v8.2.0
GPUArrays v8.2.0
Closed issues:
- Cartesian indexing issue for high dimensional arrays (#340)
- Too much specialization with permutedims: Use dynamic shared memory (#375)
permutedimswith data type UInt32 (#386)
Merged pull requests:
- new permutedims without unrolling permutation orders (#383) (@GiggleLiu)
- Overload
choleskymethod in order to prevent scalar indexing forDiagonal(#384) (@simsurace) - speed up permutedims with UInt32 (#387) (@GiggleLiu)
- Call parent when unwrapping arrays. (#388) (@maleadt)