Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Modeling energy-efficiency of neuron simulations
Extending existing compute performance models for neuroscience simulations of morphologically detailed neuronal circuits with a model for power consumption.
Modeling and simulation of morphologically detailed neuronal circuits enables us to gain a deeper understanding of biological processes on multiple scales of the brain. In order to study complex phenomena such as neuronal plasticity we need to be able to run simulations at an increased scale, both in space and time.
Understanding the various aspects of a neuron simulation that influence simulation performance and parallel scalability allows us to take decisions on hardware choice and improvements in algorithms and data structures, which are necessary to push the envelope of simulation scale. An important measure that has not been considered for this type of simulations is power consumption.
In this project we would like to extend our performance models for neuroscience simulations with a model for power consumption. Existing power consumption models (e.g. based on ECM) should be explored and extended to fit our case. The developed model can then be incorporated into the performance model, allowing us to give more detailed insights into the expected feasibility and efficiency of simulations on various hardware platforms.
Modeling and simulation of morphologically detailed neuronal circuits enables us to gain a deeper understanding of biological processes on multiple scales of the brain. In order to study complex phenomena such as neuronal plasticity we need to be able to run simulations at an increased scale, both in space and time. Understanding the various aspects of a neuron simulation that influence simulation performance and parallel scalability allows us to take decisions on hardware choice and improvements in algorithms and data structures, which are necessary to push the envelope of simulation scale. An important measure that has not been considered for this type of simulations is power consumption. In this project we would like to extend our performance models for neuroscience simulations with a model for power consumption. Existing power consumption models (e.g. based on ECM) should be explored and extended to fit our case. The developed model can then be incorporated into the performance model, allowing us to give more detailed insights into the expected feasibility and efficiency of simulations on various hardware platforms.
Not specified
This project is hosted at the Prof. Felix Schürmann Group, EPFL. To learn more, please visit https://www.epfl.ch/labs/gr-fsch/
If you are interested please contact Omar Awile <omar.awile@epfl.ch> .
This project is hosted at the Prof. Felix Schürmann Group, EPFL. To learn more, please visit https://www.epfl.ch/labs/gr-fsch/
If you are interested please contact Omar Awile <omar.awile@epfl.ch> .