Is more verbosity possible?

General
1

Hello,

I am still playing with SD7003 benchmark. When it runs, it scales really well.
I will share my results in another thread when they're ready, but I am
experiencing some strange freezing behaviors. I am almost sure this is down to
our cluster or software stack and I will be getting our system team to help me
with this on Wednesday, but I was wondering if you can suggest something to
make pyfr more verbose.

I already print everything I can from MPI and I use --verbose flag in pyfr. I
also print residuals ever 10 steps which is the only way for me to see that the
job has frozen.

Best wishes,
Robert

2

Hi Robert,

I do not believe that anything in PyFR currently makes use of the
–verbose flag. However, your best best upon noticing that the code has
frozen is to attach GDB and get a backtrace. GDB is quite good these
days about understanding Python stack frames so you’ll be able to see
exactly what is going on.

Regards, Freddie.