home.social

#petsc — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #petsc, aggregated by home.social.

  1. 🚀 #PETSc 3.21 was released today. There were a number of new contributors this release; thank you all.
    lists.mcs.anl.gov/pipermail/pe

    A few highlights:
    * VecMDot and (optionally) VecMAXPY can identify strided memory and use gemv when applicable. This is faster than home-rolled kernels on some GPUs.
    * GAMG: new filtering and smoothing options for algebraic multigrid.
    * Small subdomain (many per process) BDDC support
    * Trust region and quasi-Newton trust region improvements.

    petsc.org/release/changes/321/

  2. 🚀 #PETSc 3.21 was released today. There were a number of new contributors this release; thank you all.
    lists.mcs.anl.gov/pipermail/pe

    A few highlights:
    * VecMDot and (optionally) VecMAXPY can identify strided memory and use gemv when applicable. This is faster than home-rolled kernels on some GPUs.
    * GAMG: new filtering and smoothing options for algebraic multigrid.
    * Small subdomain (many per process) BDDC support
    * Trust region and quasi-Newton trust region improvements.

    petsc.org/release/changes/321/

  3. 🚀 3.21 was released today. There were a number of new contributors this release; thank you all.
    lists.mcs.anl.gov/pipermail/pe

    A few highlights:
    * VecMDot and (optionally) VecMAXPY can identify strided memory and use gemv when applicable. This is faster than home-rolled kernels on some GPUs.
    * GAMG: new filtering and smoothing options for algebraic multigrid.
    * Small subdomain (many per process) BDDC support
    * Trust region and quasi-Newton trust region improvements.

    petsc.org/release/changes/321/

  4. 🚀 #PETSc 3.21 was released today. There were a number of new contributors this release; thank you all.
    lists.mcs.anl.gov/pipermail/pe

    A few highlights:
    * VecMDot and (optionally) VecMAXPY can identify strided memory and use gemv when applicable. This is faster than home-rolled kernels on some GPUs.
    * GAMG: new filtering and smoothing options for algebraic multigrid.
    * Small subdomain (many per process) BDDC support
    * Trust region and quasi-Newton trust region improvements.

    petsc.org/release/changes/321/

  5. 🚀 #PETSc 3.21 was released today. There were a number of new contributors this release; thank you all.
    lists.mcs.anl.gov/pipermail/pe

    A few highlights:
    * VecMDot and (optionally) VecMAXPY can identify strided memory and use gemv when applicable. This is faster than home-rolled kernels on some GPUs.
    * GAMG: new filtering and smoothing options for algebraic multigrid.
    * Small subdomain (many per process) BDDC support
    * Trust region and quasi-Newton trust region improvements.

    petsc.org/release/changes/321/

  6. @gvwilson I mention numerical/scientific software because that's the area I work in, but also, our software lifecycle is long (many packages are over 30 years old) with a cultural appetite for excusing poor user and developer experience (cf. "Firetran" jedbrown.org/files/BrownKneple).

    Ex: For linear algebra, #PETSc represented a radical shift from the BLAS/LAPACK philosophy at the time. But PETSc has its share of baked-in architecture, as do mature packages throughout the ecosystem.

  7. @gvwilson I mention numerical/scientific software because that's the area I work in, but also, our software lifecycle is long (many packages are over 30 years old) with a cultural appetite for excusing poor user and developer experience (cf. "Firetran" jedbrown.org/files/BrownKneple).

    Ex: For linear algebra, #PETSc represented a radical shift from the BLAS/LAPACK philosophy at the time. But PETSc has its share of baked-in architecture, as do mature packages throughout the ecosystem.

  8. @gvwilson I mention numerical/scientific software because that's the area I work in, but also, our software lifecycle is long (many packages are over 30 years old) with a cultural appetite for excusing poor user and developer experience (cf. "Firetran" jedbrown.org/files/BrownKneple).

    Ex: For linear algebra, represented a radical shift from the BLAS/LAPACK philosophy at the time. But PETSc has its share of baked-in architecture, as do mature packages throughout the ecosystem.

