NVSHMEM


NVSHMEM™ is a parallel programming interface based on OpenSHMEM that provides efficient and scalable communication for NVIDIA GPU clusters. NVSHMEM creates a global address space for data that spans the memory of multiple GPUs and can be accessed with fine-grained GPU-initiated operations, CPU-initiated operations, and operations on CUDA® streams.


Download NVSHMEM Documentation Release Notes GitHub NVSHMEM API Guide

Existing communication models, such as Message-Passing Interface (MPI), orchestrate data transfers using the CPU. In contrast, NVSHMEM uses asynchronous, GPU-initiated data transfers, eliminating synchronization overheads between the CPU and the GPU.

Efficient, Strong Scaling

NVSHMEM enables long-running kernels that include both communication and computation, reducing overheads that can limit an application’s performance when strong scaling.

Low Overhead

One-sided communication primitives reduce overhead by allowing the initiating process or GPU thread to specify all information required to complete a data transfer. This low-overhead model enables many GPU threads to communicate efficiently.

Naturally Asynchronous

Asynchronous communications make it easier for programmers to interleave computation and communication, thereby increasing overall application performance.


What's New in NVSHMEM 3.6

  • Added configuration file support (similar to NCCL) for easy and repeatable environment variable management.
  • Added experimental NVSHMEM LTO-IR (Link-Time Optimization IR) library build option for improved device code optimization.
  • Added enhanced user buffer registration with preferred address support via nvshmemx_buffer_register_symmetric.​
  • Added error code return values for tile API calls to improve error handling.​
  • Added multi-NIC support for libfabric transport with round-robin NIC selection. The new environment variable NVSHMEM_LIBFABRIC_MAX_NIC_PER_PE controls the maximum number of NICs used per PE.
  • Improved version mismatch error messages to include detailed host and device library version information.

What's New in NVSHMEM4Py 0.3.0

  • Added CuTe DSL support for NVSHMEM4Py with device-side bindings for RMA, collective, AMO, and memory operations.
  • Added device-side construction of peer and multicast tensors for CuTe DSL.
  • Added helper functions to simplify NVSHMEM/CuTe DSL usage.​
  • Made Numba-CUDA an optional dependency and bumped minimum version to 0.25.​
  • Fixed peer/multimem buffer tracking assumptions for parent buffer cleanup.

Key Features


  • Combines the memory of multiple GPUs into a partitioned global address space that’s accessed through NVSHMEM APIs
  • Includes a low-overhead, in-kernel communication API for use by GPU threads
  • Includes stream-based and CPU-initiated communication APIs
  • Supports x86 and Arm processors.
  • Is interoperable with MPI and other OpenSHMEM implementations

NVSHMEM Advantages


Increase Performance

Convolution is a compute-intensive kernel that’s used in a wide variety of applications, including image processing, machine learning, and scientific computing. Spatial parallelization decomposes the domain into sub-partitions that are distributed over multiple GPUs with nearest-neighbor communications, often referred to as halo exchanges.

In the Livermore Big Artificial Neural Network (LBANN) deep learning framework, spatial-parallel convolution is implemented using several communication methods, including MPI and NVSHMEM. The MPI-based halo exchange uses the standard send and receive primitives, whereas the NVSHMEM-based implementation uses one-sided put, yielding significant performance improvements on Lawrence Livermore National Laboratory’s Sierra supercomputer.


Efficient Strong-Scaling on Sierra Supercomputer



Efficient Strong-Scaling on NVIDIA DGX SuperPOD

Accelerate Time to Solution

Reducing the time to solution for high-performance, scientific computing workloads generally requires a strong-scalable application. QUDA is a library for lattice quantum chromodynamics (QCD) on GPUs, and it’s used by the popular MIMD Lattice Computation (MILC) and Chroma codes.

NVSHMEM-enabled QUDA avoids CPU-GPU synchronization for communication, thereby reducing critical-path latencies and significantly improving strong-scaling efficiency.

Watch the GTC 2020 Talk



Simplify Development

The conjugate gradient (CG) method is a popular numerical approach to solving systems of linear equations, and CGSolve is an implementation of this method in the Kokkos programming model. The CGSolve kernel showcases the use of NVSHMEM as a building block for higher-level programming models like Kokkos.

NVSHMEM enables efficient multi-node and multi-GPU execution using Kokkos global array data structures without requiring explicit code for communication between GPUs. As a result, NVSHMEM-enabled Kokkos significantly simplifies development compared to using MPI and CUDA.


Productive Programming of Kokkos CGSolve

Resources



Ready to start developing with NVSHMEM?


Get Started