Commits

Breadcrumbs

History for

ondev

User selector

All users

All time

Commit History

Commits on Feb 11, 2026

• GPU Vulkan Display: fix DEPFILE path

  davidrohr

  committed
  a1b356b

  Copy full SHA for a1b356b

Commits on Feb 10, 2026

• GPU Parameters: Add script to generate parameter file from parameter list csv/json and architecture

  davidrohr

  committed
  2d96089

  Copy full SHA for 2d96089

Commits on Feb 9, 2026

• GPU Parameters: Support multiple csv/json files, and merge the parameters into the header on the fly

  davidrohr

  committed
  f1175e1

  Copy full SHA for f1175e1

• GPU CMake: Clean up some targets, fix if JSON contains 0 architectures

  davidrohr

  committed
  f4becde

  Copy full SHA for f4becde

• GPU CMake: Write separate headers for GPU device and non-device parameters, use same headers for compilation and for parameter file generation

  davidrohr

  committed
  6a9fd1e

  Copy full SHA for 6a9fd1e

Commits on Feb 7, 2026

• GPU CMake: Use FILE GENERATE to generate Default Parameter Headers to track changes and rerun if necessary

  davidrohr

  committed
  fe395fe

  Copy full SHA for fe395fe

• GPU CMake: Make source for GPU parameters configurable, auto-convert CSV to JSON if necessary

  davidrohr

  committed
  0f4e315

  Copy full SHA for 0f4e315

• GPU CMake: Generate optimized parameter files for all available architectures, not only for hardcoded list

  davidrohr

  committed
  e7b5a26

  Copy full SHA for e7b5a26

• GPU CMake: Avoid repetitive JSON parsing

  davidrohr

  committed
  5639312

  Copy full SHA for 5639312

Commits on Feb 5, 2026

• GPU Framework: Add OpenCL support to GPU param header generation + update CFCheckPadBaseline parameters

  

  cima22

  authored and

  davidrohr

  committed
  fe431a1

  Copy full SHA for fe431a1

• GPU Framework: refactor generation of default GPU parameters

  

  cima22

  authored and

  davidrohr

  committed
  f683cac

  Copy full SHA for f683cac

• GPU Framework: remove GPUDefParametersDefaults.h and automatically generate GPU parameters using json file and CMake

  

  cima22

  authored and

  davidrohr

  committed
  bfa44ca

  Copy full SHA for bfa44ca

Commits on May 23, 2025

• GPU CMake: Switch from IMPLICIT_DEPENDS to DEPFILE and -MT -MD -MF, as ninja does not support IMPLICIT_DEPENDS

  davidrohr

  committed
  9babfd4

  Copy full SHA for 9babfd4

Commits on Apr 22, 2025

• GPU Standalone: support build with ONNXRuntime

  davidrohr

  committed
  427e840

  Copy full SHA for 427e840

Commits on Apr 16, 2025

• GPU: Move compile-time constant parameters to new scheme with runtimeParameter struct and automatic RTC-generated defines

  Show description for efbf147
  davidrohr

  committed
  efbf147

  Copy full SHA for efbf147

Commits on Apr 13, 2025

• GPU: Fix typo in variable name, fix comments, fix debug messages

  davidrohr

  committed
  e3b82a8

  Copy full SHA for e3b82a8

Commits on Mar 27, 2025

• GPU: Count kernel number in Cmake, and pass to MACROS, to simplify the preprocessor logic

  davidrohr

  committed
  a5caa27

  Copy full SHA for a5caa27

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

  davidrohr

  committed
  6d54cfc

  Copy full SHA for 6d54cfc

Commits on Mar 26, 2025

• GPU: Propagate list of noFastMathKernels to GPU RTC and apply special compile settings

  davidrohr

  committed
  fa403b4

  Copy full SHA for fa403b4

• GPU CMake: Clean up and collect files / kernels to be compiled in deterministic mode in one place

  davidrohr

  committed
  2ddad04

  Copy full SHA for 2ddad04

Commits on Mar 13, 2025

• GPU: Remove obsolete preprocessor magic to create 2 different kernels for single-slice and multi-slice

  davidrohr

  committed
  1b8388d

  Copy full SHA for 1b8388d

Commits on Mar 10, 2025

• GPU HIP Cmake: Get rid of hipcc, use Clang --cuda-device-only instead

  davidrohr

  committed
  d2bd60f

  Copy full SHA for d2bd60f

Commits on Feb 17, 2025

• GPU: Fix linker warning with GCC 14, no need to make stack executable for including the binary code

  davidrohr

  committed
  43d48b1

  Copy full SHA for 43d48b1

Commits on Feb 1, 2025

• GPU: Remove all AliRoot code

  davidrohr

  committed
  edf695d

  Copy full SHA for edf695d

Commits on Jan 17, 2025

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

  davidrohr

  committed
  90ca34c

  Copy full SHA for 90ca34c

Commits on Sep 11, 2024

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

  davidrohr

  committed
  cc64785

  Copy full SHA for cc64785

Commits on Apr 22, 2024

• GPU: Fix standalone compilation without O2 headers

  davidrohr

  committed
  8387a8c

  Copy full SHA for 8387a8c

Commits on Apr 21, 2024

• GPU: Add HIP per-kernel compilation mode

  davidrohr

  committed
  6a8629c

  Copy full SHA for 6a8629c

Commits on Mar 31, 2024

• GPU: Fixes for running with Run 2 data

  davidrohr

  committed
  fa8dafb

  Copy full SHA for fa8dafb

Commits on Mar 27, 2024

• GPU: HIP CMake: Add generation of per-kernel compilation templates

  davidrohr

  committed
  bf8de6a

  Copy full SHA for bf8de6a

Commits on Mar 18, 2024

• GPU: Remove OpenCL workarounds that are no longer needed

  davidrohr

  committed
  30cd3c4

  Copy full SHA for 30cd3c4

• GPU: Add x_types argument to kernel call macos

  davidrohr

  committed
  5d25f14

  Copy full SHA for 5d25f14

Commits on Feb 29, 2024

• GPU: List explicitly which files are needed for which kernel

  davidrohr

  committed
  a772573

  Copy full SHA for a772573

• GPU: Can use = to significy that class file equals kernel name

  davidrohr

  committed
  baf70bf

  Copy full SHA for baf70bf

• GPU: Auto-generate include-list for kernel headers

  davidrohr

  committed
  3d923e7

  Copy full SHA for 3d923e7