Wanted to know if oneMKL performs the operations in parallel mode. For example, should I implement my own parallel matrix-matrix multiplication in SYCL or should I use the already implemented oneMKL gemm or other routines (if they work already in parallel mode)?
MKL has two versions of its library: sequential and threaded
The sequential version is thread-safe and uses 1 thread internally for MKL functions.
The threaded version is also thread-safe but uses its own thread pool (multi-threaded internally). It is the programmer's responsibility not to oversubscribe threads or at least not to have the over-subscription exhibit unnecessary interference.
The issue is closing 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.