Resolving dependencies... Starting splitmix-0.1.0.1 Starting hashable-1.3.0.0 Building splitmix-0.1.0.1 Building hashable-1.3.0.0 Completed splitmix-0.1.0.1 Starting random-1.2.0 Building random-1.2.0 Completed hashable-1.3.0.0 Completed random-1.2.0 Starting uuid-types-1.0.3 Building uuid-types-1.0.3 Completed uuid-types-1.0.3 Starting cuda-0.10.2.0 Failed to install cuda-0.10.2.0 Build log ( /home/builder/.cabal/logs/ghc-8.6.3/cuda-0.10.2.0-CRqUereo8MVIWFMzvHYKfh.log ): cabal: Entering directory '/tmp/cabal-tmp-15704/cuda-0.10.2.0' [1 of 1] Compiling Main ( /tmp/cabal-tmp-15704/cuda-0.10.2.0/dist/setup/setup.hs, /tmp/cabal-tmp-15704/cuda-0.10.2.0/dist/setup/Main.o ) Linking /tmp/cabal-tmp-15704/cuda-0.10.2.0/dist/setup/setup ... Configuring cuda-0.10.2.0... setup: ******************************************************************************** The configuration process failed to locate your CUDA installation. Ensure that you have installed both the developer driver and toolkit, available from: > http://developer.nvidia.com/cuda-downloads and make sure that `nvcc` is available in your PATH, or set the CUDA_PATH environment variable appropriately. Check the above output log and run the command directly to ensure it can be located. If you have a non-standard installation, you can add additional search paths using --extra-include-dirs and --extra-lib-dirs. Note that 64-bit Linux flavours often require both `lib64` and `lib` library paths, in that order. ******************************************************************************** cabal: Leaving directory '/tmp/cabal-tmp-15704/cuda-0.10.2.0' cabal: Error: some packages failed to install: cuda-0.10.2.0-CRqUereo8MVIWFMzvHYKfh failed during the configure step. The exception was: ExitFailure 1 nvvm-0.10.0.0-IEnWznzdOJvsFCuOMNnyA depends on nvvm-0.10.0.0 which failed to install.