[Bug target/95864] New: [11 Regression] GCN offloading execution regressions after commit f062c3f11505b70c5275e5bc0e52f3e441f8afbc "amdgcn: Switch to HSACO v3 binary format"

[Bug target/95864] New: [11 Regression] GCN offloading execution regressions after commit f062c3f11505b70c5275e5bc0e52f3e441f8afbc "amdgcn: Switch to HSACO v3 binary format"

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95864

            Bug ID: 95864
           Summary: [11 Regression] GCN offloading execution regressions
                    after commit f062c3f11505b70c5275e5bc0e52f3e441f8afbc
                    "amdgcn: Switch to HSACO v3 binary format"
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Keywords: openmp
          Severity: normal
          Priority: P3
         Component: target
          Assignee: ams at gcc dot gnu.org
          Reporter: tschwinge at gcc dot gnu.org
                CC: jules at gcc dot gnu.org
  Target Milestone: ---
            Target: gcn

I'm seeing the following GCN offloading execution regressions after commit
f062c3f11505b70c5275e5bc0e52f3e441f8afbc "amdgcn: Switch to HSACO v3 binary
format":

    PASS: libgomp.c/../libgomp.c-c++-common/for-5.c (test for excess errors)
    [-PASS:-]{+WARNING: program timed out.+}
    {+FAIL:+} libgomp.c/../libgomp.c-c++-common/for-5.c execution test

    libgomp: GCN fatal error: Asynchronous queue error
    Runtime message: HSA_STATUS_ERROR_INVALID_ISA: The instruction set
architecture is invalid.

    WARNING: program timed out.
    FAIL: libgomp.c/../libgomp.c-c++-common/for-5.c execution test

Same for C++.

    PASS: libgomp.c/../libgomp.c-c++-common/for-6.c (test for excess errors)
    [-PASS:-]{+WARNING: program timed out.+}
    {+FAIL:+} libgomp.c/../libgomp.c-c++-common/for-6.c execution test

Same as above, same for C++.

    @@ -3456,9 +3460,11 @@ PASS: libgomp.fortran/optional-map.f90   -O1 
execution test
    PASS: libgomp.fortran/optional-map.f90   -O2  (test for excess errors)
    PASS: libgomp.fortran/optional-map.f90   -O2  execution test
    PASS: libgomp.fortran/optional-map.f90   -O3 -fomit-frame-pointer
-funroll-loops -fpeel-loops -ftracer -finline-functions  (test for excess
errors)
    [-PASS:-]{+WARNING: program timed out.+}
    {+FAIL:+} libgomp.fortran/optional-map.f90   -O3 -fomit-frame-pointer
-funroll-loops -fpeel-loops -ftracer -finline-functions  execution test
    PASS: libgomp.fortran/optional-map.f90   -O3 -g  (test for excess errors)
    [-PASS:-]{+WARNING: program timed out.+}
    {+FAIL:+} libgomp.fortran/optional-map.f90   -O3 -g  execution test
    PASS: libgomp.fortran/optional-map.f90   -Os  (test for excess errors)
    PASS: libgomp.fortran/optional-map.f90   -Os  execution test

Same as above.

    @@ -4323,9 +4329,11 @@ PASS: libgomp.fortran/target1.f90   -O1  execution
test
    PASS: libgomp.fortran/target1.f90   -O2  (test for excess errors)
    PASS: libgomp.fortran/target1.f90   -O2  execution test
    PASS: libgomp.fortran/target1.f90   -O3 -fomit-frame-pointer -funroll-loops
-fpeel-loops -ftracer -finline-functions  (test for excess errors)
    [-PASS:-]{+WARNING: program timed out.+}
    {+FAIL:+} libgomp.fortran/target1.f90   -O3 -fomit-frame-pointer
-funroll-loops -fpeel-loops -ftracer -finline-functions  execution test
    PASS: libgomp.fortran/target1.f90   -O3 -g  (test for excess errors)
    [-PASS:-]{+WARNING: program timed out.+}
    {+FAIL:+} libgomp.fortran/target1.f90   -O3 -g  execution test
    PASS: libgomp.fortran/target1.f90   -Os  (test for excess errors)
    PASS: libgomp.fortran/target1.f90   -Os  execution test

Same as above.


That's testing the default multilib on "Advanced Micro Devices, Inc. [AMD/ATI]
Fiji [Radeon R9 FURY / NANO Series] (rev ca)" hardware.  (I haven't tested
anything else.)

There seems to be some bug?


But also:


The error reporting seems "strange"?


We shouldn't run into a timeout?  (Missing unlocking on error path?)

[Bug target/95864] [11 Regression] GCN offloading execution regressions after commit f062c3f11505b70c5275e5bc0e52f3e441f8afbc "amdgcn: Switch to HSACO v3 binary format"

[ams at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95864

--- Comment #1 from Andrew Stubbs <ams at gcc dot gnu.org> ---
I'm aware of these issues.

I fixed all the test failures that were definitely bugs in the HSACOv3
implementation, and the ones that remain appear to be either latent bugs
uncovered by the new driver configuration, or possibly even not our problem.

Of course, I could be mistaken, but I won't find out until I find time to look
closer.

[Bug target/95864] [11 Regression] GCN offloading execution regressions after commit f062c3f11505b70c5275e5bc0e52f3e441f8afbc "amdgcn: Switch to HSACO v3 binary format"

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95864

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |11.0

[Bug target/95864] [11 Regression] GCN offloading execution regressions after commit f062c3f11505b70c5275e5bc0e52f3e441f8afbc "amdgcn: Switch to HSACO v3 binary format"

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95864

Thomas Schwinge <tschwinge at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED
           Assignee|ams at gcc dot gnu.org             |jules at gcc dot gnu.org

--- Comment #2 from Thomas Schwinge <tschwinge at gcc dot gnu.org> ---
This apparently got fixed by Julian's commit
r11-3057-g3aee3aaf48be2d3d81e381690ae9dd305d8b505f "openacc: Fix mkoffload
SGPR/VGPR count parsing for HSACO v3".