Mario Rugiero
Mario Rugiero
Maybe those tests just don't produce run containers 🤔
Indeed. I'll add `RunOptimize` and try again.
Even forcing denser arrays I saw no gains.
They get created by adding 65k random numbers in the range 0-100000, the same for both, then optimize one. I can try with a larger one. Probably the addition is...
Maybe the result was too dense and became mostly bitmaps in the non-run-optimized version now that I think of it. I can print the stats anyway.
Yes, I stalled work on this for now, but the issue is that it doesn't know how to find the input when I build the tests in a separate step...
As @lemire pointed out, frozen bitmaps are not meant to be modified. The name attempts to make that clear. It could maybe handle that better or behave as CoW, but...
I think something like this should work (I wrote a dummy version but right now it crashes): - Allocate a slice for indices for each bitmap. - (Optionally) Sort from...
A more sophisticated version would use shotgun search for the current key, as well as skipping to the maximum found from the last iteration. But that's a reasonable draft I...
It works but doesn't change a bit 🤷 However, using a pool for the containers slice does decrease allocations and memory use by 7-8% in the benchmarks, without an execution...