From: "Tulio Magno Quites Machado Filho" <tuliom@linux.vnet.ibm.com>
To: Adhemerval Zanella <adhemerval.zanella@linaro.org>,
Rajalakshmi Srinivasaraghavan <raji@linux.vnet.ibm.com>
Cc: libc-alpha@sourceware.org, Florian Weimer <fweimer@redhat.com>
Subject: Re: [PATCH] powerpc: Use aligned stores in memset
Date: Wed, 08 Nov 2017 18:52:00 -0000 [thread overview]
Message-ID: <87vaik8uxy.fsf@linux.vnet.ibm.com> (raw)
In-Reply-To: <c4bc039f-64ab-a171-1044-d98cec120d60@linaro.org>
Adhemerval Zanella <adhemerval.zanella@linaro.org> writes:
> I think one way to provide a slight better memcpy implementation for POWER8
> and still be able to circumvent the non-aligned on non-cacheable memory
> is to use tunables.
>
> The branch azanella/memcpy-power8 [1] has a power8 memcpy optimization which
> uses unaligned load and stores that I created some time ago but never actually
> send upstream. It shows better performance on both bench-memcpy and
> bench-memcpy-random (about 10% on latter) and mixed results on bench-memcpy-large
> (which it is mainly dominated by memory throughput and on the environment I am
> using, a shared PowerKVM instance, the results does not seem to be reliable).
>
> It could use some tunning, specially on some the range I used for unrolling
> the load/stores and it also does not care for unaligned access on cross-page
> boundary (which tend to be quite slow on current hardware, but also on
> current page size of usual 64k also uncommon).
>
> This first patch does not enable this option as a default for POWER8, it just
> add on string tests as an option. The second patch changes the selection to:
>
> 1. If glibc is configure with tunables, set the new implementation as the
> default for ISA 2.07 (power8).
>
> 2. Also if tunable is active, add the parameter glibc.tune.aligned_memopt
> to disable the new implementation selection.
I think it would be safer if we don't change the default behavior.
IMHO, programs that want a performance improvement would have to set a
tunable.
In other words, the new implementation would be disabled by default.
> So programs that rely on aligned loads can set:
>
> GLIBC_TUNABLES=glibc.tune.aligned_memopt=1
I also think that we should not expose internal details of the implementation
to users, i.e. avoiding to use aligned/unaligned in the name of the function
and in the tunables.
I think that glibc.tune.cached_memopt=1 better exposes what is the optimal
use-case scenario of this implementation.
> This is a RFC patch and if the idea sounds to powerpc arch mantainers I can
> work on finishing the patch with more comments and send upstream. I tried
> to apply same unaligned idea for memset and memmove, but I could get any real
> improvement in neither.
I like the idea. Could you merge both patches and send it to libc-alpha,
please?
> [1] https://sourceware.org/git/?p=glibc.git;a=shortlog;h=refs/heads/azanella/memcpy-power8
> [ bench-memcpy-random.out: text/plain ]
> {
> "timing_type": "hp_timing",
> "functions": {
> "memcpy": {
> "bench-variant": "random",
> "ifuncs": ["__memcpy_power8",
I also suggest give a more specific name, e.g. __memcpy_power8_cached.
That would make room for a POWER8 implementation what uses only naturally
aligned loads/stores.
Your implementation uses lxvd2x and stxvd2x, which should be avoided in a
cache-inhibited scenario, i.e. glibc.tune.aligned_memopt=0.
However, after changing the tunables' name to glibc.tune.cached_memopt, I think
these instructions could stay they're executed when glibc.tune.cached_memopt=1.
Thanks!!!
--
Tulio Magno
next prev parent reply other threads:[~2017-11-08 18:52 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-08-18 5:13 Rajalakshmi Srinivasaraghavan
2017-08-18 6:21 ` Florian Weimer
2017-08-18 6:51 ` Rajalakshmi Srinivasaraghavan
2017-08-18 9:10 ` Florian Weimer
2017-08-18 12:13 ` Adhemerval Zanella
2017-09-12 10:30 ` Florian Weimer
2017-09-12 12:18 ` Zack Weinberg
2017-09-12 13:57 ` Steven Munroe
2017-09-12 14:37 ` Joseph Myers
2017-09-12 15:06 ` Zack Weinberg
2017-09-12 17:09 ` Florian Weimer
2017-09-12 13:38 ` Steven Munroe
2017-09-12 14:08 ` Florian Weimer
2017-09-12 14:16 ` Steven Munroe
2017-09-12 17:04 ` Florian Weimer
2017-09-12 19:21 ` Steven Munroe
2017-09-12 19:45 ` Florian Weimer
2017-09-12 20:25 ` Steven Munroe
2017-09-13 13:12 ` Tulio Magno Quites Machado Filho
2017-09-18 13:54 ` Florian Weimer
2017-10-03 18:29 ` Adhemerval Zanella
2017-10-05 12:13 ` Rajalakshmi Srinivasaraghavan
2017-11-08 18:52 ` Tulio Magno Quites Machado Filho [this message]
2017-12-08 19:52 ` [PATCHv2] powerpc: POWER8 memcpy optimization for cached memory Tulio Magno Quites Machado Filho
2017-12-08 20:06 ` Florian Weimer
2017-12-11 12:44 ` Tulio Magno Quites Machado Filho
2017-12-11 20:09 ` Adhemerval Zanella
2017-12-10 7:11 ` Rajalakshmi Srinivasaraghavan
2017-12-11 19:48 ` Tulio Magno Quites Machado Filho
2017-08-18 6:25 ` [PATCH] powerpc: Use aligned stores in memset Andrew Pinski
2017-08-21 2:20 ` Tulio Magno Quites Machado Filho
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=87vaik8uxy.fsf@linux.vnet.ibm.com \
--to=tuliom@linux.vnet.ibm.com \
--cc=adhemerval.zanella@linaro.org \
--cc=fweimer@redhat.com \
--cc=libc-alpha@sourceware.org \
--cc=raji@linux.vnet.ibm.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).