cudaPackages.cuda_compat: ignore missing libs provided at runtime
cudaPackages.gpus: Jetson should never build by default
cudaPackages.flags: don't build Jetson capabilities by default
cudaPackages: re-introduce filter for pre-existing CUDA redist packages in overrides
cudaPackages: only recurseIntoAttrs for the latest of each major version
cudaPackages.nvccCompatabilities: use GCC 10 through CUDA 11.5 to avoid a GLIBC incompatability
cudaPackages.cutensor: acquire libcublas through cudatoolkit prior to 11.4
cudaPackages.cuda_compat: mark as broken on aarch64-linux if not targeting Jetson
cudaPackages.cutensor_1_4: fix build
cudaPackages: adjust use of autoPatchelfIgnoreMissingDeps
cudaPackages.cuda_nvprof: remove unecessary override to add addOpenGLRunpath
cudaPackages: use getExe' to avoid patchelf warning about missing meta.mainProgram
cudaPackages: fix evaluation with Nix 2.3
cudaPackages: fix platform detection for Jetson/non-Jetson aarch64-linux
python3Packages.tensorrt: mark as broken if required packages are missing
Note: evaluating the name of the derivation will fail if tensorrt is not present,
which is why we wrap the value in `lib.optionalString`.
cudaPackages.flags.getNixSystem: add guard based on jetsonTargets
cudaPackages.cudnn: use explicit path to patchelf
cudaPackages.tensorrt: use explicit path to patchelf
This is useful for the cuda variants of packages like opencv and pytorch,
whose xxxxConfig.cmake files do find_package(CUDAToolkit REQUIRED)
regardless of whether they actually use it. With the propagated hook,
we no longer have to manually add cuda dependencies into torch/opencvs
reverse dependencies
cudaPackages.cuda_nvcc: fix setupCudaHook propagation
We don't need to add the extra nvcc flags to locate cudart when using
cudatoolkit because it comes in the merged layout and nvcc doesn't have
any trouble locating dependencies in the same prefix
We want to remove llvmPackages_6, but it is the only version GHC 8.6.5
supports. Luckily, we actually don't need LLVM in any case, since all
X86 architectures have native codegen for Darwin and Linux, as well as
POWER for Linux. Consequently, we can just pass `null` and add an extra
assert to make this more transparent to future tinkerers.
GHC has a native backend for POWER and SPARC. This is relevant to 8.6.5
binary in the case of POWER where we actually have a bindist and don't
need to include LLVM.
GHC 8.6.5 binary is essentially broken on aarch64-linux, hence the
platform has been removed from meta in the past. We can also drop the
corresponding src data to lessen confusion.
