Effect of lock free QSBR step 6

baseline commit, patch

Speedups (negative means slowdowns):

• parallel_get/1/70000: -8%
• parallel_get/2/70000: 14%
• parallel_get/4/70000: 11%
• parallel_get/8/70000: 45%
• parallel_get/16/70000: 15%
• parallel_get/1/2000000: -6%
• parallel_get/2/2000000: 8%
• parallel_get/4/2000000: 9%
• parallel_get/8/2000000:
• parallel_get/16/2000000:
• parallel_insert_disjoint_ranges/1/70000: -2%
• parallel_insert_disjoint_ranges/2/70000: 2%
• parallel_insert_disjoint_ranges/4/70000: 0%
• parallel_insert_disjoint_ranges/8/70000:
• parallel_insert_disjoint_ranges/16/70000:
• parallel_insert_disjoint_ranges/32/70000:
• parallel_insert_disjoint_ranges/1/2000000:
• parallel_insert_disjoint_ranges/2/2000000: 17%
• parallel_insert_disjoint_ranges/4/2000000:
• parallel_insert_disjoint_ranges/8/2000000: -8%
• parallel_insert_disjoint_ranges/16/2000000: -23%
• parallel_insert_disjoint_ranges/32/2000000:
• parallel_delete_disjoint_ranges/1/70000: -2%
• parallel_delete_disjoint_ranges/2/70000: 4%
• parallel_delete_disjoint_ranges/4/70000: 8%
• parallel_delete_disjoint_ranges/8/70000: 24%
• parallel_delete_disjoint_ranges/16/70000:
• parallel_delete_disjoint_ranges/32/70000:
• parallel_delete_disjoint_ranges/1/2000000: -16%
• parallel_delete_disjoint_ranges/2/2000000: 9%
• parallel_delete_disjoint_ranges/4/2000000: 9%
• parallel_delete_disjoint_ranges/8/2000000:
• parallel_delete_disjoint_ranges/16/2000000:
• parallel_delete_disjoint_ranges/32/2000000: