also rename methods exchangePostCollDataGPU27() to exchangeCollDataGPU27()

prevents accidental use of the null stream

* move cudaEventRecord to UpdateGrid27
* use variables for stream indices in UpdateGrid27
* create and destroy streams in Simulation.init() and Simulation.free()

Previously, the event was also triggered when data was written. This behavior caused crashes of VF

instead of bc nodes

bulkOnly --> split domain

Thh new behaviour can turned off by passing false for the "onlyBulk" bool

Use level instead of dridNumber in NodeIndicesMultiGPU
and instead of i in loops in MultipleGridBuilder

CUDA doesn't support std::vector

No need to pass node types anymore