[petsc-dev] Question on PETSc + CUDA configuration with MPI on cluster
Satish Balay
balay.anl at fastmail.org
Tue Sep 23 00:13:34 CDT 2025
orte-info output does suggest OpenMPI is built with cuda enabled.
Are you able to run PETSc examples? What do you get for:
>>>>
balay at petsc-gpu-01:/scratch/balay/petsc/src/snes/tutorials$ make ex19
/scratch/balay/petsc/arch-linux-c-debug/bin/mpicc -fPIC -Wall
-Wwrite-strings -Wno-unknown-pragmas -Wno-lto-type-mismatch
-Wno-stringop-overflow -fstack-protector -fvisibility=hidden -g3 -O0
-I/scratch/balay/petsc/include
-I/scratch/balay/petsc/arch-linux-c-debug/include
-I/nfs/gce/projects/petsc/soft/u22.04/spack-2024-11-27-cuda/opt/spack/linux-ubuntu22.04-x86_64/gcc-11.4.0/cuda-12.0.1-gy7foq57oi6wzltombtsdy5eqz5gkjgc/include
-Wl,-export-dynamic ex19.c
-Wl,-rpath,/scratch/balay/petsc/arch-linux-c-debug/lib
-L/scratch/balay/petsc/arch-linux-c-debug/lib
-Wl,-rpath,/nfs/gce/projects/petsc/soft/u22.04/spack-2024-11-27-cuda/opt/spack/linux-ubuntu22.04-x86_64/gcc-11.4.0/cuda-12.0.1-gy7foq57oi6wzltombtsdy5eqz5gkjgc/lib64
-L/nfs/gce/projects/petsc/soft/u22.04/spack-2024-11-27-cuda/opt/spack/linux-ubuntu22.04-x86_64/gcc-11.4.0/cuda-12.0.1-gy7foq57oi6wzltombtsdy5eqz5gkjgc/lib64
-L/nfs/gce/projects/petsc/soft/u22.04/spack-2024-11-27-cuda/opt/spack/linux-ubuntu22.04-x86_64/gcc-11.4.0/cuda-12.0.1-gy7foq57oi6wzltombtsdy5eqz5gkjgc/lib64/stubs
-Wl,-rpath,/scratch/balay/petsc/arch-linux-c-debug/lib
-L/scratch/balay/petsc/arch-linux-c-debug/lib
-Wl,-rpath,/usr/lib/gcc/x86_64-linux-gnu/11
-L/usr/lib/gcc/x86_64-linux-gnu/11 -lpetsc -llapack -lblas -lm -lcudart
-lnvToolsExt -lcufft -lcublas -lcusparse -lcusolver -lcurand -lcuda
-lX11 -lmpi_usempif08 -lmpi_usempi_ignore_tkr -lmpi_mpifh -lmpi
-lgfortran -lm -lgfortran -lm -lgcc_s -lquadmath -lstdc++ -o ex19
balay at petsc-gpu-01:/scratch/balay/petsc/src/snes/tutorials$ ./ex19 -snes_monitor -dm_mat_type seqaijcusparse -dm_vec_type seqcuda -pc_type gamg -pc_gamg_esteig_ksp_max_it 10 -ksp_monitor -mg_levels_ksp_max_it 3
lid velocity = 0.0625, prandtl # = 1., grashof # = 1.
0 SNES Function norm 2.391552133017e-01
0 KSP Residual norm 2.013462697105e-01
1 KSP Residual norm 5.027022294231e-02
2 KSP Residual norm 7.248258907839e-03
3 KSP Residual norm 8.590847505363e-04
4 KSP Residual norm 1.511762118013e-05
5 KSP Residual norm 1.410585959219e-06
1 SNES Function norm 6.812362089434e-05
0 KSP Residual norm 2.315252918142e-05
1 KSP Residual norm 2.351994603807e-06
2 KSP Residual norm 3.882072626158e-07
3 KSP Residual norm 2.227447016095e-08
4 KSP Residual norm 2.200353394658e-09
5 KSP Residual norm 1.147903850265e-10
2 SNES Function norm 3.411489611752e-10
Number of SNES iterations = 2
balay at petsc-gpu-01:/scratch/balay/petsc/src/snes/tutorials$
<<<<
So what issue are you seeing with your code? And does it go away with the option: "-use_gpu_aware_mpi 0"? for example:
>>>>
balay at petsc-gpu-01:/scratch/balay/petsc/src/snes/tutorials$ ./ex19 -snes_monitor -dm_mat_type seqaijcusparse -dm_vec_type seqcuda -pc_type gamg -pc_gamg_esteig_ksp_max_it 10 -ksp_monitor -mg_levels_ksp_max_it 3 -use_gpu_aware_mpi 0
lid velocity = 0.0625, prandtl # = 1., grashof # = 1.
