Commits

Breadcrumbs

History for

ondigitization

User selector

All users

All time

Commit history

Commits on Nov 17, 2025

• GPU CMake: Add option to override deterministic mode

  davidrohr

  committed
  ea2972eView commit details

  Copy full SHA for ea2972e

  View code at this pointBrowse repository at this point

Commits on Nov 12, 2025

• Find Thrust + fix compilation on CUDA13 (#14801)

  Show description for 72e6d01
  mpuccio

  authored
  72e6d01View commit details

  Copy full SHA for 72e6d01

  View code at this pointBrowse repository at this point

Commits on Oct 24, 2025

• GPU: Fix architecture name, we do not use MI2xx

  davidrohr

  committed
  e0c17bcView commit details

  Copy full SHA for e0c17bc

  View code at this pointBrowse repository at this point

Commits on Sep 19, 2025

• GPU CMake: Bump default CUDA architecture to 75 to support CUDA 13

  davidrohr

  committed
  17da522View commit details

  Copy full SHA for 17da522

  View code at this pointBrowse repository at this point

Commits on Aug 25, 2025

• GPU CMake: Force using alidist GCC for host compilation of CUDA/HIP code

  davidrohr

  committed
  012946bView commit details

  Copy full SHA for 012946b

  View code at this pointBrowse repository at this point

Commits on Jun 16, 2025

• GPU CUDA CMake: Do not print warnings that architectures will be depreacted

  davidrohr

  committed
  fe8111bView commit details

  Copy full SHA for fe8111b

  View code at this pointBrowse repository at this point

• GPU CMake: Switch from deprecated AMDGPU_TARGETS to GPU_TARGETS

  davidrohr

  committed
  d4fb131View commit details

  Copy full SHA for d4fb131

  View code at this pointBrowse repository at this point

Commits on Jun 4, 2025

• Add a version number to FindO2GPU.cmake and remark to keep it in sync with alidist

  davidrohr

  committed
  0c08a1fView commit details

  Copy full SHA for 0c08a1f

  View code at this pointBrowse repository at this point

Commits on May 29, 2025

• GPU CMake: Some adaptions for gpu-systems package

  davidrohr

  committed
  49882f7View commit details

  Copy full SHA for 49882f7

  View code at this pointBrowse repository at this point

Commits on May 27, 2025

• GPU RTC: Check that current compiler versions are the same as those used to create RTC code in cache

  davidrohr

  committed
  a7534dbView commit details

  Copy full SHA for a7534db

  View code at this pointBrowse repository at this point

Commits on Apr 23, 2025

• GPU CMake: Another attempt to silence the CMake warnings about architectures

  davidrohr

  committed
  1f6767cView commit details

  Copy full SHA for 1f6767c

  View code at this pointBrowse repository at this point

Commits on Apr 22, 2025

• GPU CMake: Set GPU architecture before checking the language, to suppress warnings about architecture detection failure

  davidrohr

  committed
  4204d44View commit details

  Copy full SHA for 4204d44

  View code at this pointBrowse repository at this point

Commits on Apr 18, 2025

• GPU CMake: Fix compile flags, particularly HIP deterministic mode was missing -ffp-contract=off

  davidrohr

  committed
  b856a63View commit details

  Copy full SHA for b856a63

  View code at this pointBrowse repository at this point

Commits on Mar 27, 2025

• GPU: Make launch bounds of GPUDefParamters available as C++ struct at runtime

  davidrohr

  committed
  6d54cfcView commit details

  Copy full SHA for 6d54cfc

  View code at this pointBrowse repository at this point

• GPU HIP CMake: Remove deprecated option

  davidrohr

  committed
  d1fcd96View commit details

  Copy full SHA for d1fcd96

  View code at this pointBrowse repository at this point

• GPU CMake: Cleanup and do not use Clang x86 denormals flag for device compilation

  davidrohr

  committed
  d9d78bcView commit details

  Copy full SHA for d9d78bc

  View code at this pointBrowse repository at this point

Commits on Mar 26, 2025

• GPU Compile Flags: Set denormals to zero unconditionally

  Show description for 78d0f9c
  davidrohr

  committed
  78d0f9cView commit details

  Copy full SHA for 78d0f9c

  View code at this pointBrowse repository at this point

Commits on Mar 24, 2025

• GPU: Replace GPUCA_NO_FAST_MATH by more fine-grain GPUCA_DETERMINISTIC_MODE

  davidrohr

  committed
  e836e3fView commit details

  Copy full SHA for e836e3f

  View code at this pointBrowse repository at this point

• GPU CMake: Shuffle stuff a bit and move all NO_FAST_MATH stuff to GPU/... and FindO2GPU.cmake

