Snippets Groups Projects

Compare History

Add some convenience to simulation creation. (#1904)

Thorsten Hater authored 2 years ago

Encapsulates 80% of cases:
```cxx
 rec = recipe()
 ctx = make_context()
 dec = partition_load_balance(rec, ctx)
 sim = simulation(rec, dec, ctx)
```
is now written as
```cxx
 rec = recipe()
 sim = simulation(rec)
```
In python we use keyword args to allow
both to be specified separatly.

Partially fixes #1862

d15afd39

Name	Last commit	Last update
..
CMakeLists.txt
bench.cpp
readme.md

Artificial Infrastructure Benchmark

This miniapp uses the arb::benchmark_cell to build a network of cells that have predictable cell integration run time and spike generation patterns. It is for benchmarking purposes only.

The main application of this miniapp is to benchmark the Arbor simulation architecture. The architecture is everything that isn't cell state update, including

Spike exchange.
Event sorting and merging.
Threading and MPI infrastructure.

For example, the scaling behavior of the spike exchange can be studied as factors such as fan-in, network minimum delay, spiking frequency and spiking pattern are varied, without having to tweak parameters on a model that uses "biologically realistic" cells such as LIF cells.

Usage

The model can be configured using a json configuration file:

./bench params.json

An example parameter file is:

{
    "name": "small test",
    "num-cells": 2000,
    "duration": 100,
    "fan-in": 10000,
    "min-delay": 10,
    "spike-frequency": 20,
    "realtime-ratio": 0.1
}

The parameters in the file:

name: a string with a name for the benchmark.
num-cells: the total number of cells in the model. The cell population is assumed to be homogoneous, that is the spike-frequency and cell-overhead parameters are the same for all cells.
duration: the length of the simulated time interval, in ms.
fan-in: the number of incoming connections on each cell.
min-delay: the minimum delay of the network.
spike-frequency: frequency of the independent Poisson processes that generate spikes for each cell.
realtime-ratio: the ratio between time taken to advance a single cell in the simulation and the simulated time. For example, a value of 1 indicates that the cell is simulated in real time, while a value of 0.1 indicates that 10s can be simulated in a single second.

The network is randomly connected with no self-connections and fan-in incoming connections on each cell, with every connection having delay of min-delay.