Thanks Jeff for your explanation, we are running SP132 and SP135. That bug is related to @@io_busy which over weighted IO busy, this is annoying since many DBAs still use sp_monitor to get the CPU/IO usage.
We actually use ProactiveDBA and MDA tables for monitoring, I don't think anybody would rely on sp_sysmon after ASEv12.5.
"update (index) stats with hashing" did consume much less tempdb space, however we don't see much improvement on time consumption against a 700M table.
Run Command | CPU Time (ms) | Elapse Time (ms) |
Update statistics xxxx | 1691914 | 3996041 |
Update index statistics xxxx | 2086711 | 4433156 |
Update index statistics xxx with hashing | 2018926 | 3872431 |
Update statistics xxx with hashing | 1461211 | 3278796 |