Skip to content

Latest commit

 

History

History
40 lines (33 loc) · 1.8 KB

CHANGELOG.md

File metadata and controls

40 lines (33 loc) · 1.8 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

[v2.0] - 2024-08-19

Added

  • CI via GitHub Actions
  • Support for decks with arbitrary pose count (was fixed to 65536 before)
  • Kernel runtime summary statistics (min, max, sum, avg, stdDev)
  • Partial auto-tune to find the best PPWI/WGSIZE combination
  • PPWI/WGSIZE heatmap script
  • Flag for machine-readable CSV output
  • Flag for toggling optional energy output to file
  • Executable now embeds compile commands used at build-time
  • New models: C++ std, C++20 std, Intel TBB, RAJA, Thrust, serial
  • Context allocation and transfer (if required) is now measured
  • Added optional cpu_feature library, can be disabled at build-time

Changed

  • Human-readable output now uses YAML format
  • Renamed parameter NUM_TD_PER_THREAD to PPWI for all implementations
  • Consolidated builds to use a shared CMake script, Makefiles removed
  • All implementation now share a common C++ driver with device selection based on index or name substrings
  • More robust input deck/parameter validation (bad params now gives reason at launch instead of SEGV mid-benchmark)
  • Optimised deck IO for faster poses to memory
  • OpenCL implementation now embeds kernel directly in executable
  • OpenCL implementation now use the official OpenCL C++ Bindings
  • Kokkos implementation now supports team policies
  • SYCL implementation now uses discard_write instead of read_write for storing energy results
  • HIP implementation is now standalone and not derived from the CUDA via hipify
  • Fused OpenMP and OpenMP target implementation via macros
  • OpenMP/OpenMP target now supports the team directive as wgsize

Fixed

  • SYCL and Kokkos impl. not respecting runtime pose count flag

[v1.0] - 2021-06-24

Initial public release.