Re: Performance degradation on concurrent COPY into a single relation in PG16.

Hi,

On 2023-07-25 08:50:19 -0700, Andres Freund wrote:
> One idea I had was to add a fastpath that won't parse all strings, but will
> parse the strings that we would generate, and fall back to the more general
> variant if it fails. See the attached, rough, prototype:
> 
> fix_COPY_DEFAULT.patch + fastpath.patch:
> 746.971
> 
> fix_COPY_DEFAULT.patch + fastpath.patch + isdigit.patch:
> 715.570
> 
> Now, the precise contents of this fastpath are not yet clear (wrt imul or
> not), but I think the idea has promise.

Btw, I strongly suspect that fastpath wants to be branchless SSE when it grows
up.

Greetings,

Andres Freund