Collect metrics from MemoizationTables

[majocha]

Description

MemoizationTables use ConcurrentDictionary as a backing store. If we use named Cache instances with no eviction instead, we can collect and see their metrics with --times option:

    ----------------------------------------------------------------------------------------------------------------------------
    | Cache name                          | hit-ratio |     adds |  updates |     hits |   misses | evictions | eviction-fails |
    |-------------------------------------|-----------|----------|----------|----------|----------|-----------|----------------|
    | mostSpecificOverrideMethodInfoCache |    73.26% |      365 |        0 |     1000 |      365 |         0 |              0 |
    |                typeSubsumptionCache |    69.27% |   129308 |        0 |   291450 |   129308 |         0 |              0 |
    |           rawDataValueTypeGenerator |    96.40% |       39 |        0 |     1045 |       39 |         0 |              0 |
    |           primaryTypeHierarchyCache |    50.00% |       14 |        0 |       14 |       14 |         0 |              0 |
    |             implicitConversionCache |    92.71% |     1367 |        0 |    17385 |     1367 |         0 |              0 |
    |                     namedItemsCache |    56.65% |    15841 |        0 |    20697 |    15841 |         0 |              0 |
    |                      v_memoize_file |    99.90% |      380 |        0 |   385340 |      380 |         0 |              0 |
    |                   propertyInfoCache |    38.71% |     1762 |        0 |     1113 |     1762 |         0 |              0 |
    |                     methodInfoCache |    76.99% |     5621 |        0 |    18809 |     5621 |         0 |              0 |
    |                    ilFieldInfoCache |    75.19% |       32 |        0 |       97 |       32 |         0 |              0 |
    |            entireTypeHierarchyCache |    90.04% |     3428 |        0 |    30992 |     3428 |         0 |              0 |
    |           recdOrClassFieldInfoCache |    46.21% |      795 |        0 |      683 |      795 |         0 |              0 |
    |                      eventInfoCache |     0.00% |        1 |        0 |        0 |        1 |         0 |              0 |
    ----------------------------------------------------------------------------------------------------------------------------

In cases where there are many instances with the same name, the above data is accumulated by name.

possible TODOs: Identify other uses of ConcurrentDictionary that may be also instrumented like this. Maybe collect some more useful data, currently hit ratio is computed, the rest are just raw event counts as they come from the Cache.

[github-actions]

⚠️ Release notes required, but author opted out

Warning

Author opted out of release notes, check is disabled for this pull request.
cc @dotnet/fsharp-team-msft