public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Thomas Schwinge <thomas@codesourcery.com>
To: Ilya Verbin <iverbin@gmail.com>, Jakub Jelinek <jakub@redhat.com>
Cc: Kirill Yukhin <kirill.yukhin@gmail.com>,
	Andrey Turetskiy	<andrey.turetskiy@gmail.com>,
	<gcc-patches@gcc.gnu.org>
Subject: Re: [PATCH 4/4] OpenMP 4.0 offloading to Intel MIC: non-fallback testing
Date: Thu, 18 Dec 2014 15:56:00 -0000	[thread overview]
Message-ID: <87tx0t0yh5.fsf@kepler.schwinge.homeip.net> (raw)
In-Reply-To: <20141030114001.GA34120@msticlxl57.ims.intel.com>

[-- Attachment #1: Type: text/plain, Size: 2685 bytes --]

Hi!

On Thu, 30 Oct 2014 14:40:01 +0300, Ilya Verbin <iverbin@gmail.com> wrote:
> This patch allows to run non-fallback 'make check-target-libgomp'.  It passes to
> the host compiler additional -B options with the paths to the offload compilers,
> since non-installed host compiler doesn't know where to find mkoffload tools.
> Also in case of intelmic offload targets it appends paths to liboffloadmic lib.

> --- a/libgomp/configure.ac
> +++ b/libgomp/configure.ac
> @@ -280,9 +280,13 @@ else
>    multilib_arg=
>  fi
>  
> +# Get accel target and path to install tree of accel compiler
> +offload_additional_options=
> +offload_additional_lib_paths=
>  offload_targets=
>  if test x"$enable_offload_targets" != x; then
>    for tgt in `echo $enable_offload_targets | sed -e 's#,# #g'`; do
> +    tgt_dir=`echo $tgt | grep '=' | sed 's/.*=//'`
>      tgt=`echo $tgt | sed 's/=.*//'`
>      case $tgt in
>        *-intelmic-* | *-intelmicemul-*)
> @@ -295,10 +299,20 @@ if test x"$enable_offload_targets" != x; then
>      else
>        offload_targets=$offload_targets,$tgt_name
>      fi
> +    if test x"$tgt_dir" != x; then
> +      offload_additional_options="$offload_additional_options -B$tgt_dir/libexec/gcc/\$(target_alias)/\$(gcc_version) -B$tgt_dir/bin"
> +      offload_additional_lib_paths="$offload_additional_lib_paths:$tgt_dir/lib64:$tgt_dir/lib"
> +    else
> +      offload_additional_options="$offload_additional_options -B\$(libexecdir)/gcc/\$(target_alias)/\$(gcc_version) -B\$(bindir)"
> +      offload_additional_lib_paths="$offload_additional_lib_paths:$toolexeclibdir"
> +    fi
>    done
>  fi

Hmm, maybe I'm seeing a problem where there isn't one, but in case I'm
not: how will this work if there is more than one offloading compiler
configured?  Won't you get conflicting -B paths added to
offload_additional_options in this case?

> --- a/libgomp/testsuite/lib/libgomp.exp
> +++ b/libgomp/testsuite/lib/libgomp.exp
> @@ -169,6 +186,10 @@ proc libgomp_init { args } {
>  
>      # Disable color diagnostics
>      lappend ALWAYS_CFLAGS "additional_flags=-fdiagnostics-color=never"
> +
> +    # Required to support non-fallback testing of '#pragma omp target'.
> +    # Help GCC to find target mkoffload.
> +    lappend ALWAYS_CFLAGS "additional_flags=${offload_additional_options}"
>  }

What we're doing in OpenACC offloading testing (gomp-4_0-branch), is in
the libgomp.oacc-c/c.exp (etc.) file cycle through all the available
offloading devices, and then in there -- I think -- such options should
be set, that are specific to one particular offloading device/compiler?


