Effect of lock free QSBR step 8

baseline commit, patch

Speedups (negative means slowdowns):

• parallel_get/1/70000: 7%
• parallel_get/2/70000: 2%
• parallel_get/4/70000:
• parallel_get/8/70000:
• parallel_get/16/70000: 3%
• parallel_get/1/2000000: 5%
• parallel_get/2/2000000: 2%
• parallel_get/4/2000000: +2%
• parallel_get/8/2000000: 2%
• parallel_get/16/2000000:
• parallel_insert_disjoint_ranges/1/70000: 2%
• parallel_insert_disjoint_ranges/2/70000:
• parallel_insert_disjoint_ranges/4/70000:
• parallel_insert_disjoint_ranges/8/70000:
• parallel_insert_disjoint_ranges/16/70000:
• parallel_insert_disjoint_ranges/32/70000:
• parallel_insert_disjoint_ranges/1/2000000: 33%
• parallel_insert_disjoint_ranges/2/2000000: 23%
• parallel_insert_disjoint_ranges/4/2000000: -16%
• parallel_insert_disjoint_ranges/8/2000000: -9%
• parallel_insert_disjoint_ranges/16/2000000:
• parallel_insert_disjoint_ranges/32/2000000:
• parallel_delete_disjoint_ranges/1/70000: 6%
• parallel_delete_disjoint_ranges/2/70000:
• parallel_delete_disjoint_ranges/4/70000: 5%
• parallel_delete_disjoint_ranges/8/70000: 78%
• parallel_delete_disjoint_ranges/16/70000: 68%
• parallel_delete_disjoint_ranges/32/70000: 52%
• parallel_delete_disjoint_ranges/1/2000000: 5%
• parallel_delete_disjoint_ranges/2/2000000: 3%
• parallel_delete_disjoint_ranges/4/2000000: 5%
• parallel_delete_disjoint_ranges/8/2000000: 65%
• parallel_delete_disjoint_ranges/16/2000000: 34%
• parallel_delete_disjoint_ranges/32/2000000: 50%