Need help with parallel debugging

[Kiny Wan]

Dear all,

        Recently, I have many questions about PyFR since I try to read and understand the context of codes. The structure and idea of PyFR is attractive, but difficulities remain exist. I successfully achieve serial debugging with pycharm in ubuntu16, but unfortunately I am failed to debug the code with MPI. The debugging of cuda and opencl might be more difficult. I hope you can provide some experiences in detail about how to parallel debugging during studying or developing the PyFR? To my best knowledge, the available tools for C++/fortran parallel debugging is totalview and ddt. 

        In my opinion, PyFR has great potential and can be combined with artificial intelligence and future quantum computing. So I hope to learn from and later contribute to this open source. 

Regards,

Kiny

[Freddie Witherden]

PyFR was developed entirely (both in serial and parallel) without the use of any debuggers.  In general it is much easier to get it right the first time than faff around with a debugger; this is especially true when one is dealing with GPUs, multiple threads, or anything to do with MPI.

Regards, Freddie.

On 3 Jul 2020, at 10:01, Kiny Wan <wlc...@gmail.com> wrote:



--
You received this message because you are subscribed to the Google Groups "PyFR Mailing List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pyfrmailingli...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/pyfrmailinglist/aea4e7ac-a376-4543-8543-708f9c009e4eo%40googlegroups.com.

[Kiny Wan]

Thank you, Freddie.  I will study the code structure with serial debugging and follow your way to learn from the rest.