public inbox for gcc-help@gcc.gnu.org
 help / color / mirror / Atom feed
From: Marc Glisse <marc.glisse@inria.fr>
To: Mandeep Sandhu <mandeepsandhu.chd@gmail.com>
Cc: gcc-help@gcc.gnu.org
Subject: Re: stdc++ issue: extremely long compile time with large number of string literals
Date: Thu, 9 Jul 2020 22:49:30 +0200 (CEST)	[thread overview]
Message-ID: <alpine.DEB.2.23.453.2007092243400.6317@stedding.saclay.inria.fr> (raw)
In-Reply-To: <CAC+QLdSAw2sxrbAd4tS52_BtL22mTfUTBAGD2eohxFw6vNM+SA@mail.gmail.com>

On Thu, 9 Jul 2020, Mandeep Sandhu via Gcc-help wrote:

> I have an strange (to me) issue, where trying to compile a header
> which has a single "std::unordered_set<std::string>" initialized with
> around 50K short strings is taking forever.
>
> The set is declared as:
> const std::unordered_set<std::string> my_set ({"item1", "item2", ....});
>
> (The header is auto-generated using a script which takes a JSON array
> and puts its elements in an unordered_set)
>
> I understand that creation of many strings has an overhead, but this
> issue seems to affect compilation time, not runtime.
>
> Can someone explain to me why it takes such a long time to compile?
> Keeping the strings to under 5K, makes the program compile in about 8
> secs.
>
> I'm using the following compiler on Linux:
> $ g++ --version
> g++ (Ubuntu 9.2.1-9ubuntu2) 9.2.1 20191008
> Copyright (C) 2019 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

You didn't say what flags you are using. For anything autogenerated like 
that, going above -O1 is always risky.

One first thing to try is -ftime-report

At -O0
  phase opt and generate             :   3.46 ( 84%)   0.15 ( 45%)   3.62 ( 81%)  188385 kB ( 70%)

without a specific pass that stands out.

At -O1
  phase opt and generate             :  87.96 ( 99%)   6.57 ( 97%)  94.56 ( 99%)24503776 kB (100%)
  callgraph ipa passes               :  85.93 ( 97%)   6.52 ( 96%)  92.47 ( 97%)24413253 kB ( 99%)
  tree eh                            :  84.32 ( 95%)   6.43 ( 95%)  90.78 ( 95%)24340697 kB ( 99%)

so, it looks related to the optimization of exceptions...

-- 
Marc Glisse

  reply	other threads:[~2020-07-09 20:49 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-09 20:31 Mandeep Sandhu
2020-07-09 20:49 ` Marc Glisse [this message]
2020-07-09 21:27   ` Mandeep Sandhu
2020-07-09 23:49 ` Jonathan Wakely
     [not found]   ` <CAC+QLdRqjTe4YsUpQRcLG7tpxGda0oh6H788ORZa5MQR3NqRbw@mail.gmail.com>
     [not found]     ` <CAH6eHdTRiwA3t1K3okfVc+umPM8PJfbePaO7AHdLz+SYPOU9XQ@mail.gmail.com>
2020-07-10 19:36       ` Mandeep Sandhu
2020-07-10 20:25         ` Jonathan Wakely
2020-07-11  6:27           ` Mandeep Sandhu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=alpine.DEB.2.23.453.2007092243400.6317@stedding.saclay.inria.fr \
    --to=marc.glisse@inria.fr \
    --cc=gcc-help@gcc.gnu.org \
    --cc=mandeepsandhu.chd@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).