Cuda Samples Failed Compilation

Robotics & Edge Computing Jetson & Embedded Systems Jetson AGX Orin
1

Hello Nvidia Support,
I have just re-flashed my Jetson AGX ORIN 64G with latest jetpack 6.2 and cuda 12.6 and while compiling lates cuda samples like /home/raphs/projects/Nvidia/cuda-samples/Samples/5_Domain_Specific/nbody/ it throws an error

make
[ 20%] Building CUDA object CMakeFiles/nbody_opengles.dir/bodysystemcuda.cu.o
nvcc fatal : Unsupported gpu architecture ‘compute_101’
make[2]: *** [CMakeFiles/nbody_opengles.dir/build.make:76: CMakeFiles/nbody_opengles.dir/bodysystemcuda.cu.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:83: CMakeFiles/nbody_opengles.dir/all] Error 2
make: *** [Makefile:91: all] Error 2

It used to work, I am sure there are some changes how to get this build.
Please if you can help me. I appreciate that.
NOTE : I managed to build with workaround like playing with flags set(CMAKE_CUDA_ARCHITECTURES 50 52 60 61 70 72 75 80 86 87 89 90 100 101 120).
BUT I really appreciate the basic reason and what I should be officially doing…

5

Hi,

Please check out the v12.5 branch and try it again.

$ cd Samples/5_Domain_Specific/nbody
$ sudo apt-get install freeglut3 freeglut3-dev
$ make
/usr/local/cuda/bin/nvcc -ccbin g++ -I../../../Common -m64 -ftz=true --threads 0 --std=c++11 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_90,code=sm_90 -gencode arch=compute_90,code=compute_90 -o bodysystemcuda.o -c bodysystemcuda.cu
/usr/local/cuda/bin/nvcc -ccbin g++ -I../../../Common -m64 -ftz=true --threads 0 --std=c++11 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_90,code=sm_90 -gencode arch=compute_90,code=compute_90 -o nbody.o -c nbody.cpp
/usr/local/cuda/bin/nvcc -ccbin g++ -I../../../Common -m64 -ftz=true --threads 0 --std=c++11 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_90,code=sm_90 -gencode arch=compute_90,code=compute_90 -o render_particles.o -c render_particles.cpp
/usr/local/cuda/bin/nvcc -ccbin g++ -m64 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_90,code=sm_90 -gencode arch=compute_90,code=compute_90 -o nbody bodysystemcuda.o nbody.o render_particles.o -L/usr/lib/nvidia-container -lGL -lGLU -lglut
mkdir -p ../../../bin/aarch64/linux/release
cp nbody ../../../bin/aarch64/linux/release

Thanks.

6

Thanks this works, I am interested to understand supported and unsupported GPU compute, Does this mean future release will not support this any more or there have to be an older branch ?

7

Hi,

The error indicates GPU driver in JetPack 6.2 cannot recognize the 101 GPU architecture.
This is expected as SM_101 is a new architecture just added in CUDA 12.8 last month.

Thanks.

9

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.