Grüße,
 Thomas

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 472 bytes --]

  parent reply	other threads:[~2014-12-18 15:28 UTC|newest]

Thread overview: 116+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-10-21 17:16 [PATCH 0/4] OpenMP 4.0 offloading to Intel MIC Ilya Verbin
2014-10-21 17:20 ` [PATCH 1/4] Add mkoffload for " Ilya Verbin
2014-10-21 21:58   ` Joseph S. Myers
2014-10-21 22:31     ` Ilya Verbin
2014-10-22  7:59       ` Jakub Jelinek
2014-10-22  8:27   ` Jakub Jelinek
2014-10-22 18:57     ` Ilya Verbin
2014-11-03 17:36       ` [gomp4] Mark fopenacc as LTO option Tom de Vries
2015-01-23 15:44         ` [committed][PR64707] Make fopenmp an " Tom de Vries
2015-01-23 15:53           ` [committed][PR64672] Make fopenacc " Tom de Vries
2014-11-06 18:02       ` [PATCH 1/4] Add mkoffload for Intel MIC Jakub Jelinek
2014-12-22 11:25       ` OMP builtins in offloading (was: [PATCH 1/4] Add mkoffload for Intel MIC) Thomas Schwinge
2014-12-22 11:28         ` Jakub Jelinek
2014-12-26 15:49           ` Ilya Verbin
2015-01-08 15:42             ` Thomas Schwinge
2015-01-08 15:49               ` Jakub Jelinek
2015-01-08 16:32                 ` Ilya Verbin
2015-01-08 16:39                   ` Jakub Jelinek
2015-01-09 10:40                     ` Richard Biener
2015-01-09 10:45                       ` Jakub Jelinek
2015-02-16 18:58                 ` Ilya Verbin
2021-08-04  9:46               ` OMP builtins in offloading Thomas Schwinge
2014-12-22 11:27       ` [PATCH 1/4] Add mkoffload for Intel MIC Thomas Schwinge
2014-12-22 11:48         ` Jakub Jelinek
2015-01-08 14:59           ` Thomas Schwinge
2015-01-08 15:02             ` H.J. Lu
2015-01-08 16:21               ` Thomas Schwinge
2015-08-04 11:20               ` Use gcc/coretypes.h:enum offload_abi in mkoffloads (was: [PATCH 1/4] Add mkoffload for Intel MIC) Thomas Schwinge
2015-09-28  8:26                 ` Use gcc/coretypes.h:enum offload_abi in mkoffloads Thomas Schwinge
2015-09-28 11:28                   ` Bernd Schmidt
2015-09-30 10:05                     ` Thomas Schwinge
2015-02-18 11:48       ` [PATCH 1/4] Add mkoffload for Intel MIC Thomas Schwinge
2015-02-18 11:56         ` Ilya Verbin
2015-02-18 12:23           ` [PATCH] Use automatic dependencies for mkoffload.o Jakub Jelinek
2015-02-18 13:10             ` Richard Biener
2015-02-18 12:01         ` [PATCH 1/4] Add mkoffload for Intel MIC Jakub Jelinek
2015-09-28  9:39       ` Thomas Schwinge
2015-09-28 11:26         ` Bernd Schmidt
2015-09-28 11:27           ` Bernd Schmidt
2015-09-28 12:00             ` Ilya Verbin
2015-09-28 12:05               ` Bernd Schmidt
2015-09-28 12:33                 ` Jakub Jelinek
2015-09-28 12:33                   ` Bernd Schmidt
2015-09-29 11:24                     ` Richard Biener
2015-09-29 12:11                       ` Bernd Schmidt
2015-09-30 17:55                 ` Refactor intelmic-mkoffload.c argv building to use obstacks (was: [PATCH 1/4] Add mkoffload for Intel MIC) Thomas Schwinge
2015-03-06 13:55   ` [PATCH] Fix intelmic-mkoffload " Ilya Verbin
2015-03-09 14:49     ` Jakub Jelinek
2014-10-21 17:24 ` [PATCH 2/4] Add liboffloadmic Ilya Verbin
2014-10-22  8:55   ` Jakub Jelinek
2014-10-22 17:13     ` Joseph S. Myers