  9. @gvwilson I mention numerical/scientific software because that's the area I work in, but also, our software lifecycle is long (many packages are over 30 years old) with a cultural appetite for excusing poor user and developer experience (cf. "Firetran" jedbrown.org/files/BrownKneple).

    Ex: For linear algebra, #PETSc represented a radical shift from the BLAS/LAPACK philosophy at the time. But PETSc has its share of baked-in architecture, as do mature packages throughout the ecosystem.

  10. @gvwilson I mention numerical/scientific software because that's the area I work in, but also, our software lifecycle is long (many packages are over 30 years old) with a cultural appetite for excusing poor user and developer experience (cf. "Firetran" jedbrown.org/files/BrownKneple).

    Ex: For linear algebra, #PETSc represented a radical shift from the BLAS/LAPACK philosophy at the time. But PETSc has its share of baked-in architecture, as do mature packages throughout the ecosystem.

  11. Some good news!

    After a little back and forth with intel, managed to get #PETSc built properly with the Intel #LLVM compilers AND #MPI

    in the specific case of PETSc, you want to use `./configure --with-debugging=0 --with-cc='mpiicc -cc=icx' --with-cxx='mpiicpc -cxx=icpx' --with-fc='mpiifort -fc=ifx'`

    NB however that this doesn't work well (at all?) with nested cmake, which is relied on by PETSc when using --download-kokkos=1 to mirror your build flags when building kokkos from source
    #HPC

  12. Some good news!

    After a little back and forth with intel, managed to get #PETSc built properly with the Intel #LLVM compilers AND #MPI

    in the specific case of PETSc, you want to use `./configure --with-debugging=0 --with-cc='mpiicc -cc=icx' --with-cxx='mpiicpc -cxx=icpx' --with-fc='mpiifort -fc=ifx'`

    NB however that this doesn't work well (at all?) with nested cmake, which is relied on by PETSc when using --download-kokkos=1 to mirror your build flags when building kokkos from source
    #HPC

  13. Some good news!

    After a little back and forth with intel, managed to get #PETSc built properly with the Intel #LLVM compilers AND #MPI

    in the specific case of PETSc, you want to use `./configure --with-debugging=0 --with-cc='mpiicc -cc=icx' --with-cxx='mpiicpc -cxx=icpx' --with-fc='mpiifort -fc=ifx'`

    NB however that this doesn't work well (at all?) with nested cmake, which is relied on by PETSc when using --download-kokkos=1 to mirror your build flags when building kokkos from source
    #HPC

  14. Some good news!

    After a little back and forth with intel, managed to get #PETSc built properly with the Intel #LLVM compilers AND #MPI

    in the specific case of PETSc, you want to use `./configure --with-debugging=0 --with-cc='mpiicc -cc=icx' --with-cxx='mpiicpc -cxx=icpx' --with-fc='mpiifort -fc=ifx'`

    NB however that this doesn't work well (at all?) with nested cmake, which is relied on by PETSc when using --download-kokkos=1 to mirror your build flags when building kokkos from source
    #HPC

  15. Some good news!

    After a little back and forth with intel, managed to get #PETSc built properly with the Intel #LLVM compilers AND #MPI

    in the specific case of PETSc, you want to use `./configure --with-debugging=0 --with-cc='mpiicc -cc=icx' --with-cxx='mpiicpc -cxx=icpx' --with-fc='mpiifort -fc=ifx'`

    NB however that this doesn't work well (at all?) with nested cmake, which is relied on by PETSc when using --download-kokkos=1 to mirror your build flags when building kokkos from source
    #HPC

  16. The feature I miss the most from Twitter on Mastodon is quote tweets, explicitly of my own for the use case of related but divergent technical threads.

    Working on the #GPGPU && #AVX512 accelerated build of #PETSc to enable faster execution of #OpenFOAM as of now and all the relevant threads would have to be independent threads and hard to cross reference as new threads/tangents pop up.

  17. The feature I miss the most from Twitter on Mastodon is quote tweets, explicitly of my own for the use case of related but divergent technical threads.

    Working on the #GPGPU && #AVX512 accelerated build of #PETSc to enable faster execution of #OpenFOAM as of now and all the relevant threads would have to be independent threads and hard to cross reference as new threads/tangents pop up.

