Running large simulation tests on GitHub/AWS

[WeinaJi]

In GitLab (BlueConfigs) pipeline, we are running the following integration tests in different scales. If we want to run integration tests in scale for GitHub PRs, maybe on AWS, we need to find the proper circuits and the proper size of node_set for simulation.

• Quick tests
  • neocortex
    • sonataconf-quick-sscx-O1 (SSCx O1 circuit, hex_O1_L4_100, 100 cells)
    • sonataconf-quick-keep-axon (SSCx O1 circuit, hex_O1_L4_100, 100 cells)
    • sonataconf-quick-scx-multi-circuit (custom usecase3, Mosaic_A, 3 cells)
  • ncx_plasticity
    • sonataconf-sscx-v7-plasticity (sscx-v7-plasticity, 200_L5_PCs, 200 cells)
  • ncx_ngv
    • sonataconf-quick-multiscale (ngv, testNGVSSCX_AstroMini, 10 neurons + 9 astrocytes)
  • hippocampus:
    • sonataconf-quick-hip-multipopulation (hip mooc circuit, most_central_100_SP_PC, 100 cells)
  • thalamus:
    • sonataconf-quick-thalamus (public thalamus circuit, small, 14 cells)
• Medium tests
  • neocortex
    • sonataconf-sscx-O1 (SSCx O1, hex_O1_L4_1000, 1000 cells)
    • sonataconf-scx-v5-uhill-conductance-scale (v5_sonata, Column, 1000 cells)
  • hippocampus
    • sonataconf-hippocampus (hippocampus sonata, slice10_sub1000, 1000 cells)
  • thalamus
    • sonataconf-thalamus (public thalamus, mc2_Column_sub1000, 1000 cells)
• weekly long tests
  • neocortex
    • sonataconf-sscx-O1 (SSCx O1, hex0 30K)
  • hippocampus
    • sonataconf-hippocampus (hippocampus sonata, slice10 12K)
  • thalamus
    • sonataconf-thalamus (public thalamus, mc2_Column 14K)

[WeinaJi]

Also as proposed by @ferdonline in https://bbpteam.epfl.ch/project/issues/browse/BBPBGLIB-1171 , it would be good to have a test with a combined circuit (NCX+Thalamus or NCX+Hippocampus) in the future.