// sharing_weight: variable's impact on the resource during the sharing
// if == 0, the variable is not considered by LMM
// on CPU, actions with N threads have a sharing of N
// on network, the actions with higher latency have a lesser sharing_weight
// sharing_weight: variable's impact on the resource during the sharing
// if == 0, the variable is not considered by LMM
// on CPU, actions with N threads have a sharing of N
// on network, the actions with higher latency have a lesser sharing_weight