From: Nathan Sidwell <nathan@acm.org>
To: Jakub Jelinek <jakub@redhat.com>, Alexander Monakov <amonakov@ispras.ru>
Cc: gcc-patches@gcc.gnu.org, Bernd Schmidt <bschmidt@redhat.com>,
Dmitry Melnik <dm@ispras.ru>,
Thomas Schwinge <thomas@codesourcery.com>
Subject: Re: [gomp-nvptx 2/9] nvptx backend: new "uniform SIMT" codegen variant
Date: Wed, 02 Dec 2015 15:18:00 -0000 [thread overview]
Message-ID: <565F0BB3.5020608@acm.org> (raw)
In-Reply-To: <20151202151205.GS5675@tucnak.redhat.com>
On 12/02/15 10:12, Jakub Jelinek wrote:
> If we have a reasonable IPA pass to discover which addressable variables can
> be shared by multiple threads and which can't, then we could use soft-stack
> for those that can be shared by multiple PTX threads (different warps, or
> same warp, different threads in it), then we shouldn't need to copy any
> stack, just broadcast the scalar vars.
Note the current scalar (.reg) broadcasting uses the live register set. Not
the subset of that that is actually read within the partitioned region. That'd
be a relatively straightforward optimization I think.
nathan
next prev parent reply other threads:[~2015-12-02 15:18 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-01 15:28 [gomp-nvptx 0/9] Codegen bits for NVPTX OpenMP SIMD Alexander Monakov
2015-12-01 15:28 ` [gomp-nvptx 9/9] adjust SIMD loop lowering for SIMT targets Alexander Monakov
2015-12-01 22:40 ` Alexander Monakov
2015-12-02 11:48 ` Jakub Jelinek
2015-12-02 13:54 ` Alexander Monakov
2015-12-02 14:02 ` Jakub Jelinek
2015-12-02 14:26 ` Alexander Monakov
2015-12-01 15:28 ` [gomp-nvptx 6/9] nvptx libgcc: rewrite in C Alexander Monakov
2015-12-01 23:52 ` Bernd Schmidt
2015-12-02 0:23 ` Alexander Monakov
2015-12-07 15:13 ` Nathan Sidwell
2015-12-01 15:28 ` [gomp-nvptx 7/9] nvptx mkoffload: pass -mgomp for OpenMP offloading Alexander Monakov
2015-12-01 15:28 ` [gomp-nvptx 5/9] new target hook: TARGET_SIMT_VF Alexander Monakov
2015-12-01 15:28 ` [gomp-nvptx 2/9] nvptx backend: new "uniform SIMT" codegen variant Alexander Monakov
2015-12-01 16:02 ` Bernd Schmidt
2015-12-01 16:20 ` Alexander Monakov
2015-12-07 15:09 ` Nathan Sidwell
2015-12-02 10:40 ` Jakub Jelinek
2015-12-02 13:02 ` Nathan Sidwell
2015-12-02 13:10 ` Jakub Jelinek
2015-12-02 13:39 ` Nathan Sidwell
2015-12-02 13:46 ` Jakub Jelinek
2015-12-02 14:00 ` Bernd Schmidt
2015-12-02 14:14 ` Nathan Sidwell
2015-12-02 14:22 ` Jakub Jelinek
2015-12-02 14:23 ` Nathan Sidwell
2015-12-02 14:24 ` Jakub Jelinek
2015-12-02 14:34 ` Alexander Monakov
2015-12-02 14:39 ` Nathan Sidwell
2015-12-02 14:54 ` Alexander Monakov
2015-12-02 15:12 ` Jakub Jelinek
2015-12-02 15:18 ` Nathan Sidwell [this message]
[not found] ` <CABtfrpAyUtWub2CBHKYqN0aLNTZ1QspmxyQzOU6Gr+3ogZpSNA@mail.gmail.com>
2015-12-02 16:36 ` Jakub Jelinek
2015-12-02 17:09 ` Nathan Sidwell
2015-12-02 17:09 ` Alexander Monakov
2015-12-02 17:20 ` Nathan Sidwell
2015-12-03 13:57 ` Alexander Monakov
2015-12-02 14:41 ` Alexander Monakov
2015-12-02 14:43 ` Nathan Sidwell
2015-12-01 15:28 ` [gomp-nvptx 4/9] nvptx backend: add -mgomp option and multilib Alexander Monakov
2015-12-02 10:56 ` Jakub Jelinek
2015-12-02 14:18 ` Alexander Monakov
2015-12-03 10:42 ` Alexander Monakov
2015-12-01 15:28 ` [gomp-nvptx 8/9] libgomp: update gomp_nvptx_main for -mgomp Alexander Monakov
2015-12-01 15:56 ` Bernd Schmidt
2015-12-02 11:02 ` Jakub Jelinek
2015-12-01 15:46 ` [gomp-nvptx 3/9] nvptx backend: add two more identifier maps Alexander Monakov
2015-12-01 15:47 ` [gomp-nvptx 1/9] nvptx backend: allow emitting COND_EXEC insns Alexander Monakov
2015-12-02 13:31 ` Bernd Schmidt
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=565F0BB3.5020608@acm.org \
--to=nathan@acm.org \
--cc=amonakov@ispras.ru \
--cc=bschmidt@redhat.com \
--cc=dm@ispras.ru \
--cc=gcc-patches@gcc.gnu.org \
--cc=jakub@redhat.com \
--cc=thomas@codesourcery.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).