VTK-m compilation uses too much memory
While building VTK-m (in the context of ParaView), memory usage goes up to the roof, preventing me to use all my CPUs in order to avoid swapping.
Is this expected ? If it is, would there be a way to reduce the parallelization of only VTK-m in order to be able to use all my CPU when compiling the rest of ParaView/VTK ?
Usually, one VTK-m compilation thread uses 2Gb of memory.
Edited by Mathieu Westphal (Kitware)