I am trying out some sample dpcpp programs on the intel devcloud. While creating a queue am able to create for default_selector which gives the Intel Xeon Gold CPU as the target interface. And when using accelerator_selector, it gives Intel FPGA Emulation Device. But when trying to give gpu_selector, it fails to find a device (gets a sycl::_V1::runtime_error with PI_ERROR_DEVICE_NOT_FOUND). Is there a way to request for a devcloud node with a gpu ? I tried running the 'qsub -I' to get the cpu with 24 cores and the fpu devices but doesn't seem to work for the gpu. Appreciate any help on this.
Thank you for posting Intel communities.
By using qsub -I command you will enter in interactive node and this command generates random node numbers. You can try the below command.
qsub -I -l nodes=1:gpu:ppn=2 -d .
-I (upper case i) is used to request an interactive session.
-l (lower case L) nodes=1:gpu:ppn=2 is used to assign one full GPU node to the job.
-d is used to configure the current folder as the working directory for the task.
job.sh is the script that gets executed on the compute node.
Also refer the below links :
Glad to know that your issue is resolved. If you need any additional information, please post a new question as this thread will no longer be monitored by Intel.