NVIDIA® NVSHMEM 3.4.5 Release Notes#
NVSHMEM is an implementation of the OpenSHMEM specification for NVIDIA GPUs. The NVSHMEM programming interface implements a Partitioned Global Address Space (PGAS) model across a cluster of NVIDIA GPUs. NVSHMEM provides an easy-to-use interface to allocate memory that is symmetrically distributed across the GPUs. In addition to a CPU-side interface, NVSHMEM provides a NVIDIA® CUDA® kernel-side interface that allows CUDA threads to access any location in the symmetrically-distributed memory.
The release notes describe the key features, software enhancements and improvements, and known issues for NVSHMEM 3.4.5 and earlier releases.
Key Features and Enhancements#
This NVSHMEM release includes the following key features and enhancements:
Added support for data direct NIC configurations in the IB transports. Added a new environment variable,
NVSHMEM_DISABLE_DATA_DIRECT
, to force disable data direct NIC even when present.Added support for CPU-Assisted IBGDA without the use of GDRCopy or the x86 regkey setting. Systems not supporting the other methods will automatically fall back to this new method. It enables the use of IBGDA on a broad range of systems without the need for administrator intervention.
Added a new environment variable
NVSHMEM_HCA_PREFIX
to enable IB transports on systems which name their HCA devices in a non-standard way (for example,ipb*
instead ofmlx5*
).Deprecated support for the combined
libnvshmem.a
host and device static library.
Compatibility#
NVSHMEM 3.4.5 has been tested with the following:
CUDA Toolkit:
12.2
12.6
12.9
13.0
CPUs:
x86 and NVIDIA Grace™ processors
GPUs:
NVIDIA Ampere A100
NVIDIA Hopper™
NVIDIA Blackwell®
Limitations#
NVSHMEM is not compatible with the PMI client library on Cray systems, and must use the NVSHMEM internal PMI-2 client library.
You can launch jobs with the PMI bootstrap by specifying
--mpi=pmi2
to Slurm andNVSHMEM_BOOTSTRAP_PMI=PMI-2
, or directly by using the MPI or SHMEM bootstraps.You can also set PMI-2 as the default PMI by setting
NVSHMEM_DEFAULT_PMI2=1
when you build NVSHMEM.
The
libfabric
transport does not support VMM yet, so you must disable VMM by settingNVSHMEM_DISABLE_CUDA_VMM=1
.Systems with PCIe peer-to-peer communication require one of the following:
InfiniBand to support NVSHMEM atomics APIs
Using NVSHMEM’s UCX transport that, if IB is absent, will use sockets for atomics
nvshmem_barrier*
,nvshmem_quiet
, ornvshmem_wait_until
only ensures ordering and visibility between the source and destination PEs and does not ensure global ordering and visibility.When built with GDRCopy, and when using InfiniBand on earlier versions of the 460 driver and previous branches, NVSHMEM cannot allocate the complete device memory because of the inability to reuse the BAR1 space. This has been fixed in the CUDA release 460 driver and in release 470 and later.
IBGDA does not work with CX-4 when the link layer is Ethernet (RoCE).
NVSHMEM is not supported on Grace with Ada L40 platforms.
NVSHMEM is not supported on virtualized environments (VM).
User buffers registered with
nvshmemx_buffer_register_symmetric
lack support forlibfabric
transport to perform GPU-GPU communication over Remote networks (EFA, Slingshot, etc.).When registering Extended GPU memory (EGM) user buffers with
nvshmemx_buffer_register_symmetric
, the buffers on different PEs must belong to distinct CPU sockets within a node. This can be achieved by selecting GPUs on a different NUMA domain using theCUDA_VISIBLE_DEVICES
environment variable.When using the Libfabric transport with
NVSHMEM_LIBFABRIC_PROVIDER=EFA
, you must ensure that thelibfabric
environment variableFI_EFA_ENABLE_SHM_TRANSFER
is set to0
before launching their application. While NVSHMEM does set this variable during initialization, it can be ignored by the EFA provider if it was already initialized by the launcher, for example when using mpirun.
Deprecated Features#
Support for
libnvshmem.a
is now deprecated.
Known Issues#
Complex types, which are enabled by setting
NVSHMEM_COMPLEX_SUPPORT
at compile time, are not currently supported.When enabling
libfabric
transport withNVSHMEM_LIBFABRIC_PROVIDER=EFA
, certain operations are experimental and may cause the application kernel to hang in the following operations:Device side
nvshmem_put/nvshmem_get
withnvshmem_barrier
Host side
nvshmem_put_on_stream/nvshmem_get_on_stream
When you enable UCX remote transport with
NVSHMEM_REMOTE_TRANSPORT=UCX
, you may observe a data mismatch when scaling 32 PEs or more on DGX-2 platform.