Compilation time issue

[Romain Chaffanjon]

Hello everyone, 

I'm trying to achieve a year based resolution, from 2016 to 2050. 

My problem is that, the more period you have the more input data you need.

• For Emissions activity, Efficiency, Cost invest, the number of parameters depends linearly on the number of period. 
• For cost fixed and variable this number depends quadratically on the number of period.

So as the model under development have a lot of variable, it implies heavy sql input file (more than 100 mb), and very long compilation time (Creation of Temoa instance). 

For my last run 3h were needed for compilation and 1h30 for optimization. 

So I'm wondering if there is any possibility to decrease this compilation time. I though it would be great to have the possibility to keep Temoa instance in the memory. Then import, compile and merge specific scenario data with the instance.

But I'm not sure it is realistically feasible. 

Any insights to handle this would be very helpful.

Thank a lot, 

Romain

[Katie Jordan]

Hi Romain,

We don't currently have a good option for reducing compilation time. Reducing model runtime, including model instantiation, is an ongoing goal for our team. If we discover a method that decreases compilation time, we will be sure to let you know.

Best,

Katie