Commits

Breadcrumbs

History for

ondev

User selector

All users

All time

Commit History

Commits on Sep 11, 2021

• Some CMake improvements for GPU detection

  davidrohr

  committed
  f77e8a1

  Copy full SHA for f77e8a1

Commits on Jun 23, 2021

• Updated copyright headers in all files with new text … (#6430)

  Show description for 03608ff
  ihrivnac

  authored
  03608ff

  Copy full SHA for 03608ff

Commits on Jun 7, 2021

• GPU: Enforce the CUDA Architecture we set, otherwise some other CMake module might override it

  davidrohr

  committed
  fbe5289

  Copy full SHA for fbe5289

Commits on May 14, 2021

• GPU: AMD HIP creates a cache variable, overriding our settings if not FORCEd

  davidrohr

  committed
  8acc0c8

  Copy full SHA for 8acc0c8

Commits on Apr 25, 2021

• Fix CUDA version check, apparently the CMake way has changed meanwhile

  davidrohr

  committed
  6395e6b

  Copy full SHA for 6395e6b

Commits on Apr 21, 2021

• GPU: Require clang >= 13 for OpenCL2 compilation

  davidrohr

  committed
  7263003

  Copy full SHA for 7263003

• GPU: Require at least CUDA 11.3

  davidrohr

  committed
  759c329

  Copy full SHA for 759c329

Commits on Mar 4, 2021

• Add missing find_package if OCL2 is requested but OCL1 is not

  davidrohr

  committed
  e651938

  Copy full SHA for e651938

Commits on Jan 27, 2021

• GPU: Apparently one must not do find_package(ROCM)

  davidrohr

  committed
  592ae04

  Copy full SHA for 592ae04

Commits on Jan 9, 2021

• GPU: Disable CMake warnings already earlier

  davidrohr

  committed
  b5dc383

  Copy full SHA for b5dc383

Commits on Jan 7, 2021

• GPU: Disable some ROCm CMake warnings

  davidrohr

  committed
  6bbd168

  Copy full SHA for 6bbd168

Commits on Nov 10, 2020

• Use 61 as minimum CUDA compute capability to support half precision

  davidrohr

  committed
  79250b7

  Copy full SHA for 79250b7

Commits on Aug 27, 2020

• GPU: Use CMake 3.18 style to set CUDA GPU Targets

  davidrohr

  committed
  bca6140

  Copy full SHA for bca6140

Commits on Aug 26, 2020

• GPU: Disable CUDA host compiler check

  davidrohr

  committed
  2f10fe8

  Copy full SHA for 2f10fe8

• GPU: Use C++17 in CUDA compilation with CUDA 11/CMake 3.18

  davidrohr

  committed
  63e6af7

  Copy full SHA for 63e6af7

Commits on Aug 25, 2020

• GPU: Switch to CUDA 11

  davidrohr

  committed
  1dad6ed

  Copy full SHA for 1dad6ed

• GPU: Adjust CMake for hip-clang

  davidrohr

  committed
  24bb8a1

  Copy full SHA for 24bb8a1

• GPU: Enable __device__ keyword for lambdas in CUDA

  davidrohr

  committed
  9b27ec4

  Copy full SHA for 9b27ec4

Commits on Aug 24, 2020

• GPU: Define HIP CXX flags globally and add gpu-rdc to ITS code (no longer default with hip-clang)

  davidrohr

  committed
  0643be8

  Copy full SHA for 0643be8

Commits on Apr 27, 2020

• GPU: Keep CMake CUDA compile options and extend -Werror for CUDA if set

  davidrohr

  committed
  08faadf

  Copy full SHA for 08faadf

Commits on Dec 19, 2019

• GPU: Workaround for CUDA Detection in CMake (CheckLanguage does not respect CMAKE_CUDA_HOST_COMPILER)

  

  davidrohr

  authored and

  ktf

  committed
  9b46d2b

  Copy full SHA for 9b46d2b

Commits on Dec 9, 2019

• GPU: Force flushing denormals to zero

  

  davidrohr

  authored and

  sawenzel

  committed
  f5306e6

  Copy full SHA for f5306e6

Commits on Oct 24, 2019

• GPU: If cub is not found, complain about cub not CUDA

  davidrohr

  committed
  6cee5d0

  Copy full SHA for 6cee5d0

Commits on Oct 16, 2019

• GPU: Do not try to build OpenCL2 backend when OpenCL 2.2 headers are available but Clang < 10.0 since preprocessing will fail

  davidrohr

  committed
  699f84b

  Copy full SHA for 699f84b

Commits on Oct 1, 2019

• GPU: Add setting to force checking all OpenCL2 backends for CI

  davidrohr

  committed
  75b75fe

  Copy full SHA for 75b75fe

Commits on Sep 26, 2019

• Make rocthrust target globally visible before attempting to alias it

  

  aphecetche

  authored and

  sawenzel

  committed
  b3b1726

  Copy full SHA for b3b1726

• fix stupid misplaced alias...

  

  davidrohr

  authored and

  sawenzel

  committed
  b9943ee

  Copy full SHA for b9943ee

• Fix GPU CMake namespaces

  

  davidrohr

  authored and

  sawenzel

  committed
  ef9360e

  Copy full SHA for ef9360e

Commits on Sep 24, 2019

• GPU: Add line information to device code in debug build

  davidrohr

  committed
  5ff19db

  Copy full SHA for 5ff19db

Commits on Sep 4, 2019

• GPU: Support OpenCL SPIR-V IL and build from source at runtime

  davidrohr

  committed
  675052a

  Copy full SHA for 675052a

• GPU: Fix compile warnings

  davidrohr

  committed
  41b2af0

  Copy full SHA for 41b2af0

Commits on Jul 30, 2019

• Fix ROCm/HIP autodetection

  davidrohr

  committed
  46716fb

  Copy full SHA for 46716fb

Commits on Jul 26, 2019

• GPU CMake: Add detection for thrust / hipcub / rocthrust

  davidrohr

  committed
  5472d5a

  Copy full SHA for 5472d5a

• GPU: Add OpenCL 2.x backend using ROCm Clang/LLVM >= 10.0 clang-ocl offline compiler

  davidrohr

  committed
  c6a6627

  Copy full SHA for c6a6627

Commits on Jul 17, 2019

• Forward CMake CXX Flags to CUDA (otherwise we loose warnings / gdb / ...)

  davidrohr

  committed
  12581ae

  Copy full SHA for 12581ae