v8.3.2

GPUArrays v8.3.2

Diff since v8.3.1

Closed issues:

• mapreduce broken for empty arrays (#398)

Merged pull requests:

• Improve tests, typing, and overflow behavior of norm (#399) (@danielwe)
• Support and test empty reductions. (#401) (@maleadt)
• Fix GPU compatibility of reductions in norm (#402) (@danielwe)

v8.3.1

GPUArrays v8.3.1

Diff since v8.3.0

Merged pull requests:

• Fix ldiv for 1.8. (#397) (@maleadt)

v8.3.0

GPUArrays v8.3.0

Diff since v8.2.1

Merged pull requests:

• Overload \ and ldiv! in order to prevent scalar indexing for Diagonal (#391) (@simsurace)
• Add extrema support (#392) (@N5N3)
• Use style dispatch in broadcast(!) (#393) (@N5N3)
• add methods for axpby, axpy, scal (#394) (@bjarthur)
• More checks for supported eltypes in tests. (#395) (@maleadt)

v8.2.1

GPUArrays v8.2.1

Diff since v8.2.0

Merged pull requests:

• 1.8 compatibility: fix cholesky nopivot argument. (#390) (@maleadt)

v8.2.0

GPUArrays v8.2.0

Diff since v8.1.3

Closed issues:

• Cartesian indexing issue for high dimensional arrays (#340)
• Too much specialization with permutedims: Use dynamic shared memory (#375)
• permutedims with data type UInt32 (#386)

Merged pull requests:

• new permutedims without unrolling permutation orders (#383) (@GiggleLiu)
• Overload cholesky method in order to prevent scalar indexing for Diagonal (#384) (@simsurace)
• speed up permutedims with UInt32 (#387) (@GiggleLiu)
• Call parent when unwrapping arrays. (#388) (@maleadt)

v8.1.3

GPUArrays v8.1.3

Diff since v8.1.2

Closed issues:

• Features we need (#30)
• Multiplying small Matrices (#45)
• LLVM error on A .^ n (#54)
• Julia Base GPU support (#57)
• norm on GPUArrays throws error if for more than 32 elements (#66)
• dot does not seem to be implemented for GPUArrays. (#122)
• Support for mathematical operations on complex numbers (#165)
• mapreduce of view(GPUArrays) dispatches to CPU (#182)
• AMDGPU.jl reverse-ci (#299)
• Compatibility with FillArrays (#381)

Merged pull requests:

• complex norm (#382) (@mloubout)

v8.1.2

GPUArrays v8.1.2

Diff since v8.1.1

Closed issues:

• Interface for aliasing an array (#377)

Merged pull requests:

• Allow use of CPU RNG without scalar iteration. (#378) (@maleadt)
• Make v and hcat with numbers work. (#379) (@maleadt)

v8.1.1

GPUArrays v8.1.1

Diff since v8.1.0

Merged pull requests:

• Don't test complex randn. (#376) (@maleadt)

v8.1.0

GPUArrays v8.1.0

Diff since v8.0.2

Merged pull requests:

• Allow tests to be run without Float64 and ComplexF64 types (#373) (@troels)
• Fix copying Diagonal matrices (#374) (@rossviljoen)

v8.0.2

GPUArrays v8.0.2

Diff since v8.0.1

Merged pull requests:

• Try to use the heuristic's block configuration when using grid-stride kernels. (#372) (@maleadt)