In reply to @connorbaker:matrix.org
wouldn't things like API changes between versions cause breakage?
EDIT: I guess they would cause build failures... my primary concern was that it would cause failures at runtime, but I suppose that's not really a problem for compiled targets. Relative to libc, NVIDIA's libraries change way, way more between releases (even minor versions!).

In reply to @connorbaker:matrix.org
SomeoneSerge (nix.camp): as a short-term thing, are you okay with me patching out OpenCV's requirement that CUDA version match so we can merge the CUDA fix?

Sure let's try. I'd still check something something trivial like

# test1
import torch
torch.randn(10, 10, device="cuda").sum().item()
import cv2
# do something with cv2 and cuda

# test2

import cv2
# do something with cv2 and cuda
import torch
torch.randn(10, 10, device="cuda").sum().item()

In reply to @ss:someonex.net
Sorry my availability has been limited this week

In reply to @ss:someonex.net

Sure let's try. I'd still check something something trivial like

# test1
import torch
torch.randn(10, 10, device="cuda").sum().item()
import cv2
# do something with cv2 and cuda

# test2

import cv2
# do something with cv2 and cuda
import torch
torch.randn(10, 10, device="cuda").sum().item()

In reply to @connorbaker:matrix.org
It is surprising to me that 0.3.5 works with CUDA 12 at all; I guess there were no breaking API changes on stuff they relied on?

In reply to @connorbaker:matrix.org
To clarify SomeoneSerge (nix.camp), do you want a test like that in the OpenCV PR, or is it okay if that's tracked (via https://github.com/NixOS/nixpkgs/issues/341650) and added later?

In reply to @adam:robins.wtf
manually copying it over from the build host allows ollama to successfully work

In reply to @adam:robins.wtf
it's that cudaToolkit/cuda-merged env that wasn't being included

connor (he/him) (UTC-5) tiny-cuda-nn fails on aarch64-linux with:

-- Build files have been written to: /build/source/build
cmake: enabled parallel building
cmake: enabled parallel installing
Running phase: buildPhase
build flags: -j80
[1/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o.d -x cu -c /build/source/src/encoding.cu -o CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[2/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/loss.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/loss.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/loss.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/loss.cu.o.d -x cu -c /build/source/src/loss.cu -o CMakeFiles/tiny-cuda-nn.dir/src/loss.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[3/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/reduce_sum.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/reduce_sum.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/reduce_sum.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/reduce_sum.cu.o.d -x cu -c /build/source/src/reduce_sum.cu -o CMakeFiles/tiny-cuda-nn.dir/src/reduce_sum.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[4/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/cpp_api.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/cpp_api.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/cpp_api.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/cpp_api.cu.o.d -x cu -c /build/source/src/cpp_api.cu -o CMakeFiles/tiny-cuda-nn.dir/src/cpp_api.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[5/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/object.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/object.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/object.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/object.cu.o.d -x cu -c /build/source/src/object.cu -o CMakeFiles/tiny-cuda-nn.dir/src/object.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[6/20] Building CUDA object samples/CMakeFiles/mlp_learning_an_image.dir/mlp_learning_an_image.cu.o
FAILED: samples/CMakeFiles/mlp_learning_an_image.dir/mlp_learning_an_image.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT samples/CMakeFiles/mlp_learning_an_image.dir/mlp_learning_an_image.cu.o -MF samples/CMakeFiles/mlp_learning_an_image.dir/mlp_learning_an_image.cu.o.d -x cu -c /build/source/samples/mlp_learning_an_image.cu -o samples/CMakeFiles/mlp_learning_an_image.dir/mlp_learning_an_image.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[7/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/common.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/common.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/common.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/common.cu.o.d -x cu -c /build/source/src/common.cu -o CMakeFiles/tiny-cuda-nn.dir/src/common.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[8/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/optimizer.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/optimizer.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/optimizer.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/optimizer.cu.o.d -x cu -c /build/source/src/optimizer.cu -o CMakeFiles/tiny-cuda-nn.dir/src/optimizer.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[9/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o.d -x cu -c /build/source/src/cutlass_mlp.cu -o CMakeFiles/tiny-cuda-nn.dir/src/cutlass_mlp.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[10/20] Building CUDA object benchmarks/image/CMakeFiles/bench_image_ours.dir/bench_ours.cu.o
FAILED: benchmarks/image/CMakeFiles/bench_image_ours.dir/bench_ours.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT benchmarks/image/CMakeFiles/bench_image_ours.dir/bench_ours.cu.o -MF benchmarks/image/CMakeFiles/bench_image_ours.dir/bench_ours.cu.o.d -x cu -c /build/source/benchmarks/image/bench_ours.cu -o benchmarks/image/CMakeFiles/bench_image_ours.dir/bench_ours.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[11/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/common_device.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/common_device.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/common_device.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/common_device.cu.o.d -x cu -c /build/source/src/common_device.cu -o CMakeFiles/tiny-cuda-nn.dir/src/common_device.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[12/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/network.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/network.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/network.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/network.cu.o.d -x cu -c /build/source/src/network.cu -o CMakeFiles/tiny-cuda-nn.dir/src/network.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
[13/20] Building CXX object dependencies/fmt/CMakeFiles/fmt.dir/src/os.cc.o
[14/20] Building CXX object dependencies/fmt/CMakeFiles/fmt.dir/src/format.cc.o
[15/20] Building CXX object samples/CMakeFiles/mlp_learning_an_image.dir/__/dependencies/stbi/stbi_wrapper.cpp.o
[16/20] Building CXX object benchmarks/image/CMakeFiles/bench_image_ours.dir/__/__/dependencies/stbi/stbi_wrapper.cpp.o
ninja: build stopped: subcommand failed.
got build log for '/nix/store/pq000fwn348gcig02sqbjrsb5n3f4yj8-tiny-cuda-nn-1.6.drv' from 'daemon'

