Commits

History for

Commits on Jan 27, 2022

GPU: Was not using clang and llvm-spirv from the alidist clang installation
davidrohr
committed
f3bedb4

Commits on Jan 25, 2022

GPU: Add flag requested by AMD to avoid a memory issue

David Rohr
authored and
davidrohr
committed
67c193c
GPU: Add -fgpu-defer-diag option to hipcc
davidrohr
committed
597d5e2

Commits on Sep 11, 2021

Some CMake improvements for GPU detection
davidrohr
committed
f77e8a1

Commits on Jun 23, 2021

Updated copyright headers in all files with new text … (#6430)
ihrivnac
authored
03608ff

Commits on Jun 7, 2021

GPU: Enforce the CUDA Architecture we set, otherwise some other CMake module might override it
davidrohr
committed
fbe5289

Commits on May 14, 2021

GPU: AMD HIP creates a cache variable, overriding our settings if not FORCEd
davidrohr
committed
8acc0c8

Commits on Apr 25, 2021

Fix CUDA version check, apparently the CMake way has changed meanwhile
davidrohr
committed
6395e6b

Commits on Apr 21, 2021

GPU: Require clang >= 13 for OpenCL2 compilation
davidrohr
committed
7263003
GPU: Require at least CUDA 11.3
davidrohr
committed
759c329

Commits on Mar 4, 2021

Add missing find_package if OCL2 is requested but OCL1 is not
davidrohr
committed
e651938

Commits on Jan 27, 2021

GPU: Apparently one must not do find_package(ROCM)
davidrohr
committed
592ae04

Commits on Jan 9, 2021

GPU: Disable CMake warnings already earlier
davidrohr
committed
b5dc383

Commits on Jan 7, 2021

GPU: Disable some ROCm CMake warnings
davidrohr
committed
6bbd168

Commits on Nov 10, 2020

Use 61 as minimum CUDA compute capability to support half precision
davidrohr
committed
79250b7

Commits on Aug 27, 2020

GPU: Use CMake 3.18 style to set CUDA GPU Targets
davidrohr
committed
bca6140

Commits on Aug 26, 2020

GPU: Disable CUDA host compiler check
davidrohr
committed
2f10fe8
GPU: Use C++17 in CUDA compilation with CUDA 11/CMake 3.18
davidrohr
committed
63e6af7

Commits on Aug 25, 2020

GPU: Switch to CUDA 11
davidrohr
committed
1dad6ed
GPU: Adjust CMake for hip-clang
davidrohr
committed
24bb8a1
GPU: Enable __device__ keyword for lambdas in CUDA
davidrohr
committed
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

Commits on Apr 27, 2020

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

Commits on Dec 9, 2019

GPU: Force flushing denormals to zero

davidrohr
authored and
sawenzel
committed
f5306e6

Commits on Oct 24, 2019

GPU: If cub is not found, complain about cub not CUDA
davidrohr
committed
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

Commits on Oct 1, 2019

GPU: Add setting to force checking all OpenCL2 backends for CI
davidrohr
committed
75b75fe

Commits on Sep 26, 2019

Make rocthrust target globally visible before attempting to alias it

aphecetche
authored and
sawenzel
committed
b3b1726
fix stupid misplaced alias...

davidrohr
authored and
sawenzel
committed
b9943ee
Fix GPU CMake namespaces

davidrohr
authored and
sawenzel
committed
ef9360e

Commits on Sep 24, 2019

GPU: Add line information to device code in debug build
davidrohr
committed
5ff19db

Commits on Sep 4, 2019

GPU: Support OpenCL SPIR-V IL and build from source at runtime
davidrohr
committed
675052a
GPU: Fix compile warnings
davidrohr
committed
41b2af0

Commits on Jul 30, 2019

Fix ROCm/HIP autodetection
davidrohr
committed
46716fb