Commits

Branch selector

 development

User selector

maximumcats

Datepicker

All time

Commit History

Commits on Jul 9, 2020

• Retain old timestep data during post-timestep regrid (#1122)

  maximumcats

  authored
  c2a57a2

  Copy full SHA for c2a57a2

Commits on Jul 6, 2020

• Document GPUS_PER_SOCKET and GPUS_PER_NODE (#1097)

  Show description for cc1f9d4
  

  maximumcats

  and

  WeiqunZhang

  authored
  cc1f9d4

  Copy full SHA for cc1f9d4

Commits on Jun 27, 2020

• Set backtrace threadprivate exception for PGI (#1067)

  maximumcats

  authored
  ef8ac4a

  Copy full SHA for ef8ac4a

Commits on Jun 26, 2020

• MFIter reducer belongs only to AMREX_USE_GPU_PRAGMA (#1063)

  Show description for 116ff2a
  maximumcats

  authored
  116ff2a

  Copy full SHA for 116ff2a

Commits on Jun 25, 2020

• Only add generic GNU flags to F compiler when appropriate (#1060)

  Show description for 385a0a5
  maximumcats

  authored
  385a0a5

  Copy full SHA for 385a0a5

Commits on Mar 21, 2020

• Use Fab interface in setPhysBoundaryValues if available

  maximumcats

  committed
  e02e0a9

  Copy full SHA for e02e0a9

Commits on Feb 22, 2020

• Deal with the case where extern is in a subdirectory

  maximumcats

  committed
  f63d9af

  Copy full SHA for f63d9af

Commits on Feb 15, 2020

• Add stream benchmark

  maximumcats

  committed
  1803fa5

  Copy full SHA for 1803fa5

Commits on Feb 9, 2020

• Add __managed__ attribute to extern parameters

  maximumcats

  committed
  1e47b1f

  Copy full SHA for 1e47b1f

• Merge branch 'development' into write_probin

  maximumcats

  committed
  45abae9

  Copy full SHA for 45abae9

Commits on Feb 8, 2020

• Add AMREX_GPU_CONSTANT and AMREX_GPU_MANAGED

  maximumcats

  committed
  1e29dfc

  Copy full SHA for 1e29dfc

Commits on Dec 28, 2019

• Use a 1D block/thread layout for most loops

  maximumcats

  committed
  1444804

  Copy full SHA for 1444804

• Select a better block layout for 1D and 2D

  maximumcats

  committed
  133b27a

  Copy full SHA for 133b27a

Commits on Dec 27, 2019

• Add GPU pragma option for a device sync

  maximumcats

  committed
  bf29d2b

  Copy full SHA for bf29d2b

Commits on Dec 25, 2019

• Handle an edge case where the block size is less than a warp

  maximumcats

  committed
  4c77d7a

  Copy full SHA for 4c77d7a

Commits on Dec 17, 2019

• Move some NVCC options to be included for C source files

  maximumcats

  committed
  e37e067

  Copy full SHA for e37e067

• Include C sources in the CUDA preprocessing

  maximumcats

  committed
  c2b60f2

  Copy full SHA for c2b60f2

Commits on Dec 7, 2019

• Specify in suffix if using OpenMP offload

  maximumcats

  committed
  d7e23ba

  Copy full SHA for d7e23ba

Commits on Dec 2, 2019

• Move OMP offload flags into ibm.mak

  maximumcats

  committed
  671c358

  Copy full SHA for 671c358

Commits on Nov 26, 2019

• Fix the ghost cell check

  maximumcats

  committed
  9251c9d

  Copy full SHA for 9251c9d

• Merge branch 'development' into multifab_swap

  maximumcats

  committed
  8678ba3

  Copy full SHA for 8678ba3

• Take a shortcut in MultiFab::Swap() when we can

  maximumcats

  committed
  d074714

  Copy full SHA for d074714

Commits on Nov 24, 2019

• Explicitly specify which Arena StateData uses

  maximumcats

  committed
  73d2b94

  Copy full SHA for 73d2b94

• Add a MultiFab operation that explicitly swaps dst/src data

  Show description for 7c89663
  maximumcats

  committed
  7c89663

  Copy full SHA for 7c89663

Commits on Nov 11, 2019

• Allow code to compile for either old or no-host builds

  maximumcats

  committed
  ca8b5f5

  Copy full SHA for ca8b5f5

• ifdef out CPU boundary fill functions for device-only pragma build

  maximumcats

  committed
  0cc07bb

  Copy full SHA for 0cc07bb

• Define a default value for a new make variable

  maximumcats

  committed
  56bd6ee

  Copy full SHA for 56bd6ee

• Add a define if we're using the no-host version

  maximumcats

  committed
  864ee51

  Copy full SHA for 864ee51

• Add option for GPU pragmas to generate only a device version

  maximumcats

  committed
  e76f21d

  Copy full SHA for e76f21d

• Implement amrex_fab_filcc without amrex_get_loop_bounds

  maximumcats

  committed
  1729979

  Copy full SHA for 1729979

Commits on Nov 9, 2019

• Update the default CUDA architecture to cc70 (Volta V100)

  maximumcats

  committed
  3b7e7bd

  Copy full SHA for 3b7e7bd

Commits on Nov 7, 2019

• Report numerical CUDA error

  maximumcats

  committed
  4e1e658

  Copy full SHA for 4e1e658

Commits on Oct 30, 2019

• Set OpenACC stream to be the same as the CUDA stream

  maximumcats

  committed
  da1008a

  Copy full SHA for da1008a

• For OpenACC where CCOMP is not pgc++, still include openacc.h

  maximumcats

  committed
  e225ebd

  Copy full SHA for e225ebd

Commits on Oct 29, 2019

• Add peak at OLCF

  maximumcats

  committed
  0c19db1

  Copy full SHA for 0c19db1

Pagination

Previous

Next