[Bug libgomp/65099] New: nvptx offloading: hard-coded 64-bit assumptions

[Bug libgomp/65099] New: nvptx offloading: hard-coded 64-bit assumptions

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

            Bug ID: 65099
           Summary: nvptx offloading: hard-coded 64-bit assumptions
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Keywords: openacc
          Severity: normal
          Priority: P3
         Component: libgomp
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tschwinge at gcc dot gnu.org
                CC: jakub at gcc dot gnu.org

Primarily (only?) in libgomp, there are a few hard-coded 64-bit assumptions for
nvptx offloading, so 32-bit x86 (-m32) nvptx offloading doesn't currently work.

Also, the nvptx offloading compiler (that is, the compiler support libraries)
will then have to be built for both 32-bit and 64-bit multilibs.

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

--- Comment #1 from Thomas Schwinge <tschwinge at gcc dot gnu.org> ---
Author: tschwinge
Date: Wed Jul  8 14:59:59 2015
New Revision: 225560

URL: https://gcc.gnu.org/viewcvs?rev=225560&root=gcc&view=rev
Log:
[nvptx offloading] Only 64-bit configurations are currently supported

        PR libgomp/65099
        gcc/
        * config/nvptx/mkoffload.c (main): Create an offload image only in
        64-bit configurations.
        libgomp/
        * plugin/plugin-nvptx.c (nvptx_get_num_devices): Return 0 if not
        in a 64-bit configuration.
        * testsuite/libgomp.oacc-c++/c++.exp: Don't attempt nvidia
        offloading testing if no such device is available.
        * testsuite/libgomp.oacc-c/c.exp: Likewise.
        * testsuite/libgomp.oacc-fortran/fortran.exp: Likewise.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/nvptx/mkoffload.c
    trunk/libgomp/ChangeLog
    trunk/libgomp/plugin/plugin-nvptx.c
    trunk/libgomp/testsuite/libgomp.oacc-c++/c++.exp
    trunk/libgomp/testsuite/libgomp.oacc-c/c.exp
    trunk/libgomp/testsuite/libgomp.oacc-fortran/fortran.exp

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

--- Comment #2 from Thomas Schwinge <tschwinge at gcc dot gnu.org> ---
Author: tschwinge
Date: Wed Jul 15 11:55:13 2015
New Revision: 225824

URL: https://gcc.gnu.org/viewcvs?rev=225824&root=gcc&view=rev
Log:
[PR libgomp/65099] nvptx offloading: only 64-bit configurations are currently
supported

Backport trunk r225560:

        PR libgomp/65099
        gcc/
        * config/nvptx/mkoffload.c (main): Create an offload image only in
        64-bit configurations.
        libgomp/
        * plugin/plugin-nvptx.c (nvptx_get_num_devices): Return 0 if not
        in a 64-bit configuration.
        * testsuite/libgomp.oacc-c++/c++.exp: Don't attempt nvidia
        offloading testing if no such device is available.
        * testsuite/libgomp.oacc-c/c.exp: Likewise.
        * testsuite/libgomp.oacc-fortran/fortran.exp: Likewise.

Modified:
    branches/gcc-5-branch/gcc/ChangeLog
    branches/gcc-5-branch/gcc/config/nvptx/mkoffload.c
    branches/gcc-5-branch/libgomp/ChangeLog
    branches/gcc-5-branch/libgomp/plugin/plugin-nvptx.c
    branches/gcc-5-branch/libgomp/testsuite/libgomp.oacc-c++/c++.exp
    branches/gcc-5-branch/libgomp/testsuite/libgomp.oacc-c/c.exp
    branches/gcc-5-branch/libgomp/testsuite/libgomp.oacc-fortran/fortran.exp

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[amonakov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

Alexander Monakov <amonakov at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |amonakov at gcc dot gnu.org

--- Comment #3 from Alexander Monakov <amonakov at gcc dot gnu.org> ---
As of CUDA Toolkit release 7.0 (Early 2015), NVIDIA removed support for running
CUDA applications on 32-bit x86 kernels, and deprecated support for running
32-bit applications on 64-bit x86 kernels.  However, 32-bit support is in place
for the ARM platform.

Thought I'd add this note since the need for 32-bit support in GCC offloading
in the first place is non-obvious.

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Thomas Schwinge <tschwinge@gcc.gnu.org>:

https://gcc.gnu.org/g:6106dfb9f73a33c87108ad5b2dcd4842bdd7828e

commit r11-6677-g6106dfb9f73a33c87108ad5b2dcd4842bdd7828e
Author: Thomas Schwinge <thomas@codesourcery.com>
Date:   Mon Nov 30 15:15:20 2020 +0100

    [nvptx libgomp plugin] Build only in supported configurations

    As recently again discussed in <https://gcc.gnu.org/PR97436> "[nvptx] -m32
    support", nvptx offloading other than for 64-bit host has never been
    implemented, tested, supported.  So we simply should buildn't the nvptx
libgomp
    plugin in this case.

    This avoids build problems if, for example, in a (standard) bi-arch
    x86_64-pc-linux-gnu '-m64'/'-m32' build, libcuda is available only in a
64-bit
    variant but not in a 32-bit one, which, for example, is the case if you
build
    GCC against the CUDA toolkit's 'stubs/libcuda.so' (see
    <https://stackoverflow.com/a/52784819>).

