[m-rev.] for post-commit review: speed up fat_sparse_bitset.list_to_set

[Zoltan Somogyi]

For review by anyone.

The attached benchmark results show that the new list_to_set
algorithms are consistently much faster than the old ones,
even for random data, for which the runs that the new algorithms
are designed to exploit are likely to be very short.

Zoltan.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Log.l2fs
Type: application/octet-stream
Size: 488 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20230129/aff5fef2/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: DIFF.l2fs
Type: application/octet-stream
Size: 16977 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20230129/aff5fef2/attachment-0001.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: BITSETS_SPEED
Type: application/octet-stream
Size: 1523 bytes
Desc: not available
URL: <http://lists.mercurylang.org/archives/reviews/attachments/20230129/aff5fef2/attachment-0002.obj>