Commits

Breadcrumbs

History for

ondev

User selector

All users

All time

Commit History

Commits on Mar 22, 2024

• GPU CMake: Fix FindO2GPU.cmake

  davidrohr

  committed
  5df07e6

  Copy full SHA for 5df07e6

• GPU HIP CMake: More fixes for ROCm detection, wasn't working on my laptop

  davidrohr

  committed
  c280308

  Copy full SHA for c280308

Commits on Mar 21, 2024

• GPU ROCM CMake: With HIP as language, it seems we may no longer set CMAKE_HIP_COMPILER and HIP_CLANG_PATH

  davidrohr

  committed
  03a5feb

  Copy full SHA for 03a5feb

• GPU HIP CMake: Automatically set -amdgpu-legacy-sgpr-spill-lowering=true workaround for ROCm >= 6.0

  davidrohr

  committed
  444a65a

  Copy full SHA for 444a65a

• GPU CMake: Use global hip_HIPIFY_PERL_EXECUTABLE variable

  davidrohr

  committed
  eb18b0a

  Copy full SHA for eb18b0a

• GPU Cmake: remove obsolete code

  davidrohr

  committed
  db32a53

  Copy full SHA for db32a53

• GPU HIP: -ffast-math seems to work now, and is needed for full performance

  davidrohr

  committed
  f0e9661

  Copy full SHA for f0e9661

• CMake: Release builds should not decrease the -O level below something requested externally via CXXFLAGS

  davidrohr

  committed
  b1add2b

  Copy full SHA for b1add2b

Commits on Mar 20, 2024

• Follow conventions for CMAKE_BUILD_TYPE

  Show description for 2b56f82
  

  ktf

  authored and

  davidrohr

  committed
  2b56f82

  Copy full SHA for 2b56f82

Commits on Mar 18, 2024

• Add GPUCA_NO_FAST_MATH_WHOLEO2 cmake variable to extend NO_FAST_MATH build to whole O2

  davidrohr

  committed
  a81bf56

  Copy full SHA for a81bf56

Commits on Mar 14, 2024

• GPU: Extend CUDA compute target detection

  davidrohr

  committed
  facd33b

  Copy full SHA for facd33b

Commits on Jan 31, 2024

• GPU: Automatically disable math optimizations for TPC track model encoding/decoding code

  davidrohr

  committed
  3860e7f

  Copy full SHA for 3860e7f

• GPU: Need to disable some more stuff with GPUCA_NO_FAST_MATH

  davidrohr

  committed
  f758ea1

  Copy full SHA for f758ea1

Commits on Jan 17, 2024

• Deprecate (T)Pythia6

  Show description for 95c186e
  sawenzel

  committed
  95c186e

  Copy full SHA for 95c186e

Commits on Jan 14, 2024

• Do not try to find GPU dependencies in non-O2 software that pulls in O2 cmake-config

  davidrohr

  committed
  dc6cecd

  Copy full SHA for dc6cecd

• GPU: Suppress bogus CUDA compiler warning

  davidrohr

  committed
  e8309a8

  Copy full SHA for e8309a8

• GPU: Get rid of GPUReconstructionKernels file, and auto-generate based on kernels.cmake

  davidrohr

  committed
  fb33365

  Copy full SHA for fb33365

Commits on Jan 2, 2024

• GPU: Deprecated paths /opt/rocm/hip removed in ROCm6, just use /opt/rocm/bin

  davidrohr

  committed
  b0dc1b0

  Copy full SHA for b0dc1b0

• GPU: If we specify a clang version for OpenCL, use that also for llvm-spirv

  davidrohr

  committed
  7ac484f

  Copy full SHA for 7ac484f

Commits on Dec 10, 2023

• GPU: Force HIP to C++17 compilation

  davidrohr

  committed
  a34fcd7

  Copy full SHA for a34fcd7

Commits on Nov 29, 2023

• cmake: fix FindO2GPU.cmake

  

  aphecetche

  authored and

  davidrohr

  committed
  cdee868

  Copy full SHA for cdee868

Commits on Nov 28, 2023

• GPU: Move GPUReconstructionHIP to hip language (#12255)

  Show description for 6cc46bd
  mconcas

  authored
  6cc46bd

  Copy full SHA for 6cc46bd

Commits on Oct 28, 2023

• Remove optional DDS package in CMake, since it is not used as far as I can tell

  

  davidrohr

  authored and

  ktf

  committed
  1cc195d

  Copy full SHA for 1cc195d

Commits on Oct 27, 2023

• GPU: Add GPUCA_OPENCL_CLANGBIN setting to force OpenCL spirv clang version

  davidrohr

  committed
  06ba41a

  Copy full SHA for 06ba41a

• GPU: Have a set of default compute targets that can be used via 'default' keyword

  davidrohr

  committed
  407f31a

  Copy full SHA for 407f31a

• GPU: Remove non-working AMD clang-ocl OCL compilation, was always leading to internal compiler error anyway

  davidrohr

  committed
  986fd2c

  Copy full SHA for 986fd2c

• Ignore checking for FairMQ version (#12156)

  Show description for d905b8c
  ktf

  authored
  d905b8c

  Copy full SHA for d905b8c

Commits on Oct 20, 2023

• GPU: Fix HIP compilation with modern system LLVM present

  davidrohr

  committed
  846881c

  Copy full SHA for 846881c

Commits on Oct 19, 2023

• Drop obsolete flag

  Show description for 31ff714
  ktf

  committed
  31ff714

  Copy full SHA for 31ff714

Commits on Oct 17, 2023

• GPU: CMake enable_language(HIP) (#11860)

  Show description for 2f62317
  mconcas

  authored
  2f62317

  Copy full SHA for 2f62317

Commits on Sep 7, 2023

• Add support for thread sanitizer (#11828)

  ktf

  authored
  9d31caf

  Copy full SHA for 9d31caf

Commits on Sep 2, 2023

• [rANS] New Histogram Types

  Show description for 5c4aa7e
  

  MichaelLettrich

  authored and

  shahor02

  committed
  5c4aa7e

  Copy full SHA for 5c4aa7e

• [rANS] production preparation for library

  Show description for e30a55c
  

  MichaelLettrich

  authored and

  shahor02

  committed
  e30a55c

  Copy full SHA for e30a55c

• [rANS] Implement SIMD Encoding

  Show description for 17ee27c
  

  MichaelLettrich

  authored and

  shahor02

  committed
  17ee27c

  Copy full SHA for 17ee27c

Commits on Aug 3, 2023

• GPU: Fix disabling of compile warnings

  davidrohr

  committed
  f5eac9d

  Copy full SHA for f5eac9d