Implementation of one-sided communication in Intel MPI

Analyzers

Talk to fellow users of Intel Analyzer tools (Intel VTune™ Profiler, Intel Advisor)

Implementation of one-sided communication in Intel MPI

800 Views

Hi All,

I'm working with The Intel MPI library for compiling an executing an MPI code that uses one-sided calls. I was wondering regarding the implementation details of the Intel MPI Library. Basically, how the MPI library implements the RMA semantic when the low-level network does not support the same semantic. For example, if the network does not support the Accumulate operations (just support Get and Put operations). For MPICH, I read that the implementation uses direct RMA operations for shared memory and relies on an active message-based approach to emulate the RMA operations for other networks like InfiniBand. Does the Intel implementation inherit the same thing? Are there any good reference about the Intel MPI implementation details that I can read?

Best regards

Ahmed

Link Copied

0 Replies

Community support is provided Monday to Friday. Other contact methods are available here.

Intel does not verify all solutions, including but not limited to any file transfers that may appear in this community. Accordingly, Intel disclaims all express and implied warranties, including without limitation, the implied warranties of merchantability, fitness for a particular purpose, and non-infringement, as well as any warranty arising from course of performance, course of dealing, or usage in trade.

For more complete information about compiler optimizations, see our Optimization Notice.