[rfc][be] static assert that nccl version is >= 2.4 (#142023)

c-p-i-o · pytorchmergebot · commit a9e3281e940b · 2024-12-05T22:11:14.000Z
Summary: Static assert that NCCL VERSION is greater than 2.4. This is in preparation of enabling error checking by default in PyTorch library and removal of some macros. This is in PR #141914. The rationale behind this version is: 1. 2.4 released ~2 years ago so it's unlikely that someone is still using the old library. 2. Enabling error checking is benefitial to the community as it helps debug subtle bugs in production environments. Test Plan: unit tests Differential Revision: D66737055 Pull Request resolved: #142023 Approved by: https://github.com/kwen2501
diff --git a/torch/csrc/distributed/c10d/NCCLUtils.hpp b/torch/csrc/distributed/c10d/NCCLUtils.hpp
@@ -39,6 +39,10 @@ constexpr int64_t kCommInitBusyWaitMillis = 2;
 #define NCCL_REMOTE_ERROR
 #endif
 
+static_assert(
+    (NCCL_MAJOR == 2 && NCCL_MINOR >= 7) || (NCCL_MAJOR > 2),
+    "NCCL version must be 2.7 or later");
+
 // Error checking is enabled only for NCCL versions 2.4+ since ncclCommAbort()
 // and ncclCommGetAsyncError() are not supported in earlier versions.
 #if defined(NCCL_MAJOR) && (NCCL_MAJOR == 2) && defined(NCCL_MINOR) && \
