Cannot execute CUDA clock example

Hello all,
I installed CUDA just now on my fedora 10 box.

I am trying to execute the clock example.
Heres what i did till now.

  1. i went inside the clock dir.
    2.executed make
    3.i saw “obj” dir created and some files.
  2. i went inside the “obj/release” dir
    5.i see “clock.cu.o” object file
    6.i changed the permission of this file
    7.then i typed “./clock.cu.o” and hit return
    8.it says : bash: ./clock.cu.o: cannot execute binary file

Where am i going wrong.
Thank you.

You are trying to run an object file, which is never going to work. The SDK executables are installed in NVIDIA_CUDA_SDK/bin/linux/release

Thank you for the reply.

You are trying to run an object file, which is never going to work. The SDK executables are installed in NVIDIA_CUDA_SDK/bin/linux/release

Yup got that.
So i tried executing the clock file in /root/NVIDIA_CUDA_SDK/bin/linux/release
But it said : cannot find libcudart.so.2 file.
So i deleted the NVIDIA_CUDA_SDK dir from /root
Reinstalled in the /usr/local/cuda (toolkit installation dir)
But now when i try to run make inside …/projects/clock dir, it says:
/usr/bin/ld: cannot find -lcutil
collect2: ld returned 1 exit status
make: *** […/…/bin/linux/release/clock] Error 1

Now whats the problem.

Thanks again

I ran make again inside /usr/local/cuda
then i ran make inside /usr/local/cuda/projects/clock
then i tried to execute clock inside /usr/local/cuda/bin/linux/release
i got:
./clock: error while loading shared libraries: libcudart.so.2: cannot open shared object file: No such file or directory

This is my bash_profile file:

.bash_profile

Get the aliases and functions

if [ -f ~/.bashrc ]; then
. ~/.bashrc
fi

User specific environment and startup programs

PATH=$PATH:$HOME/bin

export PATH

PATH=$PATH:/usr/local/cuda/bin
LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib ldd ./clock
export PATH
export LD_LIBRARY_PATH

I can see ibcudart.so.2 inside /usr/local/cuda/lib

This is the same thing i was getting earlier.

Thanks

After some tweaks, i dont get that error.
however, i get this now :

cudaSafeCall() Runtime API error in file <clock.cu>, line 92 : no CUDA-capable device is available.

but i do have a nvidia card (GeForce 6150 LE) and nvidia driver (NVIDIA-Linux-x86-180.44-pkg1.run) installed
the driver was not installed with the CUDA, however.

Now whats the problem?

Thank you.

This is not going to help one bit, but I’m having the same problem. I’ve installed CUDA on my machine and everything was looking peachy, even the deviceQuery returns “There is 1 device supporting CUDA” but all other examples fail with the same error, " no CUDA-capable device is available. ". Now, I’ve tried it again and some examples work while others just exit with an “out of memory” message.
It is weird, because I found a guy on the Internet posting the same issue, with programs working and not working with no apparent reason to that.
Btw, I have an ancient Geforce 8400m gs on my 1420 Inspiron Dell Laptop and I’m using Ubuntu 9.04 x64. I may give it a try on Vista 64 but I was hoping on playing around with CUDA on Linux. I’m using 190.18 drivers with cuda 2.3. Any ideas?