[petsc-dev] GPU performance of MatSOR()

Jed Brown jed at jedbrown.org
Sun Jul 31 16:31:51 CDT 2022


Do you have a test that compares this with a polynomial smoother for the original problem (like Chebyshev for SPD)?

Paul Mullowney <paulmullowney at gmail.com> writes:

> One could also approximate the SOR triangular solves with a Neumann series,
> where each term in the series is a SpMV (great for GPUs). The number of
> terms needed in the series is matrix dependent.
> We've seen this work to great effect for some problems.
>
> -Paul
>
> On Wed, Jul 27, 2022 at 8:05 PM Barry Smith <bsmith at petsc.dev> wrote:
>
>>
>>   There are multicolor versions of SOR that theoretically offer good
>> parallelism on GPUs but at the cost of multiple phases and slower
>> convergence rates. Unless someone already has one coded for CUDA or Kokkos
>> it would take a good amount of code to produce one that offers (but does
>> not necessarily guarantee) reasonable performance on GPUs.
>>
>> > On Jul 27, 2022, at 7:57 PM, Jed Brown <jed at jedbrown.org> wrote:
>> >
>> > Unfortunately, MatSOR is a really bad operation for GPUs. We can make it
>> use sparse triangular primitives from cuSPARSE, but those run on GPU at
>> about 20x slower than MatMult with the same sparse matrix. So unless MatSOR
>> reduces iteration count by 20x compared to your next-best preconditioning
>> option, you'll be better off finding a different preconditioner. This might
>> be some elements of multigrid or polynomial smoothing with point-block
>> Jacobi. If you can explain a bit about your application, we may be able to
>> offer some advice.
>> >
>> > Han Tran <hantran at cs.utah.edu> writes:
>> >
>> >> Hello,
>> >>
>> >> Running my example using VECMPICUDA for VecSetType(), and MATMPIAIJCUSP
>> for MatSetType(), I have the profiling results as shown below. It is seen
>> that MatSOR() has %F of GPU, only has GpuToCpu count and size. Is it
>> correct that PETSc currently does not have MatSOR implemented on GPU? It
>> would be appreciated if you can provide an explanation on how MatSOR()
>> currently use GPU. From this example, MatSOR takes a considerable time
>> relatively compared to other functions.
>> >>
>> >> Thank you.
>> >>
>> >> -Han
>> >>
>> >>
>> ------------------------------------------------------------------------------------------------------------------------
>> >> Event                Count      Time (sec)     Flop
>>           --- Global ---  --- Stage ----  Total   GPU    - CpuToGpu -   -
>> GpuToCpu - GPU
>> >>                   Max Ratio  Max     Ratio   Max  Ratio  Mess   AvgLen
>> Reduct  %T %F %M %L %R  %T %F %M %L %R Mflop/s Mflop/s Count   Size
>>  Count   Size  %F
>> >>
>> ---------------------------------------------------------------------------------------------------------------------------------------------------------------
>> >>
>> >> --- Event Stage 0: Main Stage
>> >>
>> >> BuildTwoSided     220001 1.0 3.9580e+02139.9 0.00e+00 0.0 2.0e+00
>> 4.0e+00 2.2e+05  4  0  0  0 20   4  0  0  0 20     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> BuildTwoSidedF    220000 1.0 3.9614e+02126.4 0.00e+00 0.0 0.0e+00
>> 0.0e+00 2.2e+05  4  0  0  0 20   4  0  0  0 20     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> VecMDot           386001 1.0 6.3426e+01 1.5 1.05e+11 1.0 0.0e+00
>> 0.0e+00 3.9e+05  1 11  0  0 35   1 11  0  0 35  3311   26012   386001
>> 1.71e+05    0 0.00e+00 100
>> >> VecNorm           496001 1.0 5.0877e+01 1.2 5.49e+10 1.0 0.0e+00
>> 0.0e+00 5.0e+05  1  6  0  0 45   1  6  0  0 45  2159    3707   110000
>> 4.87e+04    0 0.00e+00 100
>> >> VecScale          496001 1.0 7.9951e+00 1.0 2.75e+10 1.0 0.0e+00
>> 0.0e+00 0.0e+00  0  3  0  0  0   0  3  0  0  0  6869   13321      0
>> 0.00e+00    0 0.00e+00 100
>> >> VecCopy           110000 1.0 1.9323e+00 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> VecSet            330017 1.0 5.4319e+00 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> VecAXPY           110000 1.0 1.5820e+00 1.0 1.22e+10 1.0 0.0e+00
