From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.5-pre1 (2020-06-20) on
	ip-172-31-74-118.ec2.internal
X-Spam-Level: *
X-Spam-Status: No, score=1.8 required=3.0 tests=BAYES_50,TO_NO_BRKTS_PCNT
	autolearn=no autolearn_force=no version=3.4.5-pre1
X-Received: by 2002:a37:6491:: with SMTP id y139mr962654qkb.483.1616432546169;
        Mon, 22 Mar 2021 10:02:26 -0700 (PDT)
X-Received: by 2002:a25:1883:: with SMTP id 125mr565113yby.465.1616432545903;
 Mon, 22 Mar 2021 10:02:25 -0700 (PDT)
Path: eternal-september.org!reader02.eternal-september.org!weretis.net!feeder8.news.weretis.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.ada
Date: Mon, 22 Mar 2021 10:02:25 -0700 (PDT)
Injection-Info: google-groups.googlegroups.com; posting-host=73.2.251.187; posting-account=JSxOkAoAAADa00TJoz2WZ_46XrZCdXeS
NNTP-Posting-Host: 73.2.251.187
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <07a56dcc-9e17-49b2-a980-3a5a2d265cedn@googlegroups.com>
Subject: Performance of records with variant parts
From: John Perry <john.perry@usm.edu>
Injection-Date: Mon, 22 Mar 2021 17:02:26 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Xref: reader02.eternal-september.org comp.lang.ada:61626
List-Id: <comp.lang.ada>

Hello all

I encountered a significant penalty when using a record with variant parts =
in gnat community edition, on both MacOS and Linux. Suppose I have the foll=
owing. (Kind of long, sorry.)

| type Object_Type is ( Sphere, Plane, None );
| type Surface_Type is ( Shiny, Checkerboard );
| type Thing_Type(kind: Object_Type :=3D None; surface: Surface_Type :=3D S=
hiny)
| is record
|    case kind is
|       when None =3D>
|          null;
|       when Sphere =3D>
|          radius2: Long_Float :=3D 0.0;
|          center: Vector;
|       when Plane =3D>
|          offset: Long_Float :=3D 0.0;
|          norm: Vector;
|    end case;
| end record;

Noticing that the fields are actually the same types, but different names, =
I also tried this:

| type Thing_Type is record
|    Kind: Object_Type :=3D None;
|    Surface: Surface_Type :=3D Shiny;
|    Size: Long_Float :=3D 0.0;
|    Pos: Vector;
| end record;

This second version doesn't seem a good idea, but compiled output runs 15-3=
0% faster. As far as I can tell***, it's related to this test which gets ca=
lled ~3 million times:

|       case obj.kind is
|       when Sphere =3D>
|          -- do stuff
|       when Plane =3D>
|          -- do other stuff
|       when None =3D>
|          null;
|       end case;

Is there a reason that the record with variant parts runs so much slower? I=
 can make the complete source available if need be.

john perry

***"as far as I can tell": gprof  stated at one point that this procedure s=
ucks up most of the time; it's one of only two places where the discriminan=
t is evaluated, and it's the only one that gets called A LOT; and this one =
change makes for the difference in performance.