low plan cache memory - sql-server

Question

My question is: Why would an SQL server have a low amount of memory allocated to Plan Cache? And, if a correction is needed, what might be done to correct this?
We have an SQL server with an issue of Compilations per second being high indicating not enough of the execution plans are cached for use (first detected when we ran sp_AskBrent #ExpertMode=1, #Seconds=30 [from brentozar.com/askbrent/]).
We have run the SQL Live Monitor application (https://sqlmonitor.codeplex.com/) on the server and the Plan Cache results show a very low amount of memory (355.27 MB) allocated to caching execution plans and therefore a low Plan Cache Hit Ratio (varying between 5 and 50 percent).
My research shows that memory allocated to Plan Cache is not a configurable amount, but a calculation based on the memory allocated to the SQL instance. So, for this server, which has 48GB total and 40GB allocated to SQL, the calculation of (.75 * 4GB) + (.1 * 36GB) should allocate 6.6GB for Plan Cache. Did I calculate correctly?
Of note, this server has only one production database and that database is 50GB in size. We have Optimize for Ad hoc Workloads set to True and just set Parameterization at the database level to Forced.
Compared to another SQL server (that has 32GB total and 26GB allocated) the Plan Cache numbers look more reasonable (4GB in size and a Hit Ratio of above 80 percent.
Also, running the script below against both SQL servers consistently shows the problem server having a hit percentage in the mid 70% range and the other server showing a hit percentage in the high 90% range.
WITH cte1
AS ( SELECT [dopc].[object_name] ,
[dopc].[instance_name] ,
[dopc].[counter_name] ,
[dopc].[cntr_value] ,
[dopc].[cntr_type] ,
ROW_NUMBER() OVER ( PARTITION BY [dopc].[object_name], [dopc].[instance_name] ORDER BY [dopc].[counter_name] ) AS r_n
FROM [sys].[dm_os_performance_counters] AS dopc
WHERE [dopc].[counter_name] LIKE '%Cache Hit Ratio%'
AND ( [dopc].[object_name] LIKE '%Plan Cache%'
OR [dopc].[object_name] LIKE '%Buffer Cache%'
)
AND [dopc].[instance_name] LIKE '%_Total%'
)
SELECT CONVERT(DECIMAL(16, 2), ( [c].[cntr_value] * 1.0 / [c1].[cntr_value] ) * 100.0) AS [hit_pct]
FROM [cte1] AS c
INNER JOIN [cte1] AS c1
ON c.[object_name] = c1.[object_name]
AND c.[instance_name] = c1.[instance_name]
WHERE [c].[r_n] = 1
AND [c1].[r_n] = 2;

Answer 1

See:
... The maximum size for all caches is a function of the buffer pool size and cannot exceed the maximum server memory...
(https://technet.microsoft.com/en-us/library/ms181055%28v=sql.105%29.aspx)
I think optimize for ad hoc workloads option helps if you mostly run adhoc queries:
sp_configure 'show advanced options',1
GO
reconfigure
GO
sp_configure 'optimize for ad hoc workloads',1
GO
reconfigure
GO
DBCC FREEPROCCACHE
GO
Don't try it first on Prod servers, especialy FREEPROCCACHE.