiexec, mpiexec.hydra, mpirun, etc.), Slurm (i.e. srun, salloc, etc.), and PBS (i.e. qsub).

  • Note: Intel® SHMEM execution model requires applications to use a 1:1 mapping between PEs and GPU devices. Attempting to run an application without the ishmrun launch script may result in undefined behavior if this mapping is not maintained.

  1. Validate the application ran successfully; example output:
Selected device: Intel(R) Data Center GPU Max 1550
Selected vendor: Intel(R) Corporation
Selected device: Intel(R) Data Center GPU Max 1550
Selected vendor: Intel(R) Corporation
No errors
No errors

Launching Example Application w/ CTest

ctest can be used to run Intel® SHMEM tests that are generated at compile-time. To see a list of tests available via ctest, run:

ctest -N

To launch a single test, execute:

ctest -R <test_name>

Alternatively, all the tests in a directory (such as test/unit/) can be run with the following command:

ctest --test-dir <directory_name>

By default, a passed or failed test can be detected by the output:

    Start 69: sync-2-gpu
1/1 Test #69: sync-2-gpu .......................   Passed    2.29 sec

100% tests passed, 0 tests failed out of 1

To have a test's output printed to the console, add either the --verbose or --output-on-failure flag to the ctest command

Available Scheduler Wrappers for Jobs Run via CTest

The following values may be assigned to CTEST_LAUNCHER at configure-time (ex. -DCTEST_LAUNCHER=mpi) to set which scheduler will be used to run tests launched through a call to ctest:

  • srun (default)
    • Launches CTest jobs on a single node using Slurm's srun.
  • mpi
    • Uses mpirun to launch CTest jobs with the appropriate number of processes.
  • qsub
    • Launches CTest jobs on a single node using qsub. If this option is being used on a system where a reservation must be made (i.e. via pbsresnode) prior to running a test, assign the JOB_QUEUE environment variable to the queue associated with your reservation:
    export JOB_QUEUE=<queue>

Additional Resources

OpenSHMEM Specification

Intel® SHMEM Specification