connor (he/him) (UTC-5) tiny-cuda-nn fails on aarch64-linux with:

-- Build files have been written to: /build/source/build
cmake: enabled parallel building
cmake: enabled parallel installing
Running phase: buildPhase
build flags: -j80
[1/20] Building CUDA object CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o
FAILED: CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o 
/nix/store/z8g6ma876kbi5mxwq388aadn1h35yqy9-cuda-redist/bin/nvcc -forward-unknown-to-host-compiler -ccbin=/nix/store/k0kxskxvmkw97h3z3b5y4hwd56fh9x33-gcc-wrapper-13.3.0/bin/c++ -DTCNN_MIN_GPU_ARCH=60 -DTCNN_SHAMPOO -I/build/source/include -I/build/source/dependencies -I/build/source/dependencies/cutlass/include -I/build/source/dependencies/cutlass/tools/util/include -I/build/source/dependencies/fmt/include -O3 -DNDEBUG -std=c++14 "--generate-code=arch=compute_60,code=[compute_60,sm_60]" "--generate-code=arch=compute_61,code=[compute_61,sm_61]" "--generate-code=arch=compute_70,code=[compute_70,sm_70]" "--generate-code=arch=compute_75,code=[compute_75,sm_75]" "--generate-code=arch=compute_80,code=[compute_80,sm_80]" "--generate-code=arch=compute_86,code=[compute_86,sm_86]" "--generate-code=arch=compute_89,code=[compute_89,sm_89]" "--generate-code=arch=compute_90,code=[compute_90,sm_90]" "--generate-code=arch=compute_90a,code=[compute_90a,sm_90a]" -Xcompiler=-mf16c -Xcompiler=-Wno-float-conversion -Xcompiler=-fno-strict-aliasing -Xcudafe=--diag_suppress=unrecognized_gcc_pragma --extended-lambda --expt-relaxed-constexpr -MD -MT CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o -MF CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o.d -x cu -c /build/source/src/encoding.cu -o CMakeFiles/tiny-cuda-nn.dir/src/encoding.cu.o
nvcc warning : incompatible redefinition for option 'compiler-bindir', the last value of this option was used
g++: error: unrecognized command-line option '-mf16c'
...