public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "redbeard0531 at gmail dot com" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/95079] New: unorderd_map::insert_or_assign and try_emplace should only hash and mod once unless there is a rehash.
Date: Tue, 12 May 2020 11:00:56 +0000	[thread overview]
Message-ID: <bug-95079-4@http.gcc.gnu.org/bugzilla/> (raw)

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95079

            Bug ID: 95079
           Summary: unorderd_map::insert_or_assign and try_emplace should
                    only hash and mod once unless there is a rehash.
           Product: gcc
           Version: 10.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: redbeard0531 at gmail dot com
  Target Milestone: ---

Currently insert_or_assign() and try_emplace() call find(key) and fall back to
emplace(...) if that fails to find the key. The computed hash (and more
importantly in general) the modded bucket index computed by find() is thrown
away and recomputed by emplace(). Instead they should compute the hash once,
and unless there is a rehash, also only do the modulus once. This optimization
is already performed for operator[].

https://godbolt.org/z/cw82RC shows that the hasher is invoked once for
operator[] and twice for insert_or_assign().
http://quick-bench.com/ge8Suq7PcdRKm6IBQbjvwuXhW6Y shows that there is a
significant performance difference (20% in this test).

(I know std::unordered_map is always going to be less than fast on 64-bit
platforms, but it doesn't need to be slower than it needs to be...)

             reply	other threads:[~2020-05-12 11:00 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-12 11:00 redbeard0531 at gmail dot com [this message]
2020-05-12 11:26 ` [Bug libstdc++/95079] " redi at gcc dot gnu.org
2020-05-24 10:00 ` fdumont at gcc dot gnu.org
2020-05-29 11:13 ` cvs-commit at gcc dot gnu.org
2021-02-28 14:39 ` fdumont at gcc dot gnu.org
2021-03-02 13:40 ` redbeard0531 at gmail dot com
2021-03-02 16:49 ` fdumont at gcc dot gnu.org

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=bug-95079-4@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.gnu.org \
    /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).