0
votes

I have a Server running CentOS 6.0 and I'm trying to use it as a remote host for cuda debugging. In order to do this, I installed cuda-toolkit 5.5 both on the server and my notebook, which is running ubuntu 12.10 OS. I configured the two machines as the NVIDIA-cuda-instruction told me, yet when I started the Nsight eclipse edition and tried to remote debug my cuda applications, I ran into error, whcih says:

Failed to execute MI command:

-target-select remote 192.168.2.105:2345

Error message from debugger back end:

192.168.2.105:2345: Connection timed out

I googled this error, someones says it is a firewall issue and I need to open the port manually, so I modified the /etc/sysconfig/iptables file and add this:

-A INPUT -m state --state NEW -m tcp -p tcp --dport 2345 -j ACCEPT

to open the 2345 port. But it turn out to be of no use. Can anyone give me some suggestion and help me solve this problem?

By the way, I have tested local debugging using nsight eclipse edition on my server, and it works well.

1
You don't need to post a new question, just edit the old one. - Robert Crovella
This indeed seems like a firewall issues - it seems like firewall is enabled by default in newer RedHat derived distros. Please try remote debugging from the command line (you may even try regular GDB/GDBSERVER instead of CUDA ones). - Eugene
I changed the debug port(7777) and there is a new error. Now the console does not tell me that the connection timeed out, but it tells me something like: can't find file /home/***/cuda-workspace/.../a.cu - Xiangyu.Guo

1 Answers

0
votes

I have the same problem in an OpenSuse server. I have managed to make it work by disabling the firewall in the server.

#/sbin/rcSuSEfirewall2 stop

You can start later by

#/sbin/rcSuSEfirewall2 start