From: "Michael V. Zolotukhin" <michael.v.zolotukhin@gmail.com>
To: Nathan Sidwell <nathan@acm.org>
Cc: Jakub Jelinek <jakub@redhat.com>,
Kirill Yukhin <kirill.yukhin@gmail.com>,
Richard Henderson <rth@redhat.com>,
GCC Development <gcc@gcc.gnu.org>,
triegel@redhat.com
Subject: Re: [RFC] Offloading Support in libgomp
Date: Fri, 13 Sep 2013 11:21:00 -0000 [thread overview]
Message-ID: <20130913112035.GB30181@msticlxl57.ims.intel.com> (raw)
In-Reply-To: <5232F12F.6000704@acm.org>
Hi Nathan,
> This is an interesting design. It appears similar to how we'd
> envisioned implementing openacc support -- namely leverage the LTO
> machinery to communicate from the host compiler to the device
> compiler. Your design looks more detailed, which is good.
Thanks, do you have a similar description of your design? It would be pretty
interesting to take a look at it.
> Are you envisioning the device compilers to be stand alone
> compilers, built separately. Or are you envisioning extending the
> configuration machinery by adding something like
> --enable-acclerator=<list> so that:
> .../configure --target=x86_64-linux --enable-accelerator=foo,baz
> causes
> * a build of an x86_64 compiler aware of the foo and baz accelerators
> * build of an appropriate runtime support library
> * a build of a foo lto accelerator backend, assembler (and linker?)
> * (if needed) build of a foo support library
> * a build of a baz lto accelerator backend
> * (if needed) build of a baz support library, assembler (and linker?)
>
> or are you expecting something more like 3 separate configures & build?
> .../configure --target=x86_64-linux --enable-accelerator=foo,baz
> .../configure --target=foo --enable-languages=lto-accelerator
> .../configure --target=baz --enable-languages=lto-accelerator
>
> I'd been imagining the former scheme, as it provides for a more
> integrated build, fwiw.
That's an open question, and we'd like to clarify it too. We'd appreciate any
inputs on this.
Personally, I see actually one more option. Similar of how libgomp figures out
which runtimes are available (by looking for the corresponding plugins), we
could look for available target compilers at compile-time and produce as many
target images as number of compilers we have. Thus, we won't need to rebuild
host compiler to support more targets - we'd just need to place the
corresponding target compiler somewhere. That looks more like your second
option, but differs a bit from it in that we don't need to specify enabled
accelerators.
Michael
> nathan
next prev parent reply other threads:[~2013-09-13 11:21 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-08-22 22:37 Michael V. Zolotukhin
2013-08-23 0:22 ` Jakub Jelinek
2013-08-23 12:16 ` Michael V. Zolotukhin
2013-08-23 12:37 ` Jakub Jelinek
2013-08-24 6:17 ` Michael V. Zolotukhin
2013-08-25 16:24 ` Jakub Jelinek
2013-08-27 0:36 ` Michael V. Zolotukhin
2013-08-27 0:38 ` Jakub Jelinek
2013-08-27 6:16 ` Michael V. Zolotukhin
2013-08-27 8:06 ` Jakub Jelinek
2013-08-27 15:47 ` Michael V. Zolotukhin
2013-08-27 16:22 ` Jakub Jelinek
2013-08-27 19:54 ` Michael V. Zolotukhin
2013-08-28 11:21 ` Jakub Jelinek
2013-08-29 10:44 ` Michael V. Zolotukhin
2013-09-10 15:02 ` Michael V. Zolotukhin
2013-09-10 15:15 ` Jakub Jelinek
2013-09-10 15:31 ` Michael V. Zolotukhin
2013-09-10 15:36 ` Jakub Jelinek
2013-09-10 15:38 ` Michael V. Zolotukhin
2013-09-13 11:30 ` Michael V. Zolotukhin
2013-09-13 12:36 ` Jakub Jelinek
2013-09-13 13:11 ` Michael V. Zolotukhin
2013-09-13 13:16 ` Jakub Jelinek
2013-09-13 15:09 ` Ilya Tocar
2013-09-13 15:34 ` Jakub Jelinek
2014-07-17 7:52 ` Thomas Schwinge
2014-07-17 12:30 ` Ilya Verbin
2014-07-17 12:37 ` Jakub Jelinek
2014-07-17 12:58 ` Thomas Schwinge
2014-07-17 13:09 ` Thomas Schwinge
2014-07-17 13:35 ` Jakub Jelinek
2014-07-17 14:37 ` Thomas Schwinge
2013-09-13 9:35 ` Michael Zolotukhin
2013-09-13 10:52 ` Kirill Yukhin
2013-09-13 11:04 ` Nathan Sidwell
2013-09-13 11:21 ` Michael V. Zolotukhin [this message]
2013-09-16 9:35 ` Jakub Jelinek
2013-09-17 12:05 ` Michael V. Zolotukhin
2013-09-17 12:30 ` Jakub Jelinek
2013-10-28 10:43 ` Ilya Verbin
2013-10-29 8:04 ` Jakub Jelinek
2014-01-31 18:03 ` Ilya Verbin
2014-01-31 19:43 ` Jakub Jelinek
2014-02-14 15:24 ` Ilya Verbin
2014-02-14 15:43 ` Jakub Jelinek
2014-02-14 18:54 ` Richard Henderson
2014-02-17 15:59 ` Ilya Verbin
2014-02-17 16:03 ` Jakub Jelinek
2013-08-28 12:56 ` Richard Biener
2013-08-28 15:26 ` Jakub Jelinek
2013-08-28 17:03 ` Richard Biener
2013-08-28 17:15 ` Jakub Jelinek
2013-08-29 21:09 ` Richard Biener
2013-08-28 18:54 ` Torvald Riegel
2013-08-28 18:43 ` Torvald Riegel
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=20130913112035.GB30181@msticlxl57.ims.intel.com \
--to=michael.v.zolotukhin@gmail.com \
--cc=gcc@gcc.gnu.org \
--cc=jakub@redhat.com \
--cc=kirill.yukhin@gmail.com \
--cc=nathan@acm.org \
--cc=rth@redhat.com \
--cc=triegel@redhat.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).