From: Noah Goldstein via Libc-alpha <libc-alpha@sourceware.org>
To: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cc: GNU C Library <libc-alpha@sourceware.org>
Subject: Re: [PATCH v3 3/7] stdlib: Optimization qsort{_r} swap implementation (BZ #19305)
Date: Fri, 15 Oct 2021 12:45:16 -0500 [thread overview]
Message-ID: <CAFUsyfK8LLTuXNa5B74VGScYLMLBKb++LEimxKXNdJsxxCwqvg@mail.gmail.com> (raw)
In-Reply-To: <ee6580d8-8929-a83d-c561-59db1827fb34@linaro.org>
On Fri, Oct 15, 2021 at 12:34 PM Adhemerval Zanella
<adhemerval.zanella@linaro.org> wrote:
>
>
>
> On 15/10/2021 14:17, Noah Goldstein wrote:
> > On Fri, Oct 15, 2021 at 8:29 AM Adhemerval Zanella
> > <adhemerval.zanella@linaro.org> wrote:
> >>
> >>
> >>
> >> On 13/10/2021 00:39, Noah Goldstein wrote:
> >>>
> >>>
> >>> On Tue, Oct 12, 2021 at 11:29 PM Noah Goldstein <goldstein.w.n@gmail.com <mailto:goldstein.w.n@gmail.com>> wrote:
> >>>
> >>>
> >>>
> >>> On Fri, Sep 3, 2021 at 1:14 PM Adhemerval Zanella via Libc-alpha <libc-alpha@sourceware.org <mailto:libc-alpha@sourceware.org>> wrote:
> >>>
> >>> It optimizes take in consideration both the most common elements are
> >>> either 32 or 64 bit in size [1] and inputs are aligned to the word
> >>> boundary. This is similar to the optimization done on lib/sort.c
> >>> from Linux.
> >>>
> >>> This patchs adds an optimized swap operation on qsort based in previous
> >>> msort one. Instead of byte operation, three variants are provided:
> >>>
> >>> 1. Using uint32_t loads and stores.
> >>> 2. Using uint64_t loads and stores.
> >>> 3. Generic one with a temporary buffer and memcpy/mempcpy.
> >>>
> >>> The 1. and 2. options are selected only either if architecture defines
> >>> _STRING_ARCH_unaligned or if base pointer is aligned to required type.
> >>>
> >>> It also fixes BZ#19305 by checking input size against number of
> >>> elements 1 besides 0.
> >>>
> >>> Checked on x86_64-linux-gnu.
> >>>
> >>> [1] https://sourceware.org/pipermail/libc-alpha/2018-August/096984.html <https://sourceware.org/pipermail/libc-alpha/2018-August/096984.html>
> >>> ---
> >>> stdlib/qsort.c | 109 +++++++++++++++++++++++++++++++++++++++++--------
> >>> 1 file changed, 91 insertions(+), 18 deletions(-)
> >>>
> >>> diff --git a/stdlib/qsort.c b/stdlib/qsort.c
> >>> index 23f2d28314..59458d151b 100644
> >>> --- a/stdlib/qsort.c
> >>> +++ b/stdlib/qsort.c
> >>> @@ -24,20 +24,85 @@
> >>> #include <limits.h>
> >>> #include <stdlib.h>
> >>> #include <string.h>
> >>> +#include <stdbool.h>
> >>>
> >>> -/* Byte-wise swap two items of size SIZE. */
> >>> -#define SWAP(a, b, size) \
> >>> - do \
> >>> - { \
> >>> - size_t __size = (size); \
> >>> - char *__a = (a), *__b = (b); \
> >>> - do \
> >>> - { \
> >>> - char __tmp = *__a; \
> >>> - *__a++ = *__b; \
> >>> - *__b++ = __tmp; \
> >>> - } while (--__size > 0); \
> >>> - } while (0)
> >>> +/* Swap SIZE bytes between addresses A and B. These helpers are provided
> >>> + along the generic one as an optimization. */
> >>> +
> >>> +typedef void (*swap_func_t)(void * restrict, void * restrict, size_t);
> >>> +
> >>> +/* Return trues is elements can be copied used word load and sortes.
