About query cache vs CommandFlags.NoCache running speed？

[znyet]

How many times does the operating speed differ.
use CommandFlags.NoCache, Does it have a serious impact on the application?

[mgravell]

race your horses

Here's some simple numbers

|                 ORM |                         Method | Return |      Mean |   StdDev |     Error |  Gen 0 |  Gen 1 | Gen 2 | Allocated |
|-------------------- |------------------------------- |------- |----------:|---------:|----------:|-------:|-------:|------:|----------:|
| Dapper cache impact |      ExecuteNoParameters_Cache |   Void |  93.48 us | 1.151 us |  1.740 us | 0.1250 |      - |     - |   1.42 KB |
| Dapper cache impact |    ExecuteNoParameters_NoCache |   Void |  92.81 us | 1.050 us |  1.765 us | 0.1250 |      - |     - |   1.42 KB |
|
| Dapper cache impact |        ExecuteParameters_Cache |   Void |  64.47 us | 0.721 us |  1.212 us | 0.5000 | 0.0625 |     - |   3.13 KB |
| Dapper cache impact |      ExecuteParameters_NoCache |   Void | 370.75 us | 2.988 us |  4.517 us | 2.0000 | 1.0000 |     - |  13.16 KB |
|
| Dapper cache impact |   QueryFirstNoParameters_Cache |   Void |  94.85 us | 8.402 us | 12.702 us | 0.2500 | 0.1250 |     - |   2.23 KB |
| Dapper cache impact | QueryFirstNoParameters_NoCache |   Void | 102.12 us | 0.802 us |  1.213 us | 0.2500 | 0.1250 |     - |   2.28 KB |
|
| Dapper cache impact |     QueryFirstParameters_Cache |   Void |  67.48 us | 1.064 us |  1.609 us | 0.5000 | 0.1250 |     - |   3.83 KB |
| Dapper cache impact |   QueryFirstParameters_NoCache |   Void | 388.73 us | 5.431 us |  8.210 us | 2.0000 | 1.0000 |     - |  13.87 KB |
|

Note that with AOT, there is functionally no cache - which is not to say "it sucks", but rather the opposite "the need for that concept goes away completely, and it works great/equally with known/unknown SQL "

But yes: in vanilla Dapper, this has an impact (except for the "no parameters, no row materialization" case, ExecuteNoParameters...)