About | Features | Technologies | How to use | License | Author | Acknowledgement
The Modular Science Framework provides a low entry threshold software framework to ease the deployment of complex interactive workflows on HPC resources. In particular, it tackles the following two challenges:
- the robust online deployment of a complex workflow comprises multiple applications
- the orchestration of data flows in an interactive and reproducible fashion.
The framework comprises a set of following micro-services like independent applications:
- Provides a handle on the deployment of complex co-simulation workflows and their execution on high-performance computing (HPC) systems.
- Offers steering of the execution of the workflows
- Offers insights on both levels, system (HPC) and individual subsystem (application/simulator), in order to monitor the resource usage (e.g. CPU, Memory).
- Provides health & status information of the whole system and individual applications to detect failures and bottlenecks for effienct use of resources and dubugging purposes with minimal overhead on performance.
- Keeps track of steering and state transitions for provenance tracking.
The following tools were used in this project:
Modular Science Framework v2.0 includes the following driving usecases as demonstration of multi-scale co-simulations. Follow below link to the repository of these usecase for a more detailed description, installation guide and how to run it.
- TVB NEST Co-simulation Usecase 1: A multi-scale neuroscience use-case which connects a spiking neural network with a neural mass model using the simulators NEST and TVB respectively.
- NEST LFPy: This usecase demonstrates how to use the Co-simulation framework to calculate Local Field Potentials (LFPs) in real time, based on spike events streamed from the NEST simulator. The LFP signals are calculated from the spike events by applying the so-called kernel approach, as outlined in Hagen et al. (2022): https://doi.org/10.1371/journal.pcbi.1010353. This usecase can be used as a starting point for applying the Co-simulation framework to other network models in NEST or TVB, and also for simulating other brain signals like EEG, MEG, or ECoG signals.
- Cosim-NEST Desktop-Insite: It demonstrates the seemsless integration of Modular Science with external tools such as NEST Desktop (a web based GUI application) and the Insite pipeline. The integration is based on the TVB NEST Co-simulation Usecase 1 which serves as the foundation. This allows users of NEST Desktop to:
- Create a spiking neural network for NEST simulator.
- Observe the neuronal or network activity with Insite during the simulation.
This project is under license from BSD 3-Clause. For more details, see the LICENSE file.
Made by the Multiscale Simulation and Design team of the Simulation and Data Lab (SDL) Neuroscience at Forschungszentrum Jülich.
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 785907 (HBP SGA2).
This project/research has received funding from the European Union’s Horizon 2020 Framework Programme for Research and Innovation under the Specific Grant Agreement No. 945539 (HBP SGA3). Specific Grant Agreement ‘Interactive Computing E-Infrastructure for the Human Brain Project – ICEI’ (01/2018-03/2023).
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 800858 (HBP SGA ICEI).