0 SNES Function norm 2.391552133017e-01
0 KSP Residual norm 2.013462697105e-01
1 KSP Residual norm 5.027022294231e-02
2 KSP Residual norm 7.248258907839e-03
3 KSP Residual norm 8.590847505363e-04
4 KSP Residual norm 1.511762118013e-05
5 KSP Residual norm 1.410585959219e-06
1 SNES Function norm 6.812362089434e-05
0 KSP Residual norm 2.315252918142e-05
1 KSP Residual norm 2.351994603807e-06
2 KSP Residual norm 3.882072626158e-07
3 KSP Residual norm 2.227447016095e-08
4 KSP Residual norm 2.200353394658e-09
5 KSP Residual norm 1.147903850265e-10
2 SNES Function norm 3.411489611752e-10
Number of SNES iterations = 2
balay at petsc-gpu-01:/scratch/balay/petsc/src/snes/tutorials$
<<<<
Satish
On Tue, 23 Sep 2025, 岳新海 wrote:
> I get:
> [mae_yuexh at login01 ~]$ orte-info |grep 'MCA btl'
> MCA btl: smcuda (MCA v2.1, API v3.1, Component v4.1.5)
> MCA btl: tcp (MCA v2.1, API v3.1, Component v4.1.5)
> MCA btl: self (MCA v2.1, API v3.1, Component v4.1.5)
> MCA btl: vader (MCA v2.1, API v3.1, Component v4.1.5)
>
>
>
> Xinhai
>
>
>
>
>
> 岳新海
>
>
>
> 南方科技大学/学生/研究生/2023级研究生
>
>
>
> 广东省深圳市南山区学苑大道1088号
>
>
>
>
>
>
>
> ------------------ Original ------------------
> From: "Satish Balay"<balay.anl at fastmail.org>;
> Date: Tue, Sep 23, 2025 03:25 AM
> To: "岳新海"<12332508 at mail.sustech.edu.cn>;
> Cc: "petsc-dev"<petsc-dev at mcs.anl.gov>;
> Subject: Re: [petsc-dev] Question on PETSc + CUDA configuration with MPI on cluster
>
>
>
>
> What do you get for (with your openmpi install) :orte-info |grep 'MCA btl'
>
> With cuda built openmpi - I get:
> balay at petsc-gpu-01:/scratch/balay/petsc$ ./arch-linux-c-debug/bin/orte-info |grep 'MCA btl'
> MCA btl: smcuda (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: openib (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: self (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: tcp (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: vader (MCA v2.1, API v3.1, Component v4.1.6)
>
> And without cuda:
> balay at petsc-gpu-01:/scratch/balay/petsc.x$ ./arch-test/bin/orte-info | grep 'MCA btl'
> MCA btl: openib (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: self (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: tcp (MCA v2.1, API v3.1, Component v4.1.6)
> MCA btl: vader (MCA v2.1, API v3.1, Component v4.1.6)
>
> i.e "smcuda" should be listed for a cuda enabled openmpi.
>
> Its not clear if GPU-aware MPI makes a difference for all MPI impls (or versions) - so good to verify. [its a performance issue anyway - so primarily useful when performing timing measurements]
>
> Satish
>
> On Mon, 22 Sep 2025, 岳新海 wrote:
>
> > Dear PETSc Team,
> >
> > I am encountering an issue when running PETSc with CUDA support on a cluster. When I set the vector type to VECCUDA, PETSc reports that my MPI is not GPU-aware. However, the MPI library (OpenMPI 4.1.5) I used to configure PETSc was built with the --with-cuda option enabled.
> >
> >
> > Here are some details:
> > PETSc version: 3.20.6
> > MPI: OpenMPI 4.1.5, configured with --with-cuda
> > GPU: RTX3090
> > CUDA version: 12.1
> > I have attached both my PETSc configure command and OpenMPI configure command for reference.
> >
> > My questions are:
> >
> >
> >
> >
> > Even though I enabled --with-cuda in OpenMPI, why does PETSc still report that MPI is not GPU-aware?
> >
> >
> >
> > Are there additional steps or specific configuration flags required (either in OpenMPI or PETSc) to ensure GPU-aware MPI is correctly detected?
> >
> >
> > Any guidance or suggestions would be greatly appreciated.
> >
> >
> >
> > Best regards,
> >
> > Xinhai Yue
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > 岳新海
> >
> >
> >
> > 南方科技大学/学生/研究生/2023级研究生
> >
> >
> >
> > 广东省深圳市南山区学苑大道1088号
> >
> >
> >
> >
> >
More information about the petsc-dev
mailing list