Optimizing NEURON Simulation Environment Using Remote Memory Access with Recursive Doubling on Distributed Memory Systems.

Danish Shehzad,Zeki Bozkuş

doi:10.1155/2016/3676582

Abstract

Increase in complexity of neuronal network models escalated the efforts to make NEURON simulation environment efficient. The computational neuroscientists divided the equations into subnets amongst multiple processors for achieving better hardware performance. On parallel machines for neuronal networks, interprocessor spikes exchange consumes large section of overall simulation time. In NEURON for communication between processors Message Passing Interface (MPI) is used. MPI_Allgather collective is exercised for spikes exchange after each interval across distributed memory systems. The increase in number of processors though results in achieving concurrency and better performance but it inversely affects MPI_Allgather which increases communication time between processors. This necessitates improving communication methodology to decrease the spikes exchange time over distributed memory systems. This work has improved MPI_Allgather method using Remote Memory Access (RMA) by moving two-sided communication to one-sided communication, and use of recursive doubling mechanism facilitates achieving efficient communication between the processors in precise steps. This approach enhanced communication concurrency and has improved overall runtime making NEURON more efficient for simulation of large neuronal network models.

Highlights

The brains complex computational behavior necessitated developing large neuronal computational models
The advantages of assorted simulation environments are that each simulator has a broad range of potency and this miscellany contributes to better development and understanding of large neuronal models simulation processes
Message Passing Interface (MPI) Allgather is enhanced to Remote Memory Access (RMA) Allgather one-sided communication using recursive doubling for efficient spikes exchange between processors

Summary

Introduction

The brains complex computational behavior necessitated developing large neuronal computational models. The valuable aspect of diverse simulation environment is its sundry nature and wide-ranging strengths enabling better understanding of computational behavior of neuronal networks. This diversity has resulted in improving the simulating environments capability of computations unfolding the novel perspectives in overall computation and simulation technology. One of the key beneficial features is that fast and efficient architectures of computers can be achieved by the help of these computational simulation environments and neuronal models These have the ability to provide parallel, speedy, and efficient processing. The NEURON was extended from single CPU to multiple CPUs to support complex computational models simulation It can run parallel simulations on small clusters with 10–50 processors to large scale Blue Gene Super Computer with thousands of processors [9]. Multiple processors use MPI Allgather collective communication method to exchange spikes between processors [9]

Methods

Results

Conclusion