Repository navigation
Release v1.4.0b3
New:
- support for python 3.14
- fill= parameter for ndarray, field, tensor, during allocation
- added qd.block_idx()
- add cpu_per_worker_min_block_dim for cpu range loops
Backwards incompatible changes:
- removed SourceBuilder
- removed Python backend
- removed support for Python 3.10 and 3.11
- dropped support for SPIRV < 1.5
What's Changed
CI
- [CI] Report the Genesis check of a pull request closed before its run as skipped. by @duburcqa in #937
- [CI] Describe how the Genesis CI shares its runners. by @duburcqa in #938
- [CI] Move Genesis test and benchmark results out of Linters and enable shared status publishing by @hughperkins in #1008
- [CI] Link coverage and files changed reports from commit statuses by @hughperkins in https://github.com/Genesis-Embodied
- [CI] Limit manylinux default contents permission to read by @hughperkins in #1058
Misc
- [Misc] Remove benchmarks folder by @hughperkins in #942
- [Misc] Expose portable qd.block_idx() for CPU, CUDA, AMDGPU by @hughperkins in #949
- [Misc] Remove the Python backend by @hughperkins in #1021
- [Misc] Remove optional SourceBuilder external-code API by @hughperkins in #1053
CPU
- [CPU] Clarify runtime task and loop-body names by @hughperkins in #951
- [CPU] Configurable scheduling: revamp naming and add cpu_per_worker_min_block_dim by @hughperkins in #933
AMDGPU
- [AMDGPU] Fix direct HIP tensor access and require HIP 6+ by @hughperkins in #955
AI/quadrants/pull/1034 - [AMDGPU] Cache AMD runtime HSACO bytes across resets by @hughperkins in #1023
- [AMDGPU] Remove OpenCL and empty DAZ bitcode inputs; select ABI 600 by @hughperkins in #1047
- [AMDGPU] Centralize device library selection and ABI configuration in CMake by @hughperkins in #1049
SPIRV
- [SPIRV] Clarify SPIR-V module assembly and separate header storage by @hughperkins in #960
- [SPIRV] Separate operations from IR builder tooling by @hughperkins in #961
- [SPIRV] Extend block_idx to Vulkan and Metal through GLSL helpers by @hughperkins in #952
- [SPIRV] Require SPIR-V 1.3 and Vulkan 1.1 by @hughperkins in #964
- [SPIRV] Use SPIR-V 1.5 for Metal and require a 1.5 minimum by @hughperkins in #965
- [SPIRV] Order ndarray host writes through the graphics command queue by @hughperkins in #1046
Vulkan
- [Vulkan] Clear the adstack sizer overflow word by @hughperkins in #972
- [Vulkan] Disable MoltenVK placement heaps by default on Apple by @hughperkins in #971
Docs
- [Docs] Clarify specialized user-guide terminology in matrix_vector.md by @hughperkins in #1022
- [Docs] Isolate doc previews and delete them after seven days by @hughperkins in #1031
- [Docs] Link docs preview from a commit status by @hughperkins in #1032
Tensor
- [Tensor] Fix layout permutation for list tensor indices by @hughperkins in #989
- [Tensor] Move test_tensor_pr560_fixes.py tests into other test files by @hughperkins in #996
- [Tensor] Support shape-axis layouts for vector and matrix tensors by @hughperkins in #991
Ndarray
- [Ndarray] Rename ndarray internal helpers by @hughperkins in #994
- [Ndarray] Clarify ndarray member-walk cache name by @hughperkins in #1010
Build
- [Build] Use GitHub fork for Eigen submodule by @hughperkins in #995
- [Build] Retry interrupted dependency downloads by @hughperkins in #998
- [Build] Support Python 3.14 and retire 3.10/3.11 by @hughperkins in #990
- [Build] Replace deprecated executable lookup with shutil.which by @hughperkins in #1004
- [Build] Require LLVM 22.1.0 and remove obsolete build settings by @hughperkins in #1052
Lang
- [Lang] Use weak keys for argument disposition cache by @hughperkins in #1005
- [Lang] Add fill= initialization to ndarray and tensor factories by @hughperkins in #1015
Tests
- [Tests] Move test-only helpers out of the Quadrants package by @hughperkins in #1006
- [Tests] Use [()] for scalar indexing throughout code and tests by @hughperkins in #1009
- [Tests] Move test_utils into tests.helpers by @hughperkins in #1007
Field
- [Field] Minor refactor in field(...) function by @hughperkins in #1016
- [Field] Add deferred fill= initialization by @hughperkins in #1012
Graph
- [Graph] CUDA: upload per-launch graph args with async HtoD on the launch stream by @arcusbuilds in #950
Template
- [Template] Fix qd.Template function annotations and migrate template tests by @hughperkins in #1055
GPU
- [GPU] Align allocation streams and replace exhaustion test with mocks by @hughperkins in #1042
New Contributors
- @arcusbuilds made their first contribution in #950
Full Changelog: v1.4.0b3...v1.4.0