We want to benchmark how quickly we can select out (e.g. array[1:10]) of and materialize the full altrep vector (i.e. array[]). We should write benchmarks that do both of these so we can measure them.
Additionally, we should confirm that the more macro benchmarks that can are taking advantage of altrep to see how it helps real workflows.
We want to benchmark how quickly we can select out (e.g.
array[1:10]) of and materialize the full altrep vector (i.e.array[]). We should write benchmarks that do both of these so we can measure them.Additionally, we should confirm that the more macro benchmarks that can are taking advantage of altrep to see how it helps real workflows.