topic Re:MKL in Intel® oneAPI Math Kernel Library

MKL

John_B — Fri, 03 Sep 2021 10:28:25 GMT

I have an application which uses pthreads, and those threads call BLAS routines to do number crunching.

At present the application seems to be deadlocking, but this deadlocking does not happen if we use our own BLAS routines (ie code which does the same job but is not part of a library).

We have tried using "MKL Sequential" but the deadlocking persists.

My questions are:

1. Is MKL sequential thread-safe?

2. Does MKL sequential use threads of any type?

Thanks for any inputs that anybody can make!

Re:MKL

Gennady_F_Intel — Sun, 05 Sep 2021 04:25:42 GMT

yes, the all MKL's routines are thread-safe excetp onlu ony exilarly routine.

https://software.intel.com/content/www/us/en/develop/documentation/onemkl-linux-developer-guide/top/managing-performance-and-memory/improving-performance-with-threading.html :

Intel® oneAPI Math Kernel Library is thread-safe, which means that all Intel® oneAPI Math Kernel Library functions (except the LAPACK deprecated routine?lacon)work correctly during simultaneous execution by multiple threads. In particular, any chunk of threaded Intel® oneAPI Math Kernel Library code provides access for multiple threads to the same shared data, while permitting only one thread at any given time to access a shared piece of data. Therefore, you can call Intel® oneAPI Math Kernel Library from multiple threads and not worry about the function instances interfering with each other.

Re:MKL

Gennady_F_Intel — Fri, 10 Sep 2021 08:49:33 GMT

This issue has been resolved and we will no longer respond to this thread. If you require additional assistance from Intel, please start a new thread. Any further interaction in this thread will be considered community only.