From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Jeff Davis <pgsql(at)j-davis(dot)com> |
Cc: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, pgsql-hackers(at)postgresql(dot)org, Peter Geoghegan <pg(at)bowt(dot)ie> |
Subject: | Re: v13: Performance regression related to FORTIFY_SOURCE |
Date: | 2020-06-05 01:41:44 |
Message-ID: | 1443262.1591321304@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Jeff Davis <pgsql(at)j-davis(dot)com> writes:
> On Thu, 2020-06-04 at 16:35 -0400, Alvaro Herrera wrote:
>> If it is something worth worrying about, let's discuss what's a good
>> fix for it.
> I did post a fix for it, but it's not a very clean fix. I'm slightly
> inclined to proceed with that fix, but I was hoping someone else would
> have a better suggestion.
> How about if I wait another week, and if we still don't have a better
> fix, I will commit this one.
TBH, I don't think we should do this, at least not on the strength of the
evidence you posted so far. It looks to me like you are micro-optimizing
for one compiler on one platform. Moreover, you're basically trying to
work around a compiler codegen bug that might not be there next year.
I think what'd make more sense is to file this as a gcc bug ("why doesn't
it remove the useless object size check?") and see what they say about
that. If the answer is that this isn't a gcc bug for whatever reason,
then we could think about whether we should work around it on the
source-code level. Even then, I'd want more evidence than has been
presented about this not causing a regression on other toolchains/CPUs.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Andres Freund | 2020-06-05 01:57:58 | Re: significant slowdown of HashAggregate between 9.6 and 10 |
Previous Message | Jeff Davis | 2020-06-05 01:22:03 | Re: significant slowdown of HashAggregate between 9.6 and 10 |