    This amends PR65099 commit a92defdab79a1268f4b9dcf42b937e4002a4cf15
(r225560)
    "[nvptx offloading] Only 64-bit configurations are currently supported" to
    match the way we're doing this for the HSA/GCN plugins.

            libgomp/
            PR libgomp/65099
            * plugin/configfrag.ac (PLUGIN_NVPTX): Restrict to supported
            configurations.
            * configure: Regenerate.
            * plugin/plugin-nvptx.c (nvptx_get_num_devices): Remove 64-bit
            check.

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Thomas Schwinge
<tschwinge@gcc.gnu.org>:

https://gcc.gnu.org/g:1e56a7c9a6631b217299b2ddcd5c4d497bb3445e

commit r10-9270-g1e56a7c9a6631b217299b2ddcd5c4d497bb3445e
Author: Thomas Schwinge <thomas@codesourcery.com>
Date:   Mon Nov 30 15:15:20 2020 +0100

    [nvptx libgomp plugin] Build only in supported configurations

    As recently again discussed in <https://gcc.gnu.org/PR97436> "[nvptx] -m32
    support", nvptx offloading other than for 64-bit host has never been
    implemented, tested, supported.  So we simply should buildn't the nvptx
libgomp
    plugin in this case.

    This avoids build problems if, for example, in a (standard) bi-arch
    x86_64-pc-linux-gnu '-m64'/'-m32' build, libcuda is available only in a
64-bit
    variant but not in a 32-bit one, which, for example, is the case if you
build
    GCC against the CUDA toolkit's 'stubs/libcuda.so' (see
    <https://stackoverflow.com/a/52784819>).

    This amends PR65099 commit a92defdab79a1268f4b9dcf42b937e4002a4cf15
(r225560)
    "[nvptx offloading] Only 64-bit configurations are currently supported" to
    match the way we're doing this for the HSA/GCN plugins.

            libgomp/
            PR libgomp/65099
            * plugin/configfrag.ac (PLUGIN_NVPTX): Restrict to supported
            configurations.
            * configure: Regenerate.
            * plugin/plugin-nvptx.c (nvptx_get_num_devices): Remove 64-bit
            check.

    (cherry picked from commit 6106dfb9f73a33c87108ad5b2dcd4842bdd7828e)

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Thomas Schwinge
<tschwinge@gcc.gnu.org>:

https://gcc.gnu.org/g:0f1e1069a753e912b058f0d4bf599f0edde28408

commit r9-9182-g0f1e1069a753e912b058f0d4bf599f0edde28408
Author: Thomas Schwinge <thomas@codesourcery.com>
Date:   Mon Nov 30 15:15:20 2020 +0100

    [nvptx libgomp plugin] Build only in supported configurations

    As recently again discussed in <https://gcc.gnu.org/PR97436> "[nvptx] -m32
    support", nvptx offloading other than for 64-bit host has never been
    implemented, tested, supported.  So we simply should buildn't the nvptx
libgomp
    plugin in this case.

    This avoids build problems if, for example, in a (standard) bi-arch
    x86_64-pc-linux-gnu '-m64'/'-m32' build, libcuda is available only in a
64-bit
    variant but not in a 32-bit one, which, for example, is the case if you
build
    GCC against the CUDA toolkit's 'stubs/libcuda.so' (see
    <https://stackoverflow.com/a/52784819>).

    This amends PR65099 commit a92defdab79a1268f4b9dcf42b937e4002a4cf15
(r225560)
    "[nvptx offloading] Only 64-bit configurations are currently supported" to
    match the way we're doing this for the HSA/GCN plugins.

            libgomp/
            PR libgomp/65099
            * plugin/configfrag.ac (PLUGIN_NVPTX): Restrict to supported
            configurations.
            * configure: Regenerate.
            * plugin/plugin-nvptx.c (nvptx_get_num_devices): Remove 64-bit
            check.

    (cherry picked from commit 6106dfb9f73a33c87108ad5b2dcd4842bdd7828e)

[Bug libgomp/65099] nvptx offloading: hard-coded 64-bit assumptions

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65099

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Thomas Schwinge
<tschwinge@gcc.gnu.org>:

https://gcc.gnu.org/g:f9267925c648f2ccd9e4680b699e581003125bcf

commit r8-10728-gf9267925c648f2ccd9e4680b699e581003125bcf
Author: Thomas Schwinge <thomas@codesourcery.com>
Date:   Mon Nov 30 15:15:20 2020 +0100

    [nvptx libgomp plugin] Build only in supported configurations

    As recently again discussed in <https://gcc.gnu.org/PR97436> "[nvptx] -m32
    support", nvptx offloading other than for 64-bit host has never been
    implemented, tested, supported.  So we simply should buildn't the nvptx
libgomp
    plugin in this case.

    This avoids build problems if, for example, in a (standard) bi-arch
    x86_64-pc-linux-gnu '-m64'/'-m32' build, libcuda is available only in a
64-bit
    variant but not in a 32-bit one, which, for example, is the case if you
build
    GCC against the CUDA toolkit's 'stubs/libcuda.so' (see
    <https://stackoverflow.com/a/52784819>).

    This amends PR65099 commit a92defdab79a1268f4b9dcf42b937e4002a4cf15
(r225560)
    "[nvptx offloading] Only 64-bit configurations are currently supported" to
    match the way we're doing this for the HSA/GCN plugins.

            libgomp/
            PR libgomp/65099
            * plugin/configfrag.ac (PLUGIN_NVPTX): Restrict to supported
            configurations.
            * configure: Regenerate.
            * plugin/plugin-nvptx.c (nvptx_get_num_devices): Remove 64-bit
            check.

    (cherry picked from commit 6106dfb9f73a33c87108ad5b2dcd4842bdd7828e)