Since I made a similar pull requests to FastBitSet and TypedFastBitSet, I updated the benchmarks to include the new methods, and also found a couple of small bugs and opportunities for optimization. Interestingly, the new in-place method difference2 appears to be faster than the original difference.
FastBitSet (inplace) x 3,414 ops/sec ±0.57% (98 runs sampled)
FastBitSet (inplace2) x 3,702 ops/sec ±0.28% (99 runs sampled)
@lemire the new benchmarks for TypedFastBitSet will not work until the NPM repository is updated. thanks!!!
Since I made a similar pull requests to FastBitSet and TypedFastBitSet, I updated the benchmarks to include the new methods, and also found a couple of small bugs and opportunities for optimization. Interestingly, the new in-place method
difference2
appears to be faster than the originaldifference
.@lemire the new benchmarks for
TypedFastBitSet
will not work until the NPM repository is updated. thanks!!!