RE: [Patch] Optimize dropping of relation buffers using dlist - Mailing list pgsql-hackers

From: Jamison, Kirk/ジャミソン カーク <k.jamison@fujitsu.com>
> Recovery performance measurement results below.
> But it seems there are overhead even with large shared buffers.
>
> | s_b   | master | patched | %reg  |
> |-------|--------|---------|-------|
> | 128MB | 36.052 | 39.451  | 8.62% |
> | 1GB   | 21.731 | 21.73   | 0.00% |
> | 20GB  | 24.534 | 25.137  | 2.40% |
> | 100GB | 30.54  | 31.541  | 3.17% |

Did you really check that the optimization path is entered and the traditional path is never entered?

With the following code, when the main fork does not meet the optimization criteria, other forks are not optimized as
well. You want to determine each fork's optimization separately, don't you? 

+        /* If blocks are invalid, exit the optimization and execute full scan */
+        if (nTotalBlocks == InvalidBlockNumber)
+            break;


+        else
+            break;
+    }
     for (i = 0; i < NBuffers; i++)


Regards
Takayuki Tsunakawa