mirror of
https://github.com/NixOS/nixpkgs.git
synced 2026-08-27 02:34:53 +00:00
67e5bee8a6
autoPatchelfHook only resolves DT_NEEDED entries, so a library reached only through dlopen never lands in any RUNPATH -- listing its provider in buildInputs contributes nothing. Several components dlopen by bare soname and so silently lose those libraries at runtime. Add the missing paths with appendRunpaths for the redists, and an explicit postFixup for the two source builds, which do not go through buildRedist: - cudnn, libcublas, libcufft, libcusparse, libcusparse_lt: NVRTC and nvJitLink, dlopen'd to compile kernels at runtime. libcublas gained the dlopen in 12.8 and libcusparse_lt in 0.8, so both are version-gated. libcusparse's libnvJitLink moved from DT_NEEDED to dlopen at 12.7.3.1; the entry is left unconditional because it is a no-op before that. - libcufile: libudev, libmount, liburcu and libnuma. - nccl, libnvshmem: libibverbs, libmlx5 and libgdrapi. NCCL builds with RDMA_CORE=0 and MLX5DV=0, so none of these are DT_NEEDED; without this NCCL logs at INFO and silently falls back to the socket transport. Both also gain autoAddDriverRunpath, which nothing was supplying -- libnccl dlopens libnvidia-ml.so.1, and the statically linked CUDA runtime dlopens libcuda.so.1. Also: - cuda_cupti: ship the static archives as their own output on linux-sbsa, where they were previously left in the propagated lib output. Gate on hostRedistSystem, since linux-aarch64 ships no archives at all and declaring an empty output fails the build. - libcuobjclient: declare rdma-core, which had been resolved only by accident through libcufile's runpath. - libcusolver: note the vendored METIS in meta.license. Verified across cudaPackages_12_6 through _13_3 on linux-x86_64 and linux-sbsa, and on linux-aarch64 for the CUDA 12 releases where that redist system exists. nccl is unavailable on linux-aarch64, since pre-Thor Jetsons are unsupported upstream, so libnvshmem builds there with withNccl disabled. Known and deliberately left alone: appendRunpaths is applied to every ELF in every output, so it also writes into stubs outputs, where the entries are inert. Nothing in-tree consumes the affected stubs outputs. Assisted-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
43 lines
1.4 KiB
Nix
43 lines
1.4 KiB
Nix
{
|
|
_cuda,
|
|
buildRedist,
|
|
cuda_cudart,
|
|
cuda_nvrtc,
|
|
lib,
|
|
}:
|
|
buildRedist (finalAttrs: {
|
|
redistName = "cusparselt";
|
|
pname = "libcusparse_lt";
|
|
|
|
outputs = [
|
|
"out"
|
|
"dev"
|
|
"include"
|
|
"lib"
|
|
"static"
|
|
];
|
|
|
|
# libcusparseLt dlopens NVRTC (the CASK kernel compiler) from 0.8.
|
|
appendRunpaths = lib.optionals (lib.versionAtLeast finalAttrs.version "0.8") [
|
|
"${lib.getLib cuda_nvrtc}/lib" # libnvrtc.so.%s
|
|
];
|
|
|
|
# NOTE: libcusparseLt does not reference libcublas at all, so it is deliberately not an input.
|
|
buildInputs =
|
|
# For some reason, the 1.4.x release of cusparselt requires the cudart library.
|
|
lib.optionals (lib.hasPrefix "1.4" finalAttrs.version) [ (lib.getLib cuda_cudart) ];
|
|
|
|
meta = {
|
|
description = "High-performance CUDA library dedicated to general matrix-matrix operations in which at least one operand is a structured sparse matrix with 50% sparsity ratio";
|
|
longDescription = ''
|
|
NVIDIA cuSPARSELt is a high-performance CUDA library dedicated to general matrix-matrix operations in which at
|
|
least one operand is a structured sparse matrix with 50% sparsity ratio.
|
|
'';
|
|
homepage = "https://developer.nvidia.com/cusparselt-downloads";
|
|
changelog = "https://docs.nvidia.com/cuda/cublasmp/release_notes";
|
|
|
|
maintainers = [ lib.maintainers.sepiabrown ];
|
|
license = _cuda.lib.licenses.cusparse_lt;
|
|
};
|
|
})
|