  18. The feature I miss the most from Twitter on Mastodon is quote tweets, explicitly of my own for the use case of related but divergent technical threads.

    Working on the #GPGPU && #AVX512 accelerated build of #PETSc to enable faster execution of #OpenFOAM as of now and all the relevant threads would have to be independent threads and hard to cross reference as new threads/tangents pop up.

  19. The feature I miss the most from Twitter on Mastodon is quote tweets, explicitly of my own for the use case of related but divergent technical threads.

    Working on the #GPGPU && #AVX512 accelerated build of #PETSc to enable faster execution of #OpenFOAM as of now and all the relevant threads would have to be independent threads and hard to cross reference as new threads/tangents pop up.

  20. The feature I miss the most from Twitter on Mastodon is quote tweets, explicitly of my own for the use case of related but divergent technical threads.

    Working on the #GPGPU && #AVX512 accelerated build of #PETSc to enable faster execution of #OpenFOAM as of now and all the relevant threads would have to be independent threads and hard to cross reference as new threads/tangents pop up.

  21. I'll be speaking during the OneAPI dev summit tomorrow, specifically the panel discussion on accelerated computing.

    Partially as a meme/Sanity check the challenge of this afternoon:

    Can I got from no installs to a GPU accelerated simulation of OpenFOAM using an Arc a770 on Linux?

    Plan is #OpenFoam 2212, #petsc 3.19, #OneAPI 2023.1 and #mesa 23 (in case I need to fall back to #OpenCL)

  22. I'll be speaking during the OneAPI dev summit tomorrow, specifically the panel discussion on accelerated computing.

    Partially as a meme/Sanity check the challenge of this afternoon:

    Can I got from no installs to a GPU accelerated simulation of OpenFOAM using an Arc a770 on Linux?

    Plan is #OpenFoam 2212, #petsc 3.19, #OneAPI 2023.1 and #mesa 23 (in case I need to fall back to #OpenCL)

  23. I'll be speaking during the OneAPI dev summit tomorrow, specifically the panel discussion on accelerated computing.

    Partially as a meme/Sanity check the challenge of this afternoon:

    Can I got from no installs to a GPU accelerated simulation of OpenFOAM using an Arc a770 on Linux?

    Plan is #OpenFoam 2212, #petsc 3.19, #OneAPI 2023.1 and #mesa 23 (in case I need to fall back to #OpenCL)

  24. I'll be speaking during the OneAPI dev summit tomorrow, specifically the panel discussion on accelerated computing.

    Partially as a meme/Sanity check the challenge of this afternoon:

    Can I got from no installs to a GPU accelerated simulation of OpenFOAM using an Arc a770 on Linux?

    Plan is #OpenFoam 2212, #petsc 3.19, #OneAPI 2023.1 and #mesa 23 (in case I need to fall back to #OpenCL)

  25. I'll be speaking during the OneAPI dev summit tomorrow, specifically the panel discussion on accelerated computing.

    Partially as a meme/Sanity check the challenge of this afternoon:

    Can I got from no installs to a GPU accelerated simulation of OpenFOAM using an Arc a770 on Linux?

    Plan is #OpenFoam 2212, #petsc 3.19, #OneAPI 2023.1 and #mesa 23 (in case I need to fall back to #OpenCL)

  26. Now is a good time to register for #PETSc 2023 in Chicago, June 5-7.
    petsc.org/release/community/me

    We'd love to hear what you're doing with PETSc. Submit an abstract before May 1. docs.google.com/forms/d/e/1FAI

  27. Now is a good time to register for #PETSc 2023 in Chicago, June 5-7.
    petsc.org/release/community/me

    We'd love to hear what you're doing with PETSc. Submit an abstract before May 1. docs.google.com/forms/d/e/1FAI

  28. Now is a good time to register for 2023 in Chicago, June 5-7.
    petsc.org/release/community/me

    We'd love to hear what you're doing with PETSc. Submit an abstract before May 1. docs.google.com/forms/d/e/1FAI

  29. Now is a good time to register for #PETSc 2023 in Chicago, June 5-7.
    petsc.org/release/community/me

    We'd love to hear what you're doing with PETSc. Submit an abstract before May 1. docs.google.com/forms/d/e/1FAI

