Strategy Tester issues in Version 5 Build 2085

 

The strategy tester of the past versions did a ways better balancing of the load across the Local Agents Farm.

All calculation tasks are now equally distributed to the available cores (95 in the example of below).

The slower agents calculating their tasks while the faster agents are idle and delay the process in a big way. The only workaround is to stop the optimization and to restart it again as the job continues at the point of abortion. Sometimes you have to repeat this workaround several times...

Another worsening is that you have no "live" update of the optimization results. You have to click to refresh the view, and click, and click,...

Martin


 
Taeng:

I'm wondering that the problem is not yet fixed. Fortunately my saved old version 5 Build 2007 from February is working fine as the back testing in the current version is BS...