Re: JIT breaks PostGIS

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Christoph Berg <myon(at)debian(dot)org>
Cc: Andres Freund <andres(at)anarazel(dot)de>, Darafei "Komяpa" Praliaskouski <me(at)komzpa(dot)net>, PostgreSQL Developers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: Re: JIT breaks PostGIS
Date: 2018-07-22 19:47:58
Message-ID: 27298.1532288878@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Christoph Berg <myon(at)debian(dot)org> writes:
> The question will be coming up eventually, though, and I think the
> options on the packaging side are:

> 1) Disable jit completely
> 2) Compile --with-llvm, but disable jit in the config by default
> 3) Compile --with-llvm, but disable jit for older llvm versions
> 4) Enable jit everywhere where llvm >= 3.9 is available

> Option 4 is what the Debian packages implement now, but it might make
> sense to go to 2 or 3 for PG11 (only).

Well, right now JIT is certainly beta-quality code, so you ought
to expect bugs. We have an open item at
https://wiki.postgresql.org/wiki/PostgreSQL_11_Open_Items
to decide whether to ship v11 with JIT enabled by default or not,
but I don't expect that decision will be taken until much closer
to release. Until then, I think you should be doing (4) so that
we can gather data to inform the eventual decision.

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Fabien COELHO 2018-07-22 20:13:20 Re: pgbench: improve --help and --version parsing
Previous Message Christoph Berg 2018-07-22 18:57:01 Re: JIT breaks PostGIS