  30. Now is a good time to register for #PETSc 2023 in Chicago, June 5-7.
    petsc.org/release/community/me

    We'd love to hear what you're doing with PETSc. Submit an abstract before May 1. docs.google.com/forms/d/e/1FAI

  31. #PETSc 3.19 has been released. Some highlights:

    * Sparse matrices have good perf w/out user-provided preallocation.
    * Optimal simplex quadrature up to 20th.
    * -dm_plex_shape zbox creates a "born parallel" mesh with Z-order partition.
    * "Isoperiodicity": late mapping enables simple traversal of manifolds adjacent to periodic connections.
    * Scalable CGNS output: high order elements and flexible batching of time series.
    * Perf/GPU improvements.
    * Coordinate-based SF graph.

    petsc.org/release/changes/319/

  32. #PETSc 3.19 has been released. Some highlights:

    * Sparse matrices have good perf w/out user-provided preallocation.
    * Optimal simplex quadrature up to 20th.
    * -dm_plex_shape zbox creates a "born parallel" mesh with Z-order partition.
    * "Isoperiodicity": late mapping enables simple traversal of manifolds adjacent to periodic connections.
    * Scalable CGNS output: high order elements and flexible batching of time series.
    * Perf/GPU improvements.
    * Coordinate-based SF graph.

    petsc.org/release/changes/319/

  33. 3.19 has been released. Some highlights:

    * Sparse matrices have good perf w/out user-provided preallocation.
    * Optimal simplex quadrature up to 20th.
    * -dm_plex_shape zbox creates a "born parallel" mesh with Z-order partition.
    * "Isoperiodicity": late mapping enables simple traversal of manifolds adjacent to periodic connections.
    * Scalable CGNS output: high order elements and flexible batching of time series.
    * Perf/GPU improvements.
    * Coordinate-based SF graph.

    petsc.org/release/changes/319/

  34. #PETSc 3.19 has been released. Some highlights:

    * Sparse matrices have good perf w/out user-provided preallocation.
    * Optimal simplex quadrature up to 20th.
    * -dm_plex_shape zbox creates a "born parallel" mesh with Z-order partition.
    * "Isoperiodicity": late mapping enables simple traversal of manifolds adjacent to periodic connections.
    * Scalable CGNS output: high order elements and flexible batching of time series.
    * Perf/GPU improvements.
    * Coordinate-based SF graph.

    petsc.org/release/changes/319/

  35. #PETSc 3.19 has been released. Some highlights:

    * Sparse matrices have good perf w/out user-provided preallocation.
    * Optimal simplex quadrature up to 20th.
    * -dm_plex_shape zbox creates a "born parallel" mesh with Z-order partition.
    * "Isoperiodicity": late mapping enables simple traversal of manifolds adjacent to periodic connections.
    * Scalable CGNS output: high order elements and flexible batching of time series.
    * Perf/GPU improvements.
    * Coordinate-based SF graph.

    petsc.org/release/changes/319/

  36. @jannem You got me there, as I’m not that acquainted with these topics. So I didn’t remember the exact suggestions given.

    Citing from @mhagdorn’s summary (hash signs by me):
    “Essentially use #OpenMP or #SYCL or a suitable library such as #PETSc or #pytorch depending on the application.”

    The slides and videos are supposed to get uploaded very soon.

  37. @jannem You got me there, as I’m not that acquainted with these topics. So I didn’t remember the exact suggestions given.

    Citing from @mhagdorn’s summary (hash signs by me):
    “Essentially use #OpenMP or #SYCL or a suitable library such as #PETSc or #pytorch depending on the application.”

    The slides and videos are supposed to get uploaded very soon.

  38. @jannem You got me there, as I’m not that acquainted with these topics. So I didn’t remember the exact suggestions given.

    Citing from @mhagdorn’s summary (hash signs by me):
    “Essentially use #OpenMP or #SYCL or a suitable library such as #PETSc or #pytorch depending on the application.”

    The slides and videos are supposed to get uploaded very soon.

  39. @jannem You got me there, as I’m not that acquainted with these topics. So I didn’t remember the exact suggestions given.

