NEST:

- Install NEST with hdf5 load module implementation
	- Checkout Mike2 branch from nest repo
		https://github.com/tillschumann/nest-simulator.git
	- Adapt library pathes in the first lines in build-nest.sh (inside) the set ones are valid for bbpbg1.cscs.ch
	- Adapt NEST file path and build folder
	- Build NEST using script: build-nest.sh
	- NEST is installation is now placed in INSTALL_PATH (from build-nest.sh script)
	- To use NEST, all necessary paths can be loaded via: 'source INSTALL_PATH/bin/nest_vars.sh'

How to run a nest simulation:

	nest script.sli (resp. srun nest script.sli)

	script.sli contains the full simulatoin workflow. (http://www.nest-simulator.org/neural_simulations/)
	
	How to use neuron and synapse load module see thesis.
	

NEST Models:
	https://github.com/tillschumann/models

	Needed changes: Adapt nest install path in slurm scripts (eventually adapt batch script to given cluster)

	Info: Sli script and execute script is nessecary.
		Because nest parameter is still changing, look at newest update. to adapt script to it.
		Currently it is 1per300mouse
	     In the sli file the location of the circuit has to be adapted if it changes.
	     Further the number of neurons is hardcoded inside.
	     Before running make sure specified output folder is existing.


GenBrain:
	 Info: Generate Ciruit with long range and short range connectivity

	- See manual for Csaba

Tools:

- Paraview
	
	Used to visulize voxelized densities in 3D space
	The data can be loaded from MHD files.
	MHD files can be generated via BBP viz tools are my python scripts
	
	

Python scripts for analysis during thesis:
	https://github.com/tillschumann/thesis.git
	
	located in subfolder: scripts:



	Merging of gdf spike files written by nest spikedetector
	- mergeSpikeDetectors.py
	  Input: path, number of spiking neurons
	  Output: neurons.npy spikes.npy

	  Info: Set number of spiking neurons too high and run the script. The last number which is displayed is the correct one.
		 If the number is too low the run aboards.

	Visualize own written trace files (ex. traceFile_000.csv)
	- viewTraceFile.py
	  Input: path to file, number of shown ranks
	  Output: matplotlib windows

	  Info: You can see scatter plots, showing the ranks on the y axis and iterations on the x axis. The color shows the measured timing in seconds


	Plotting nest memory consumption based on "Spiking network simulation code for petascale computers"
	- calc_Mem_consumption_Circuit.py
	- calc_Mem_consumption_Nest.py


	Easy voxelizing scripts to visualize point neurons with paraview.
	Scripts only creates raw data. mhd has to be added by hand.
	- voxelizeSpikingNeurons.py
	- voxelizeTargetNeurons.py

	Info: ex. mhd file content: 	ObjectType = Image
					NDims = 3
					BinaryData = True
					BinaryDataByteOrderMSB = False
					CompressedData = False
					TransformMatrix = 1 0 0 0 1 0 0 0 1
					Offset = 0 0 0
					CenterOfRotation = 0 0 0
					AnatomicalOrientation = RAI
					ElementSpacing = 100 100 100
					DimSize = 133 81 115
					ElementType = MET_INT
					ElementDataFile = filename.raw



	Create spiking plots from neurons.npy spikes.npy files
	- spiking.py
	  
	  Info: Specify input path on line 82 and 83.
		Included functions like cut, sort, compare, dotplot can be used to interact with data


Visualizing tools from bbp viz team:
	 git: ssh://[email protected]/viz/Fivox
	      branch: origin/allen_loader

	      Example usage written in doc/UserGuide.md in repo 


Circuit analysis repo (an example):

	Used to perform easy statistical analysis on the big dataset
	- https://github.com/tillschumann/STATS_SYN.git
		 H5SynapseLoader.h Class to load synapses (extracted from NEST synapse load module)
		 NESTNodeSynapse.* Class to store synapse information in memory
		 main.cpp Applying easy math operation to all synapses iterativly	 



Used Cluster:

- bbplviz  bbpviz1.cscs.ch
- BGQ      bbpbg1.cscs.ch
- bbpviz   bbplviz1.cscs.ch 
- juqueen  juqueen.fz-juelich.de