  davidrohr

  committed
  ee497d7View commit details

  Copy full SHA for ee497d7

  View code at this pointBrowse repository at this point

Commits on Mar 14, 2025

• GPU CMake: Auto-detect ROCm even if CMake prefix path not set

  davidrohr

  committed
  b183c55View commit details

  Copy full SHA for b183c55

  View code at this pointBrowse repository at this point

Commits on Mar 7, 2025

• GPU: Bump required LLVM version for OpenCL

  davidrohr

  committed
  f34afc2View commit details

  Copy full SHA for f34afc2

  View code at this pointBrowse repository at this point

Commits on Jan 17, 2025

• GPU: Rename OpenCL2 to OpenCL, now that we have only 1 OpenCL implementation

  davidrohr

  committed
  843479bView commit details

  Copy full SHA for 843479b

  View code at this pointBrowse repository at this point

• GPU:Remove OpenCL 1.2, now that OpenCL 2 supports all its functionality

  davidrohr

  committed
  90ca34cView commit details

  Copy full SHA for 90ca34c

  View code at this pointBrowse repository at this point

Commits on Sep 11, 2024

• GPU: GPUCA_BUILD_DEBUG makes sure all possible source files are compiled for all kernels

  davidrohr

  committed
  cc64785View commit details

  Copy full SHA for cc64785

  View code at this pointBrowse repository at this point

• GPU CMake: Add check for minimum ROCm version

  davidrohr

  committed
  158e6f8View commit details

  Copy full SHA for 158e6f8

  View code at this pointBrowse repository at this point

Commits on Sep 7, 2024

• GPU: More WORKAROUNDs for CUDA bug exposing device code via host symbols

  davidrohr

  committed
  b1438a9View commit details

  Copy full SHA for b1438a9

  View code at this pointBrowse repository at this point

• GPU CMake: Raise required CUDA version to 12.6

  davidrohr

  committed
  1a27e5dView commit details

  Copy full SHA for 1a27e5d

  View code at this pointBrowse repository at this point

Commits on Aug 23, 2024

• GPU Compilation: print CUDA resource usage only if requested by GPUCA_KERNEL_RESOURCE_USAGE_VERBOSE flag, also print HIP resource usage

  davidrohr

  committed
  3e82457View commit details

  Copy full SHA for 3e82457

  View code at this pointBrowse repository at this point

• GPU CMake: With new ROCm/Cmake, GCC usage seems to work, no need any more to enforce gcc-toolchain

  davidrohr

  committed
  9b6e47aView commit details

  Copy full SHA for 9b6e47a

  View code at this pointBrowse repository at this point

Commits on Aug 20, 2024

• GPU CMake: Remove unsed O2_HIP_CMAKE_LINK_FLAGS and create HIP --offload-arch arguments only when needed

  davidrohr

  committed
  c481ad4View commit details

  Copy full SHA for c481ad4

  View code at this pointBrowse repository at this point

Commits on Aug 9, 2024

• GPU CMake: Set amdgpu-function-calls=true, in case not set by the ROCm scripts / cmake files

  davidrohr

  committed
  d40ccfaView commit details

  Copy full SHA for d40ccfa

  View code at this pointBrowse repository at this point

• GPU Cmake: allow unsupported GCC only if we don't autodetect CUDA

  davidrohr

  committed
  ded3e55View commit details

  Copy full SHA for ded3e55

  View code at this pointBrowse repository at this point

Commits on Aug 8, 2024

• GPU CMake: Make compatible to ROCm >= 6.2

  davidrohr

  committed
  6c44a3fView commit details

  Copy full SHA for 6c44a3f

  View code at this pointBrowse repository at this point

Commits on Aug 5, 2024

• GPU CMake: Workaround for CMake adding old cmake library paths implicitly if we request an old GCC for CUDA

  davidrohr

  committed
  bf35985View commit details

  Copy full SHA for bf35985

  View code at this pointBrowse repository at this point

• GPU CMake: Bugfix: fix hardcoded old llvm version in case of manually specified clang binary

  davidrohr

  committed
  0457082View commit details

  Copy full SHA for 0457082

  View code at this pointBrowse repository at this point