nestmc
play

NestMC A new multi-compartment neuronal network simulator Alexander - PowerPoint PPT Presentation

Apr 24, 2023 •276 likes •614 views

NestMC A new multi-compartment neuronal network simulator Alexander Peyser (FZ-J) & Sam Yates (CSCS) November 3, 2016 NestMC NestMC is a project to develop: a new multi-compartmental neuronal network simulator, that is optimized for HPC

  1. NestMC A new multi-compartment neuronal network simulator Alexander Peyser (FZ-J) & Sam Yates (CSCS) November 3, 2016

  2. NestMC NestMC is a project to develop: a new multi-compartmental neuronal network simulator, that is optimized for HPC systems, and is easy to integrate into existing workflows. See current development at https://github.com/eth-cscs/nestmc-proto NestMC | 2

  3. Who are we? Cross-institutional collaboration As part of NestMC | 3

  4. Why? Why develop a new simulator? There are problems and models that we can’t explore with current software and systems. New HPC architectures. Adapting existing simulators to new architectures is hard . NestMC | 4

  5. Hard problems Examples Near real-time multi-compartment simulations. ‘Large’ networks: long simulations, parameter search, statistical validation. Field potential calculations: large networks, volume visualization. NestMC | 5

  6. New architectures Processor clock speed growth suddenly slowed around 2004. 10’000 1’000 frequency (MHz) 100 10 1 0.1 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 year power ∝ frequency 3 Problem: NestMC | 6

  7. New HPC architectures New performance gains primarily from: Highly parallel architectures (e.g. Intel KNL). Wider vector operations (e.g. AVX512). Specialized accelerator hardware: GPU, FPGA. NestMC | 7

  8. New HPC architectures Prototype Human Brain Project HPC systems at J¨ ulich Julia Juron Intel many core KNL blade IBM Power8+GPU ‘fat’ node Efficient use demands new approaches. We no longer get good performance improvement ‘for free’. NestMC | 8

  9. Prototype design Modular: components can be substituted according to internal API. Internal API: ‘thin’ API; type parameterization allows components to determine low-overhead API data structures. CPU implementation cell simulation GPU model implementation model description execution (NMODL & loop MPI recipes) implementation spike exchange thread parallel implementation API API API NestMC | 9

  10. Prototype design — backends Cell simulation modules share computational backends for channel and synapse state evolution. CPU-hosted finite volume cell simulation CPU scalar kernels F.V.M. solver NMODL CPU vector kernels specifications F.D. solver GPU kernels API API NestMC | 10

  11. Prototype benchmarks Test case 500 ms simulation. Each cell has 350 compartments and 2000 exponential excitatory synapses. H–H mechanism on cell somas, passive dendrites. Random network. Approximately 50 Hz spiking rate. Benchmarks run on Pitz Dora , a Cray XC-40 system with 36 Broadwell cores per node. NestMC | 11

  12. Prototype benchmarks — strong scaling 10000 147,456 cells 18,432 cells 70’03” – 1.17 nh 1000 wall time (s) 18” – 1.25 nh 100 8’45” – 0.146 nh 10 5” – 0.158 nh 1 1 2 4 8 16 32 64 128 256 nodes NestMC | 12

  13. Prototype benchmarks — weak scaling 275 2,359,296 cells 9,216 cells 270 wall time (s) 265 260 255 1 2 4 8 16 32 64 128 256 nodes NestMC | 13

  14. Prototype status Currently implemented Finite-volume based discretization. Distributed model instantiation. Spike and voltage trace output. x66 multi-core and Intel KNL support. Synapse and ion-channel descriptions in NMODL. Unit and validation testing suite. GPU support NestMC | 14

  15. How does this relate to NEST? Expertise: developers and experience from NEST goes into NestMC and what we learn from NestMC feeds back to NEST Infrastructure: Community and legal infrastructure can be leveraged for multiple products Interface: similar Python interfaces can reduce time for users to use multiple tools NestMC | 15

  16. How does this relate to NEST? Components: libraries such as for connectivity can be shared between projects Formats: commonality of formats and communications such as NESTml and I/O formats Multiscaling: hybrid simulations may be built across the spectrum from neural mass models to point models to compartment models... NestMC | 16

