Benchmarks

In this section we provide some benchmarks about the execution times for some simulation cases. This information is both to give you an idea of how long your simulations may take and to show how variable can the running times be depending on your hardware resources.

If you time your own simulations do not hesitate to send us your reports! As a general note, if you are running simulations for benchmark purposes, check if there are other processes that might be competing for resources. That might result in biased results. Also do not forget to provide the system info:

  • Operating System.

  • Processor.

  • Memory Available.

  • TVB version.

  • Web Browser (if applicable).

  • Python version (if applicable).

Timeline

Performance evolution from a TVB version to another.

../../_images/benchmarks-neotraits.png

Only tvb-library Version 1.* vs 2.0 (* neo) shows a factor 2 improvement here.

../../_images/benchmarks-evolution-mac.png

Recent versions - Mac station. Version 2.0 does not do very good with storage attached, but this was expected, we will work on optimizing this in future releases

../../_images/benchmarks-evolution.png

Older Versions - Linux station

../../_images/benchmarks-evolution.png

Recent Version - Mac station