gvoskuilen

@PhilippKasgen That's definitely not intended! If you can get a minimal example that would be great, but it's likely some uncommon race in the coherence protocol so may be hard...

Can you double check that the L2 caches and the directory have their coherence parameter set to MESI right? A mismatch could cause this. If the config is OK, it...

If you are not tied to using Merlin, an alternative would be Kingsley for the network. The high level differences are that Kingsley only supports mesh, doesn't have output buffering...

Also, for Merlin I have done it this way (pseudo-code): class PopulatedPort: def build(self, nodeID, extraKeys): # build component(s) and return the network-connected port class UnpopulatedPort: def build(self, nodeID, extraKeys):...

Are you trying to run this on a specific SST CPU model? Do you have a config file? If not, the simplest method would be to create a python config...

Can you post the Ariel output before the "Ariel Memory Management Statistics:"? Is there an 'ariel_enable()' call in the application you are running?

@aravindneela The above shows that ariel is configured to wait until it finds an 'ariel_enable()' call in the program - "ARIEL: Tool is configured to suspend profiling until program control"....

@aravindneela That output is from HybridSim so I would take a look at their documentation. Four repetitions is probably one from each of the HybridSim instances.

@taniabanerjee Ariel doesn't support applications running MPI because it conflicts with SST's use of MPI. Earlier versions of Ariel haven't supported it either and so your simulations would have been...

@taniabanerjee Yes we were running OpenMP versions of those.