  17. How is this different from NEST? Models: solving point model ODEs is very distinct from large coupled compartment models Performance: NEST’s point neurons are memory bound, NestMC is computationally bound Flexibility: NEST commits to maximum flexibility across platforms — NestMC is HPC focused on a subset of the “easy” 90% of cases Science: NEST is a highly reduced model for maximizing the size of simulations which are most amenable to mathematical analysis, while NestMC will be useful for morphologically detailed simulations NestMC | 17

  18. Thank you! CSCS: Ben Cumming, Vasileios Karakasis, Stuart Yates FZ-J: Wouter Klijn BSC: Ivan Martinez bcumming@cscs.ch Contact a.peyser@fz-juelich.de https://eth-cscs.github.io/nestmc https://github.com/eth-cscs/nestmc-proto

  19. Participate Success depends on facilitating use cases! Do you have a use case which is hard to simulate with current tools? Are there computational experiments that you wish to run, but currently cannot? We want to work closely with researchers and research groups to ensure that our designs meet real needs in the community. NestMC | 18

  20. The limits of frequency Microprocessor transistor counts have grown exponentially over a very long time frame. 10’000’000’000 1’000’000’000 transistors 100’000’000 10’000’000 1’000’000 100’000 10’000 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 year NestMC | 19

  21. The limits of frequency Processor clock speed growth suddenly slowed around 2004. 10’000 1’000 frequency (MHz) 100 10 1 0.1 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 year power ∝ frequency 3 Problem: NestMC | 20

  22. The limits of frequency Single-thread performance growth also slows. normalized single-thread specFP 300 100 30 10 3 1 0.3 0.1 0.03 1995 2000 2005 2010 2015 year Analysis thanks to Jeff Preshing, http://preshing.com NestMC | 21

  23. Why? Why develop a new simulator? There are problems and models that we can’t explore with current software and systems. New HPC architectures. Adapting existing simulators to new architectures is hard . NestMC | 22

  24. Existing simulators N EURON and G ENESIS have had a very long development. N EURON in particular is very large with many features. Newer simulators such as M OOSE and Brian still primarily target the workstation. GPU support for these simulators still under development. Adapting existing large applications to highly parallel and hardware-accelerated architectures is non-trivial. NestMC | 23

  25. Opportunity A new development project can: target contemporary and future HPC architectures, be co-designed to faciliatate new and difficult use cases. NestMC | 24

  26. Opportunity A new development project can: target contemporary and future HPC architectures, be co-designed to faciliatate new and difficult use cases. In addition, much easier to adopt modern software development processes from the start. NestMC | 24

  27. NestMC Two year initial project to design and develop a multi-compartmental simulator for HPC systems. Goals Interoperability Extensibility Performance Simulator as library Modular internal API HPC targeted Highly parallel Visualization New integration schemes GPU and vector Multi-physics targets Custom spike Multi-scale communication Design for scalability Specialized cells NestMC | 25

  28. Development timeline Now Prototype development. 12/2016 Wind-up prototype. Finalize design of initial release. 4/2017 First public release of simulator. Move to open development model. NestMC | 26

  29. Prototype The prototype development allows us to explore the design space. “Plan one to throw one away” — Fred Brookes Currently implementing features and use cases to refine our design: LFP live visualization → interoperability features. Gap junctions → extensibility, internal API design. GPU execution → performance, modularity. NestMC | 27

  30. Prototype design Modular: components can be substituted according to internal API. Internal API: ‘thin’ API; type parameterization allows components to determine low-overhead API data structures. CPU implementation cell simulation GPU model implementation model description execution (NMODL & loop MPI recipes) implementation spike exchange thread parallel implementation API API API NestMC | 28

  31. Prototype design — backends Cell simulation modules share computational backends for channel and synapse state evolution. CPU-hosted finite volume cell simulation CPU scalar kernels F.V.M. solver NMODL CPU vector kernels specifications F.D. solver GPU kernels API API NestMC | 29

  32. Prototype observations Design Abstractions over communication and threading → greatly simplified testing and validation. Component architecture → rapid prototyping, → use-case driven API changes are limited in scope. Functional model description → reproducible across different systems, → distributed instantiation. NestMC | 30

  33. Prototype observations Development practices Unit and validation testing → limits exposure to bugs and design errors. Large matrix of compilers and hardware targets → continuous integration a necessity ‘Agile’ iterative refinement of development practices allows us to adapt our processes to our team distributed across multiple institutions and countries. NestMC | 31

Recommend

More recommend