Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
CC: marxin at gcc dot gnu.org
Target Milestone: ---
I have a linking error when building miniQMC.
The offload functions are first archived in libqmcwfs.a and then linked to the
final binary
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92553
--- Comment #2 from Ye Luo ---
PR92357 mentioned that you backported a fix to the gcc 9 branch and how I'm
getting the following error
/home/yeluo/opt/gcc-9-branch/bin/g++ -fopenmp -foffload=nvptx-none
-foffload=-lm -fomit-frame-pointer
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95023
--- Comment #2 from Ye Luo ---
I got the following error when building the new lib. Tried gcc 8 and 10 with
master and newlib-3.3.0. No difference
/home/yeluo/opt/build-amdgcn-gcc/./gcc/xgcc
-B/home/yeluo/opt/build-amdgcn-gcc/./gcc/ -nostdinc
: target
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
The wiki says
Next insert a symbolic to the Newlib source's "newlib" directory into the
directory containing the gcc sources. The Newlib version needs to be
cont
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
https://github.com/ye-luo/openmp-target/blob/master/hands-on/tests/map/map_class_member.cpp
$ g++ -fopenmp map_class_member.cpp
map_class_member.cpp: In instantiation of ‘maptest
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95474
--- Comment #2 from Ye Luo ---
@jakub could you mention which page and line in 4.5 spec this code violates?
https://www.openmp.org/wp-content/uploads/OpenMP-API-Specification-5.0.pdf
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95474
--- Comment #4 from Ye Luo ---
Thanks. Hopefully this part of 5.0 feature will be implemented in the near
future.
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
C version code works. The corresponding Fortran version is blocked by the
compiler errror.
C version:
#include
int main()
{
int a[1];
int* b = a;
a[0] = 0;
#pragma omp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98476
--- Comment #2 from Ye Luo ---
@jakub thanks for the quick reply. However, even if I add map(to: a) which is
almost a no-op to satisfy the compiler for 4.5 spec. The printout result is
still wrong. It should be 0 and 1 but the fortran case
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98476
--- Comment #3 from Ye Luo ---
I verified that current master has removed the syntax restriction. However, the
printout remains incorrect.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98858
--- Comment #1 from Ye Luo ---
GNU Fortran (GCC) 11.0.0 20210127 (experimental)
: fortran
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
Getting ICE
yeluo@ryzen-box:~/opt/openmp-target/hands-on/tests/fortran_use_device_ptr$
gfortran -fopenmp test_use_device_ptr_target.f90
test_use_device_ptr_target.f90
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
Using std::complex type in offload region is highly desired.
$ g++ -fopenmp complex_reduction.cpp
ptxas /tmp/cceLNaYr.o, line 484; error : Label expected for argument 0
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
this restriction seems unnecessary.
/home/yeluo/opt/qmcpack/src/Particle/SoaDistanceTableABOMPTarget.h:63:42:
error: ‘this’ allowed in OpenMP only in ‘declare simd’ clauses
63
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98869
--- Comment #1 from Ye Luo ---
Full source code at
https://github.com/QMCPACK/qmcpack/blob/a23584738485bb7c948d0ba1841c669fe76500b6/src/Particle/SoaDistanceTableABOMPTarget.h#L63
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98862
--- Comment #3 from Ye Luo ---
Cool. -foffload=-latomic resolved my problem.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66756
Ye Luo changed:
What|Removed |Added
CC||xw111luoye at gmail dot com
--- Comment #18
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102204
--- Comment #2 from Ye Luo ---
User code has been written in mostly 5.0 fashion as 4.5 has too many
restrictions. For most users, they only need a subset of 5.0 feature with
up-to-date compilers and fully implemented 5.0 is not expected.
Back
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98869
Ye Luo changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98869
--- Comment #3 from Ye Luo ---
This doesn't work with gcc 11.2 but works on devel/omp/gcc-11 branch.
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
With branch devel/omp/gcc-11
I'm getting
/home/yeluo/opt/qmcpack/build_rtx3060_gcc_offload_real/src/config.h:42:29:
error: array section does not have mappable type in ‘map’ clause
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
x86_64-pc-linux-gnu-accel-nvptx-none-gcc needs both -misa and -mptx to select a
specific target.
When I use OpenMP offload.
-foffload
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
miniQMC https://github.com/ye-luo/miniqmc
ships a set of OpenMP offload tests.
```
cmake -DCMAKE_CXX_COMPILER=g++ -DENABLE_OFFLOAD=ON ..
make -j32
```
When those tests are compiled
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102204
--- Comment #3 from Ye Luo ---
I tried today's main 3f0fcda37f58d4108feb67de08f181a32bcb6388.
This issue persists. Any chance this will be resolved in 12 release?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104285
--- Comment #1 from Ye Luo ---
Here is a minimal reproducer.
https://github.com/ye-luo/openmp-target/tree/master/tests/linking/two_identical_templates
$ g++ -fopenmp -foffload=nvptx-none -O3 -c test_a.cpp
$ g++ -fopenmp -foffload=nvptx-none
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
See reproducer
https://github.com/ye-luo/openmp-target/blob/master/tests/map/struct_with_const.cpp
$ g++ -fopenmp -foffload=nvptx-none struct_with_const.cpp
struct_with_const.cpp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104493
Ye Luo changed:
What|Removed |Added
CC||burnus at gcc dot gnu.org
--- Comment #1 from
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104493
--- Comment #3 from Ye Luo ---
Thank you for the answer with well explained development procedure of gcc. I
was intended to keep this bug report active. I just adopted OpenMP offload
options adjusted in gcc 12 but OG11 doesn't have them due to
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104493
--- Comment #4 from Ye Luo ---
When I tried master and devel/omp/gcc-12, the fix has not been applied.
Just keep a note here. Probably this has been planned.
++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
My application runs in parallel on Power9 RHEL 8.2 using gcc 9.3. The failure
doesn't show up in small scale runs. However, when running with 24576(4096*6)
MPI processes,
code:
void* ArgMappers
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105299
--- Comment #3 from Ye Luo ---
Unfortunately there is no easy reproducer.
Even without the printf. I got the same failure.
Will try with a debugger.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106829
Ye Luo changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
Resolution|---
: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: xw111luoye at gmail dot com
Target Milestone: ---
Created attachment 53537
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53537=edit
reproducer source c
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98862
Ye Luo changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resolution|---
34 matches
Mail list logo