    Citing from @mhagdorn’s summary (hash signs by me):
    “Essentially use #OpenMP or #SYCL or a suitable library such as #PETSc or #pytorch depending on the application.”

    The slides and videos are supposed to get uploaded very soon.

  40. When does a fluids mini-app cease to be "mini"? What separates Colorful Fluid Dynamics (CFD; h/t @kpdooty) from a tool for science?

    I'd argue it's when you collect turbulent statistics and get careful about statistical convergence. Our fluids mini-app was created to drive #libCEED and #PETSc library integration, but has now taken this step as it transforms into an instrument to explain how RANS models fail (and what they need to succeed). With a 10x strong scaling boost so far.

    #HPC

  41. When does a fluids mini-app cease to be "mini"? What separates Colorful Fluid Dynamics (CFD; h/t @kpdooty) from a tool for science?

    I'd argue it's when you collect turbulent statistics and get careful about statistical convergence. Our fluids mini-app was created to drive #libCEED and #PETSc library integration, but has now taken this step as it transforms into an instrument to explain how RANS models fail (and what they need to succeed). With a 10x strong scaling boost so far.

    #HPC

  42. When does a fluids mini-app cease to be "mini"? What separates Colorful Fluid Dynamics (CFD; h/t @kpdooty) from a tool for science?

    I'd argue it's when you collect turbulent statistics and get careful about statistical convergence. Our fluids mini-app was created to drive and library integration, but has now taken this step as it transforms into an instrument to explain how RANS models fail (and what they need to succeed). With a 10x strong scaling boost so far.

  43. When does a fluids mini-app cease to be "mini"? What separates Colorful Fluid Dynamics (CFD; h/t @kpdooty) from a tool for science?

    I'd argue it's when you collect turbulent statistics and get careful about statistical convergence. Our fluids mini-app was created to drive #libCEED and #PETSc library integration, but has now taken this step as it transforms into an instrument to explain how RANS models fail (and what they need to succeed). With a 10x strong scaling boost so far.

    #HPC

  44. When does a fluids mini-app cease to be "mini"? What separates Colorful Fluid Dynamics (CFD; h/t @kpdooty) from a tool for science?

    I'd argue it's when you collect turbulent statistics and get careful about statistical convergence. Our fluids mini-app was created to drive #libCEED and #PETSc library integration, but has now taken this step as it transforms into an instrument to explain how RANS models fail (and what they need to succeed). With a 10x strong scaling boost so far.

    #HPC

  45. I have a Trilinos lib build. How can I use it to interface with Trilinos? Anyone with a cmake file?

    PetsC should work maybe

    #cmake #PETSc #Trilinos

  46. I have a Trilinos lib build. How can I use it to interface with Trilinos? Anyone with a cmake file?

    PetsC should work maybe

    #cmake #PETSc #Trilinos

  47. I have a Trilinos lib build. How can I use it to interface with Trilinos? Anyone with a cmake file?

    PetsC should work maybe

    #cmake #PETSc #Trilinos

  48. 📢LAST VIDEO OF THE YEAR 😇

    Choosing a proper solver can be quite important for #FiniteElement models. In this video, we discuss a bit about various available solvers in #FreeFEM, including #PETSc with which we have access to tremendous number of tools, solvers, and preconditioners to solve the derived linear system of equations more efficiently.

    youtube.com/watch?v=wGYez2-3rv

    #Modeling #Simulation #FEA #OpenSource

  49. 📢LAST VIDEO OF THE YEAR 😇

    Choosing a proper solver can be quite important for models. In this video, we discuss a bit about various available solvers in , including with which we have access to tremendous number of tools, solvers, and preconditioners to solve the derived linear system of equations more efficiently.

    youtube.com/watch?v=wGYez2-3rv

  50. 📢LAST VIDEO OF THE YEAR 😇

    Choosing a proper solver can be quite important for #FiniteElement models. In this video, we discuss a bit about various available solvers in #FreeFEM, including #PETSc with which we have access to tremendous number of tools, solvers, and preconditioners to solve the derived linear system of equations more efficiently.

    youtube.com/watch?v=wGYez2-3rv

    #Modeling #Simulation #FEA #OpenSource