From 27f2256eede2949c04a94136500ec6c1b7fbbc29 Mon Sep 17 00:00:00 2001 From: James Lamb Date: Tue, 24 Sep 2024 23:25:18 -0500 Subject: [PATCH] bump NCCL floor to 2.18.1.1, include nccl.h where it's needed (#4661) Contributes to https://github.com/rapidsai/build-planning/issues/102 Some RAPIDS libraries are using `ncclCommSplit()`, which was introduced in `nccl==2.18.1.1`. This is part of a series of PRs across RAPIDS updating libraries' pins to `nccl>=2.18.1.1` to ensure they get a new-enough version that supports that. Authors: - James Lamb (https://github.com/jameslamb) Approvers: - Chuck Hastings (https://github.com/ChuckHastings) - Vyas Ramasubramani (https://github.com/vyasr) - https://github.com/jakirkham URL: https://github.com/rapidsai/cugraph/pull/4661 --- conda/environments/all_cuda-118_arch-x86_64.yaml | 2 +- conda/environments/all_cuda-125_arch-x86_64.yaml | 2 +- conda/recipes/libcugraph/conda_build_config.yaml | 2 +- cpp/include/cugraph/mtmg/instance_manager.hpp | 2 ++ cpp/include/cugraph/mtmg/resource_manager.hpp | 2 ++ cpp/tests/mtmg/multi_node_threaded_test.cu | 1 + dependencies.yaml | 2 +- 7 files changed, 9 insertions(+), 4 deletions(-) diff --git a/conda/environments/all_cuda-118_arch-x86_64.yaml b/conda/environments/all_cuda-118_arch-x86_64.yaml index 533f23cd7ac..fd91edd8adc 100644 --- a/conda/environments/all_cuda-118_arch-x86_64.yaml +++ b/conda/environments/all_cuda-118_arch-x86_64.yaml @@ -33,7 +33,7 @@ dependencies: - libraft==24.10.*,>=0.0.0a0 - librmm==24.10.*,>=0.0.0a0 - nbsphinx -- nccl>=2.9.9 +- nccl>=2.18.1.1 - networkx>=2.5.1 - networkx>=3.0 - ninja diff --git a/conda/environments/all_cuda-125_arch-x86_64.yaml b/conda/environments/all_cuda-125_arch-x86_64.yaml index 084a6adfd31..19da750601b 100644 --- a/conda/environments/all_cuda-125_arch-x86_64.yaml +++ b/conda/environments/all_cuda-125_arch-x86_64.yaml @@ -39,7 +39,7 @@ dependencies: - libraft==24.10.*,>=0.0.0a0 - librmm==24.10.*,>=0.0.0a0 - nbsphinx -- nccl>=2.9.9 +- nccl>=2.18.1.1 - networkx>=2.5.1 - networkx>=3.0 - ninja diff --git a/conda/recipes/libcugraph/conda_build_config.yaml b/conda/recipes/libcugraph/conda_build_config.yaml index 26aa428d7f5..6b50d0aad23 100644 --- a/conda/recipes/libcugraph/conda_build_config.yaml +++ b/conda/recipes/libcugraph/conda_build_config.yaml @@ -17,7 +17,7 @@ doxygen_version: - ">=1.8.11" nccl_version: - - ">=2.9.9" + - ">=2.18.1.1" c_stdlib: - sysroot diff --git a/cpp/include/cugraph/mtmg/instance_manager.hpp b/cpp/include/cugraph/mtmg/instance_manager.hpp index a2111804997..759635b4a34 100644 --- a/cpp/include/cugraph/mtmg/instance_manager.hpp +++ b/cpp/include/cugraph/mtmg/instance_manager.hpp @@ -20,6 +20,8 @@ #include +#include + #include namespace cugraph { diff --git a/cpp/include/cugraph/mtmg/resource_manager.hpp b/cpp/include/cugraph/mtmg/resource_manager.hpp index a9e4b81f894..e9d25c4576b 100644 --- a/cpp/include/cugraph/mtmg/resource_manager.hpp +++ b/cpp/include/cugraph/mtmg/resource_manager.hpp @@ -27,6 +27,8 @@ #include #include +#include + #include namespace cugraph { diff --git a/cpp/tests/mtmg/multi_node_threaded_test.cu b/cpp/tests/mtmg/multi_node_threaded_test.cu index 06ccd4a7fa1..374c432aac5 100644 --- a/cpp/tests/mtmg/multi_node_threaded_test.cu +++ b/cpp/tests/mtmg/multi_node_threaded_test.cu @@ -39,6 +39,7 @@ #include #include +#include #include #include diff --git a/dependencies.yaml b/dependencies.yaml index 76048be2010..cd34fcf2f70 100644 --- a/dependencies.yaml +++ b/dependencies.yaml @@ -530,7 +530,7 @@ dependencies: - aiohttp - fsspec>=0.6.0 - requests - - nccl>=2.9.9 + - nccl>=2.18.1.1 - ucx-proc=*=gpu - &ucx_py_unsuffixed ucx-py==0.40.*,>=0.0.0a0 - output_types: pyproject