The revised thin_ls for the upcoming thinp-tools v1.3.1 adopts the new Aggregator. While this component successfully improves parallelism, its current memory consumption is significantly higher than the previous implementation. For example, in the 16GB metadata test case (containing 6k thin devices), the new thin_ls consumes 554MB, compared to 271MB for the prior version.

The goal is to optimize this memory usage to be comparable to the previous version. This could be achieved by either revising the Aggregator for better memory efficiency or introducing an alternative space map.