2014-10-22 19:31     ` Ilya Verbin
2014-10-29 16:31       ` Ilya Verbin
2014-11-06 18:21       ` Jakub Jelinek
2014-11-12 10:52         ` Jakub Jelinek
2014-11-12 13:29           ` Ilya Verbin
2014-12-12 10:46   ` Thomas Schwinge
2015-02-04 17:45     ` Ilya Verbin
2015-02-04 17:46       ` Jakub Jelinek
2015-07-09 10:00   ` Thomas Schwinge
2015-07-13 14:31     ` Ilya Verbin
2014-10-21 17:28 ` [PATCH 3/4] Add libgomp plugin for Intel MIC Ilya Verbin
2014-10-22  9:47   ` Jakub Jelinek
2014-10-23 16:00     ` Ilya Verbin
2014-10-24 14:57       ` Jakub Jelinek
2014-10-24 15:12         ` Ilya Verbin
2014-10-24 15:19           ` Jakub Jelinek
2014-10-27 14:24             ` Ilya Verbin
2014-11-06 18:25               ` Jakub Jelinek
2014-11-10 14:32                 ` Ilya Verbin
2014-11-11  7:07                   ` Jakub Jelinek
2014-12-12  9:42                   ` Thomas Schwinge
2015-01-08 14:48                     ` Thomas Schwinge
2015-07-08 14:16   ` Thomas Schwinge
2015-07-08 15:14     ` Ilya Verbin
2015-07-08 15:52       ` Thomas Schwinge
2015-07-23 19:05     ` Ilya Verbin
2015-07-24  8:06       ` Jakub Jelinek
2015-07-24 14:27         ` David Malcolm
2015-07-28 15:51           ` Maxim Blumental
2015-08-03 10:24             ` Maxim Blumental
2015-08-04 17:40               ` David Malcolm
2015-08-06 14:35             ` Fwd: " Maxim Blumental
2015-08-11 12:27               ` Maxim Blumental
2015-08-24  8:51               ` Fwd: " Jakub Jelinek
2014-10-30 12:45 ` [PATCH 4/4] OpenMP 4.0 offloading to Intel MIC: non-fallback testing Ilya Verbin
2014-11-06 17:55   ` Jakub Jelinek
2014-11-10 14:51     ` Ilya Verbin
2014-11-11  7:10       ` Jakub Jelinek
2014-11-12  9:18       ` Jakub Jelinek
2014-12-17 22:53       ` Thomas Schwinge
2014-12-18 10:46         ` Jakub Jelinek
2014-12-22 16:37           ` Thomas Schwinge
2014-12-18 15:56   ` Thomas Schwinge [this message]
2014-12-18 17:43     ` Ilya Verbin
2014-12-18 17:56       ` Jakub Jelinek
2014-12-22 11:49       ` Thomas Schwinge
2014-12-22 12:50         ` Jakub Jelinek
2015-01-15 19:21           ` Ilya Verbin
2015-01-15 19:25             ` Jakub Jelinek
2015-01-28 17:28               ` Ilya Verbin
2015-01-28 17:42                 ` Jakub Jelinek
2015-01-28 17:51                   ` Ilya Verbin
2015-02-02 14:03                     ` Ilya Verbin
2014-12-26 20:53         ` Ilya Verbin
2015-01-08 16:03           ` Thomas Schwinge
2016-03-13 19:10   ` Thomas Schwinge
2014-12-22 12:08 ` [PATCH 0/4] OpenMP 4.0 offloading to Intel MIC Thomas Schwinge
2015-10-22 18:28   ` Ilya Verbin
2015-10-23  8:10     ` Jakub Jelinek
2015-10-26 14:39       ` Ilya Verbin
2014-11-21 18:44 [PATCH 4/4] OpenMP 4.0 offloading to Intel MIC: non-fallback testing Bernd Edlinger
2014-11-21 18:59 ` Ilya Verbin
2014-11-21 19:01   ` Bernd Edlinger
2014-11-21 19:34     ` Ilya Verbin
2014-11-21 20:07       ` Jakub Jelinek

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=87tx0t0yh5.fsf@kepler.schwinge.homeip.net \
    --to=thomas@codesourcery.com \
    --cc=andrey.turetskiy@gmail.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=iverbin@gmail.com \
    --cc=jakub@redhat.com \
    --cc=kirill.yukhin@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).