>> 0.0e+00 0.0e+00  0  1  0  0  0   0  1  0  0  0 15399   35566      0
>> 0.00e+00    0 0.00e+00 100
>> >> VecMAXPY          496001 1.0 1.1505e+01 1.0 1.48e+11 1.0 0.0e+00
>> 0.0e+00 0.0e+00  0 16  0  0  0   0 16  0  0  0 25665   39638      0
>> 0.00e+00    0 0.00e+00 100
>> >> VecAssemblyBegin  110000 1.0 1.2021e+00 1.2 0.00e+00 0.0 0.0e+00
>> 0.0e+00 1.1e+05  0  0  0  0 10   0  0  0  0 10     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> VecAssemblyEnd    110000 1.0 1.5988e-01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> VecScatterBegin   496001 1.0 1.3002e+01 1.0 0.00e+00 0.0 9.9e+05
>> 1.3e+04 1.0e+00  0  0100100  0   0  0100100  0     0       0   110000
>> 4.87e+04    0 0.00e+00  0
>> >> VecScatterEnd     496001 1.0 1.8988e+01 1.3 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> VecNormalize      496001 1.0 5.8797e+01 1.1 8.24e+10 1.0 0.0e+00
>> 0.0e+00 5.0e+05  1  9  0  0 45   1  9  0  0 45  2802    4881   110000
>> 4.87e+04    0 0.00e+00 100
>> >> VecCUDACopyTo     716001 1.0 3.4483e+01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  1  0  0  0  0   1  0  0  0  0     0       0   716001
>> 3.17e+05    0 0.00e+00  0
>> >> VecCUDACopyFrom  1211994 1.0 5.1752e+01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  1  0  0  0  0   1  0  0  0  0     0       0      0
>> 0.00e+00 1211994 5.37e+05  0
>> >> MatMult           386001 1.0 4.8436e+01 1.0 1.90e+11 1.0 7.7e+05
>> 1.3e+04 0.0e+00  1 21 78 78  0   1 21 78 78  0  7862   16962      0
>> 0.00e+00    0 0.00e+00 100
>> >> MatMultAdd        110000 1.0 6.2666e+01 1.1 6.03e+10 1.0 2.2e+05
>> 1.3e+04 1.0e+00  1  7 22 22  0   1  7 22 22  0  1926   16893   440000
>> 3.39e+05    0 0.00e+00 100
>> >> MatSOR            496001 1.0 5.1821e+02 1.1 2.83e+11 1.0 0.0e+00
>> 0.0e+00 0.0e+00 10 31  0  0  0  10 31  0  0  0  1090       0      0
>> 0.00e+00 991994 4.39e+05  0
>> >> MatAssemblyBegin  110000 1.0 3.9732e+02109.2 0.00e+00 0.0 0.0e+00
>> 0.0e+00 1.1e+05  4  0  0  0 10   4  0  0  0 10     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> MatAssemblyEnd    110000 1.0 5.3015e-01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 4.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> MatZeroEntries    110000 1.0 1.3179e+01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> MatCUSPARSCopyTo  220000 1.0 3.2805e+01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  1  0  0  0  0   1  0  0  0  0     0       0   220000
>> 2.41e+05    0 0.00e+00  0
>> >> KSPSetUp          110000 1.0 3.5344e-02 1.3 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> KSPSolve          110000 1.0 6.8304e+02 1.0 8.20e+11 1.0 7.7e+05
>> 1.3e+04 8.8e+05 13 89 78 78 80  13 89 78 78 80  2401   14311   496001
>> 2.20e+05 991994 4.39e+05 66
>> >> KSPGMRESOrthog    386001 1.0 7.2820e+01 1.4 2.10e+11 1.0 0.0e+00
>> 0.0e+00 3.9e+05  1 23  0  0 35   1 23  0  0 35  5765   30176   386001
>> 1.71e+05    0 0.00e+00 100
>> >> PCSetUp           110000 1.0 1.8825e-02 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> PCApply           496001 1.0 5.1857e+02 1.1 2.83e+11 1.0 0.0e+00
>> 0.0e+00 0.0e+00 10 31  0  0  0  10 31  0  0  0  1090       0      0
>> 0.00e+00 991994 4.39e+05  0
>> >> SFSetGraph             1 1.0 2.0936e-05 1.1 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> SFSetUp                1 1.0 2.5347e-03 1.0 0.00e+00 0.0 4.0e+00
>> 3.3e+03 1.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> SFPack            496001 1.0 3.0026e+00 1.1 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >> SFUnpack          496001 1.0 1.1296e-01 1.0 0.00e+00 0.0 0.0e+00
>> 0.0e+00 0.0e+00  0  0  0  0  0   0  0  0  0  0     0       0      0
>> 0.00e+00    0 0.00e+00  0
>> >>
>> ---------------------------------------------------------------------------------------------------------------------------------------------------------------
>>
>>


More information about the petsc-dev mailing list