Hi
In my case there is a problem with default a cache. (high spinlok contention& hogh ytilization cpu)
I thought that, as for example MSSQL, memory for a cache and for aggregate requests jointly and they can compete for it.(In MSSQL memory under requests (aggregation) and a buffer pool undertakes from one pool of memory)
My procedure cache=5gb, but I know what my query(with aggregation) can request a lot of memory.
And i have't recompilation in procedeire cache.
how I can calculate or find memory size requested statement(query) for aggregate?
And how 5 gb(procedure cache) in case of requests which aggregate very large number of data are enough.
All aggregate work in tempdb?
Also
From time to time we had high values in sysmon
“Object Manager Spinlock Contention”
When we had peak utilizaion cpu (98%)
“Object Manager Spinlock Contention”= 95%.