JuliaDynamics / ABMFrameworksComparison

Benchmarks and comparisons of leading ABM frameworks
Other
13 stars 8 forks source link

Speed-up NetLogo benchmarks start-up #47

Closed Tortar closed 1 year ago

Tortar commented 1 year ago

We shouldn't start-up netlogo-headless for each repetition because it takes nearly 20 secs each time as discussed in #46.

We could do so by providing all seeds at once at start-up as it is done here with the variable number:

<experiments>
  <experiment name="understand_xml" repetitions="1" runMetricsEveryStep="true">
    <setup>setup</setup>
    <go>go</go>
    <metric>count turtles</metric>
    <enumeratedValueSet variable="color-by">
      <value value="&quot;success&quot;"/>
    </enumeratedValueSet>
    <enumeratedValueSet variable="memory">
      <value value="6"/>
    </enumeratedValueSet>
    <enumeratedValueSet variable="number">
      <value value="593"/>
      <value value="344"/>
      <value value="442"/>
    </enumeratedValueSet>
    <enumeratedValueSet variable="strategies-per-agent">
      <value value="5"/>
    </enumeratedValueSet>
  </experiment>
</experiments>

and we should also change the way timings are retrieved from the .txt file created by Netlogo due to the new way of doing the benchmark.