As noted in #739 (comment) the performance of some tests are significantly slower than others now optimisation have been applied.
Specifically some of the thrust kernels could do with a performance review:
thrust_statevec_calcProbOfMultiQubitOutcome_subthrust_densmatr_calcProbOfMultiQubitOutcome_subthrust_statevec_multiQubitProjector_subthrust_densmatr_multiQubitProjector_sub
This needs to be thought about in detail as functor_insertBits is used with pointers to data passed around in its contractor and we need to be careful of what is expected by thrust in thrust::make_transform_iterator.
As noted in #739 (comment) the performance of some tests are significantly slower than others now optimisation have been applied.
Specifically some of the thrust kernels could do with a performance review:
thrust_statevec_calcProbOfMultiQubitOutcome_subthrust_densmatr_calcProbOfMultiQubitOutcome_subthrust_statevec_multiQubitProjector_subthrust_densmatr_multiQubitProjector_subThis needs to be thought about in detail as functor_insertBits is used with pointers to data passed around in its contractor and we need to be careful of what is expected by thrust in thrust::make_transform_iterator.