v1.0.0

What's Changed

• Use explicit type for dims in ParallelReduce by @PhilipFackler in #292
• Fix backend bits by @PhilipFackler in #293
• Fix type instability causing crash in 2D parallel_for on AMDGPU by @PhilipFackler in #299
• Add to_device and create_stream by @PhilipFackler in #300
• Versions of array() for allocating uninitialized arrays by @PhilipFackler in #301
• Add Apple GPU CI on ExCL by @williamfgc in #305
• Fix runners in CI by @williamfgc in #307
• Metal backend by @williamfgc in #306
• Use correct arch label in CI by @williamfgc in #308
• Update README by @williamfgc in #310
• Correct dimensions for JACC.shared by @PhilipFackler in #309
• Use explicit type for workspace member to avoid type instability by @PhilipFackler in #313
• Add basic macro syntax by @PhilipFackler in #312
• Refactored ParallelReduce outer constructors into JACC.reducer by @PhilipFackler in #314
• Update AMDGPU perf-test kernel by @luraess in #315
• Custom ranges for parallel_for and parallel_reduce by @PhilipFackler in #316
• Repo GPU CI JuliaORNL to JuliaGPU by @williamfgc in #318
• Fix ReadMe by @williamfgc in #317
• Fix documentation link in badge by @williamfgc in #319
• Update deploydocs org to JuliaGPU by @williamfgc in #321
• Add API documentation by @williamfgc in #320
• Update api_usage.md by @PhilipFackler in #322

New Contributors

• @luraess made their first contribution in #315

Full Changelog: v0.6.0...v1.0.0

v0.6.0

What's Changed

• Skip set_backend if passed the same backend by @PhilipFackler in #251
• Add ubuntu arm runner by @PhilipFackler in #250
• Add project documentation by @williamfgc in #252
• Add docs badge and acknowledgement by @williamfgc in #253
• Add missing synchronize implementation by @PhilipFackler in #256
• Improvements to "threads" backend by @PhilipFackler in #257
• Occupancy with oneAPI by @PhilipFackler in #260
• Fixed incorrect condition in AMDGPU LaunchSpec parallel_for by @PhilipFackler in #263
• Update NVIDIA CI runner by @williamfgc in #269
• Managed/Unmanaged reduce workspace for GPU backends by @PhilipFackler in #270
• Use -1 to signal default shmem_size by @PhilipFackler in #272
• Added to_host function by @PhilipFackler in #273
• Add GTX1080 CI workflow by @williamfgc in #274
• Docs badge by @williamfgc in #275
• Implement JACC.Multi for oneAPI by @PhilipFackler in #279
• Add @inline to default parallel_for by @williamfgc in #282
• N-dimensional versions and API update by @PhilipFackler in #276
• Allow more manipulation of backends by @PhilipFackler in #277
• Add JACC.Async implementations by @PhilipFackler in #278
• Simplify parallel_reduce implementation by @PhilipFackler in #284
• Test cleanup and benchmarks by @PhilipFackler in #246
• Add Op type parameter to ParallelReduce by @PhilipFackler in #287
• Updated version to 0.6.0 by @PhilipFackler in #288
• Fixed compat entries by @PhilipFackler in #289

Full Changelog: v0.5.0...v0.6.0

v0.5.0

What's Changed

• Added JACC.Async for CUDA backend. JACC.Async.copy does not work XD by @pedrovalerolara in #227
• Fix scoping for parallel_for in threads impl by @PhilipFackler in #229
• Fix conditions in 2d reduce kernel by @PhilipFackler in #231
• shared(::AbstractArray) and sync_workgroup() by @PhilipFackler in #202
• JACC.Multi API updates by @PhilipFackler in #228
• Use max shmem device props by @PhilipFackler in #235
• Added API functions for do-style syntax by @PhilipFackler in #241
• Scope synchronize properly for threads by @PhilipFackler in #243
• Added fill function by @PhilipFackler in #244
• Fix AMDGPU version by @williamfgc in #239
• Update README by @williamfgc in #238
• Refactor parallel_reduce for do syntax by @PhilipFackler in #248
• Updated version to 0.5.0 by @PhilipFackler in #249

Full Changelog: v0.4.0...v0.5.0

v0.4.0

What's Changed

• Add LaunchSpec versions of parallel_reduce by @PhilipFackler in #216
• Better occupancy for 2D parallel_for and prevent oversubscription by @PhilipFackler in #224
• Release v0.4.0 by @PhilipFackler in #225