> >>> + The size must be a multiple of the alignment, and the base address. */
> >>> +static inline bool
> >>> +is_aligned_to_copy (const void *base, size_t size, size_t align)
> >>> +{
> >>> + unsigned char lsbits = size;
> >>> +#if !_STRING_ARCH_unaligned
> >>> + lsbits |= (unsigned char)(uintptr_t) base;
> >>> +#endif
> >>> + return (lsbits & (align - 1)) == 0;
> >>> +}
> >>> +
> >>> +#define SWAP_WORDS_64 (swap_func_t)0
> >>> +#define SWAP_WORDS_32 (swap_func_t)1
> >>> +#define SWAP_BYTES (swap_func_t)2
> >>> +
> >>> +static void
> >>> +swap_words_64 (void * restrict a, void * restrict b, size_t n)
> >>> +{
> >>> + do
> >>> + {
> >>> + n -= 8;
> >>> + uint64_t t = *(uint64_t *)(a + n);
> >>> + *(uint64_t *)(a + n) = *(uint64_t *)(b + n);
> >>> + *(uint64_t *)(b + n) = t;
> >>> + } while (n);
> >>> +}
> >>> +
> >>> +static void
> >>> +swap_words_32 (void * restrict a, void * restrict b, size_t n)
> >>> +{
> >>> + do
> >>> + {
> >>> + n -= 4;
> >>> + uint32_t t = *(uint32_t *)(a + n);
> >>> + *(uint32_t *)(a + n) = *(uint32_t *)(b + n);
> >>> + *(uint32_t *)(b + n) = t;
> >>> + } while (n);
> >>> +}
> >>>
> >>>
> >>> I'm not certain swap_words_32 / swap_words_8 will be optimal for larger
> >>> key sizes. Looking at GCC's implementation of swap_generic on modern x86_64:
> >>> https://godbolt.org/z/638h3Y9va <https://godbolt.org/z/638h3Y9va>
> >>> It's able to optimize the temporary buffer out of the loop and use xmm registers which
> >>> will likely win out for larger sizes.
> >>
> >> It is probably not the most optimized code compiler can generated since
> >> I tried to go for a more generic code that should results in a somewhat
> >> better code in a architecture agnostic code. I trying to mimic some
> >> of optimization Linux did on 37d0ec34d111acfdb, and the swap_bytes I
> >> used the same strategy used on d3496c9f4f27d (Linux did not optimize
> >> anything for the byte version).
> >>
> >> I think it would be possible to tune it for an specific architecture
> >> and/or compiler, but I would prefer to use a good enough algorithm that
> >> work reasonable on multiple architectures.
> >
> > That's fair. Although I still think there are some improvements.
> >
> > Looking at the assembly for all three in fact it seems GCC optimizes all of them
> > to larger register copies: https://godbolt.org/z/bd9nnnoEY
> >
> > The biggest difference seems to be the setup / register spills for
> > the generic version so for the common case of a relatively small key
> > the special case for 4/8 makes sense.
> >
> > Have you checked that GCC is able to use the conditions for selecting
> > the swap function to optimize the functions themselves? In the godbolt
> > link I got reasonable value out of adding the invariants to swap_64/swap_32.
>
> Not yet, but I think the generated code for both x86_64 and aarch64 seems
> simple enough and should cover the common case (keys with size 4 or 8)
> fast enough [1].
swap is in the inner loop. Seems like a pretty critical component to have fully
optimized. The aarch64 version looks good, but the x86_64 version seems
to be lacking. Not arguing for an arch specific version, but if the directives
can add value to x86_64 without detracting from aarch64 seems like a zero
cost improvement.
>
> And since for v4 my plan is not to remove mergesort anymore, but limit it
> to a static buffer; it might be possible to use the same swap routines
> for both mergesort and quicksort.
>
> >
> > It also may be worth it to write a custom memcpy implementation for
> > size = 0..SWAP_GENERIC_SIZE so it can be inlined (and probably
> > more optimized than what generic memcpy can get away with).
