Arm MAP is a unique scalable low-overhead profiler, available standalone or as part of the Arm Forge debug and profile suite. It helps server and HPC code developers to accelerate their software by revealing the causes of slow performance.

It is used from multicore Linux workstations through to supercomputers. You can profile realistic test cases that you care most about with typically under 5% runtime overhead. The interactive user interface is clear and intuitive, designed for developers and computational scientists. It is architected for the biggest clusters and supercomputers - and the biggest and most complex codes! It profiles C++, C, Fortran and Python with no relinking, instrumentation or code changes required.

MAP exposes a wide set of performance problems and bottlenecks by measuring:

• Computation - with self and child and call tree representations over time
• Thread activity - to identify over-subscribed cores and sleeping threads that waste available CPU time for OpenMP and pthreads
• Instruction types (for x86_64) - to show use of vector-units or other performance extensions
• Synchronization, communication and workload imbalance for MPI or multi-process usage
• I/O performance and time spent in I/O - to identify bottlenecks in shared or local file systems