Re: Reducing opr_sanity test's runtime under CLOBBER_CACHE_ALWAYS

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
Cc: pgsql-hackers@lists.postgresql.org
Date: 2021-05-10T18:38:05Z
Lists: pgsql-hackers
Andres Freund <andres@anarazel.de> writes:
> On 2021-05-10 14:06:16 -0400, Tom Lane wrote:
>> I wonder if there's anything we could do to make ResetCatalogCache
>> faster?  It wouldn't help much for normal execution of course,
>> but it might do something to bring CCA testing time down out of
>> the stratosphere.

> We could make the hashtables shrink, not just grow...

Maybe ...

> There's also the issue that most people, I assume, run CCA tests with -O0. In
> a quick test that does make a big difference in e.g. ResetCatalogCache(). I
> just added a function specific annotation to optimize just that function and
> the overall time in my test shrank 10% or so.

If they do I think they're nuts ;-).  CCA is slow enough already without
hobbling it.

hyrax appears to use the usual -O2, as does/did avocet.  Not sure
if we have any other CCA buildfarm members right now.

			regards, tom lane



Commits

  1. Replace opr_sanity test's binary_coercible() function with C code.

  2. Clean up/tighten up coercibility checks in opr_sanity regression test.