> >
>
> I *really* do not want to go for this way. My hope is with small sizes
> compiler can inline the memcpy (which gcc does for most common
> architectures).
Since size is non-constant for the tail I don't see how we are going
avoid 3x memcpy calls. Although that can be another patch if it
gets values.
>
>
> [1] https://godbolt.org/z/v7e4xxqGa
next prev parent reply other threads:[~2021-10-15 17:45 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-09-03 17:11 [PATCH v3 0/7] Use introsort for qsort Adhemerval Zanella via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 1/7] benchtests: Add bench-qsort Adhemerval Zanella via Libc-alpha
2021-09-04 9:09 ` Alexander Monakov via Libc-alpha
2021-09-06 18:30 ` Adhemerval Zanella via Libc-alpha
2021-10-13 3:19 ` Noah Goldstein via Libc-alpha
2021-10-15 12:52 ` Adhemerval Zanella via Libc-alpha
2021-10-15 16:39 ` Noah Goldstein via Libc-alpha
2021-10-15 17:19 ` Adhemerval Zanella via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 2/7] support: Fix getopt_long with CMDLINE_OPTIONS Adhemerval Zanella via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 3/7] stdlib: Optimization qsort{_r} swap implementation (BZ #19305) Adhemerval Zanella via Libc-alpha
2021-10-13 3:29 ` Noah Goldstein via Libc-alpha
2021-10-13 3:39 ` Noah Goldstein via Libc-alpha
2021-10-15 13:29 ` Adhemerval Zanella via Libc-alpha
2021-10-15 17:17 ` Noah Goldstein via Libc-alpha
2021-10-15 17:34 ` Adhemerval Zanella via Libc-alpha
2021-10-15 17:45 ` Noah Goldstein via Libc-alpha [this message]
2021-10-15 17:56 ` Adhemerval Zanella via Libc-alpha
2021-10-15 13:12 ` Adhemerval Zanella via Libc-alpha
2021-10-15 16:45 ` Noah Goldstein via Libc-alpha
2021-10-15 17:21 ` Adhemerval Zanella via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 4/7] stdlib: Move insertion sort out qsort Adhemerval Zanella via Libc-alpha
2021-09-06 20:35 ` Fangrui Song via Libc-alpha
2021-09-06 20:48 ` Fangrui Song via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 5/7] stdlib: qsort: Move some macros to inline function Adhemerval Zanella via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 6/7] stdlib: Implement introsort with qsort Adhemerval Zanella via Libc-alpha
2021-09-04 9:17 ` Alexander Monakov via Libc-alpha
2021-09-06 18:43 ` Adhemerval Zanella via Libc-alpha
2021-09-06 20:23 ` Fangrui Song via Libc-alpha
2021-10-13 3:53 ` Noah Goldstein via Libc-alpha
2021-09-03 17:11 ` [PATCH v3 7/7] stdlib: Remove use of mergesort on qsort (BZ #21719) Adhemerval Zanella via Libc-alpha
2021-09-03 19:18 ` [PATCH v3 0/7] Use introsort for qsort Paul Eggert
2021-09-06 14:13 ` Carlos O'Donell via Libc-alpha
2021-09-06 17:03 ` Zack Weinberg via Libc-alpha
2021-09-06 18:19 ` Adhemerval Zanella via Libc-alpha
2021-09-07 0:14 ` Paul Eggert
2021-09-07 14:32 ` Adhemerval Zanella via Libc-alpha
2021-09-07 17:39 ` Paul Eggert
2021-09-07 18:07 ` Adhemerval Zanella via Libc-alpha
2021-09-07 19:28 ` Paul Eggert
2021-09-08 11:56 ` Adhemerval Zanella via Libc-alpha
2021-09-09 0:39 ` Paul Eggert
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
List information: https://www.gnu.org/software/libc/involved.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAFUsyfK8LLTuXNa5B74VGScYLMLBKb++LEimxKXNdJsxxCwqvg@mail.gmail.com \
--to=libc-alpha@sourceware.org \
--cc=adhemerval.zanella@linaro.org \
--cc=goldstein.w.n@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).