3
$\begingroup$

I have done a ParallelTable computation in script mode. After execution, the output is fine and I've got all the data I needed. but the kernel seems pending, usually it should exit after they finished job.

I am working on a workstation with 16 cores, I used 14 cores to do the job. The kernel has pending almost a day, and one cores is fully loaded. Here is the pending message, I don't know what does it mean.

Launching kernels...

LinkObject::linkd: Unable to communicate with closed link LinkObject["/usr/local/Wolfram/Mathematica/10.3/Executables/wolfram" -subkernel -noinit -wstp, 97, 5].

Kernels::rdead: Subkernel connected through KernelObject[5, local] appears dead.

Parallel`Developer`QueueRun::req: Requeueing evaluations {81} assigned to KernelObject[5, local, <defunct>].

LaunchKernels::clone: Kernel KernelObject[5, local, <defunct>] resurrected as KernelObject[15, local].

LinkObject::linkd: Unable to communicate with closed link LinkObject["/usr/local/Wolfram/Mathematica/10.3/Executables/wolfram" -subkernel -noinit -wstp, 318, 5].

Kernels::rdead: Subkernel connected through KernelObject[15, local] appears dead.

Parallel`Developer`QueueRun::req: Requeueing evaluations {81} assigned to KernelObject[15, local, <defunct>].

LaunchKernels::clone: Kernel KernelObject[15, local, <defunct>] resurrected as KernelObject[16, local].

LinkObject::linkd: Unable to communicate with closed link LinkObject["/usr/local/Wolfram/Mathematica/10.3/Executables/wolfram" -subkernel -noinit -wstp, 106, 14].

General::stop: Further output of LinkObject::linkd will be suppressed during this calculation.

Kernels::rdead: Subkernel connected through KernelObject[14, local] appears dead.

General::stop: Further output of Kernels::rdead will be suppressed during this calculation.

Parallel`Developer`QueueRun::req: Requeueing evaluations {81} assigned to KernelObject[14, local, <defunct>].

General::stop: Further output of Parallel`Developer`QueueRun::req will be suppressed during this calculation.

LaunchKernels::clone: Kernel KernelObject[14, local, <defunct>] resurrected as KernelObject[17, local].

General::stop: Further output of LaunchKernels::clone will be suppressed during this calculation.
$\endgroup$
  • 1
    $\begingroup$ The message says that "the kernel appears dead". This usually happens when the subkernel crashes (you can verify if it does). It appears that Mathematica automatically launches a new kernel to take the place of the dead one. If the subkernels area really crashing (likely yes), the next step is to find out why they crash, what code causes the crash. Also verify that you're not running out of memory. $\endgroup$ – Szabolcs Jan 8 '16 at 12:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.