920dbe5b35
Optimization for Tree allreduce on A100. Improve aggregation performance. Use shared buffers for inter-node send/recv. Add NVTX profiling hooks. Accelerate alltoall connections by merging communication for all channels. Add support for one hop communication through NVLink, for faster send/recv communication on cubemesh topologies like DGX-1. Improve alltoall scheduling to better balance intra/inter node communication. Increase send/recv parallelism by 8x, each warp sending or receiving to a different peer. Net: move to v4. Net: make flush operation asynchronous to accelerate alltoall. Net: define maximum number of requests. Fix hang when using LL128 protocol after 2^31 steps. Fix #379 : topology injection failing when using less GPUs than described in the XML. Fix #394 : protocol mismatch causing hangs or crashes when using one GPU per node.
NCCL
Optimized primitives for collective multi-GPU communication.
Introduction
NCCL (pronounced "Nickel") is a stand-alone library of standard collective communication routines for GPUs, implementing all-reduce, all-gather, reduce, broadcast, and reduce-scatter. It has been optimized to achieve high bandwidth on platforms using PCIe, NVLink, NVswitch, as well as networking using InfiniBand Verbs or TCP/IP sockets. NCCL supports an arbitrary number of GPUs installed in a single node or across multiple nodes, and can be used in either single- or multi-process (e.g., MPI) applications.
For more information on NCCL usage, please refer to the NCCL documentation.
What's inside
At present, the library implements the following collectives operations:
- all-reduce
- all-gather
- reduce-scatter
- reduce
- broadcast
These operations are implemented using ring algorithms and have been optimized for throughput and latency. For best performance, small operations can be either batched into larger operations or aggregated through the API.
Requirements
NCCL requires at least CUDA 7.0 and Kepler or newer GPUs. For PCIe based platforms, best performance is achieved when all GPUs are located on a common PCIe root complex, but multi-socket configurations are also supported.
Build
Note: the official and tested builds of NCCL can be downloaded from: https://developer.nvidia.com/nccl. You can skip the following build steps if you choose to use the official builds.
To build the library :
$ cd nccl
$ make -j src.build
If CUDA is not installed in the default /usr/local/cuda path, you can define the CUDA path with :
$ make src.build CUDA_HOME=<path to cuda install>
NCCL will be compiled and installed in build/ unless BUILDDIR is set.
By default, NCCL is compiled for all supported architectures. To accelerate the compilation and reduce the binary size, consider redefining NVCC_GENCODE (defined in makefiles/common.mk) to only include the architecture of the target platform :
$ make -j src.build NVCC_GENCODE="-gencode=arch=compute_70,code=sm_70"
Install
To install NCCL on the system, create a package then install it as root.
Debian/Ubuntu :
$ # Install tools to create debian packages
$ sudo apt install build-essential devscripts debhelper fakeroot
$ # Build NCCL deb package
$ make pkg.debian.build
$ ls build/pkg/deb/
RedHat/CentOS :
$ # Install tools to create rpm packages
$ sudo yum install rpm-build rpmdevtools
$ # Build NCCL rpm package
$ make pkg.redhat.build
$ ls build/pkg/rpm/
OS-agnostic tarball :
$ make pkg.txz.build
$ ls build/pkg/txz/
Tests
Tests for NCCL are maintained separately at https://github.com/nvidia/nccl-tests.
$ git clone https://github.com/NVIDIA/nccl-tests.git
$ cd nccl-tests
$ make
$ ./build/all_reduce_perf -b 8 -e 256M -f 2 -g <ngpus>
Copyright
All source code and accompanying documentation is copyright (c) 2015-2019, NVIDIA CORPORATION. All rights reserved.