Full Changelog: v0.3.1...v0.4.0

v0.3.1

What's Changed

• Replaced while loops in reduce kernels by @PhilipFackler in #206
• Change shmem_size to use thread count (like AMDGPUExt) by @PhilipFackler in #210
• Release v0.3.1 by @PhilipFackler in #211

Full Changelog: v0.3.0...v0.3.1

v0.3.0

What's Changed

• Bump to Julia 1.11.3 in cousteau by @williamfgc in #197
• Reorganize source code and modules by @PhilipFackler in #193
• Release v0.3.0 by @PhilipFackler in #204
• Make sure comparison is in host memory in shared test case by @PhilipFackler in #208

Full Changelog: v0.2.1...v0.3.0

v0.2.1

What's Changed

• Install backend on set_backend by @PhilipFackler in #195
• Release v0.2.1 by @PhilipFackler in #196

Full Changelog: v0.2.0...v0.2.1

v0.2.0

What's Changed

• Added JACCASYNC API for threads.jl backend. Note that JACC.Async work… by @pedrovalerolara in #161
• Bump Atomix to 1.0.1 by @williamfgc in #165
• Remove rocm by @williamfgc in #166
• Bump AMDGPU CI version by @williamfgc in #168
• Update CI labels by @williamfgc in #169
• Update README info by @williamfgc in #164
• Added @init_backend for user convenience. by @PhilipFackler in #170
• Update parallel_reduce by @PhilipFackler in #173
• Remove Array type and add array function by @PhilipFackler in #177
• Fix bug in oneAPI 1D reduce by @PhilipFackler in #182
• WIP: AMDGPU compute occupancy by @PhilipFackler in #178
• WIP: Better blocks/threads calculations for CUDA backend by @PhilipFackler in #136
• Add parallel_for API with keyword struct by @PhilipFackler in #188
• Use computed occupancy for amdgpu parallel_reduce by @PhilipFackler in #190
• Release v0.2.0 by @PhilipFackler in #191

Full Changelog: v0.1.1...v0.2.0

v0.1.1

What's Changed

• Update to macos-latest for github actions by @PhilipFackler in #152
• Update to macos-latest for github actions by @PhilipFackler in #153
• Switched to TestItemRunner and enabled selecting tests by name or tag by @PhilipFackler in #154
• Updated JACC.BLAS test by @PhilipFackler in #155
• Update oneAPI testing by @PhilipFackler in #151
• Fixed @maybe_threaded to work as intended with precompilation by @PhilipFackler in #158
• Release v0.1.1 by @PhilipFackler in #159

Full Changelog: v0.1.0...v0.1.1

JACC v0.1.0

What's Changed

• Reorder GPU grid indices by @williamfgc in #104
• Reorder AMDGPU gridsize by @williamfgc in #105
• Revert reordering by @williamfgc in #106
• swapped thread dimension by @ygtangg in #107
• Changed thread dimension to: 1-32-32 by @ygtangg in #108
• Reverted GPU thread dimension: 32-32-1 by @ygtangg in #109
• Promoted JACC.shared for CUDA backend. Added test case in tests_cuda by @pedrovalerolara in #111
• Promoted JACC.shared for AMDGPU backend. Test added by @pedrovalerolara in #112
• Promoted JACC.shared OneAPI implementation. Added testing. by @pedrovalerolara in #113
• Blas level1 by @hetmankad in #110
• Added JACC.multi for CUDA. Testing doesn't work, I got a segmentation… by @pedrovalerolara in #116
• Added JACC.multi implementation for the AMDGPU backend, non-included … by @pedrovalerolara in #124
• Update JACCMULTI.jl by @pedrovalerolara in #125
• Update to latest checkout action by @williamfgc in #126
• Added stencil-aware funtions for JACC.multi on CUDA back ends. There … by @pedrovalerolara in #127
• Integrate other PRs for fixing extensions by @PhilipFackler in #123
• Bump Julia to 1.10 on AMD GPU CI on cousteau by @williamfgc in #128
• Added ghost cells support for JACC.multi on AMDGPU backend. There is … by @pedrovalerolara in #132
• Added range checks to parallel_for implementations by @PhilipFackler in #131
• Moved and refactored most tests as common portable versions by @PhilipFackler in #135
• Custom operators for parallel_reduce by @PhilipFackler in #120
• Capitalized modules Multi and Experimental and applied formatting by @PhilipFackler in #146
• Release v0.1.0 by @PhilipFackler in #147
• Fixed [compat] entry for julia by @PhilipFackler in #148

New Contributors

• @ygtangg made their first contribution in #107
• @hetmankad made their first contribution in #110

Full Changelog: v0.0.5...v0.1.0