[0.10] add get_backend for StaticArrays

[vchuravy]

Port of #621

[christiangnrd]

Guess we need to find an alternative to MArray for KA.scratchpad or re-add StaticArrays.jl as a full dependency.

KernelAbstractions.jl/src/pocl/backend.jl

Lines 224 to 226 in 4409708

	@device_override @inline function KA.Scratchpad(ctx, ::Type{T}, ::Val{Dims}) where {T, Dims}
	StaticArrays.MArray{KA.__size(Dims), T}(undef)
	end

[github-actions]

Benchmark Results

Show table

	main	9f47797...	main / 9f47797...
saxpy/default/Float32/1024	0.087 ± 0.011 ms	0.0868 ± 0.0092 ms	1 ± 0.16
saxpy/default/Float32/1048576	0.506 ± 0.027 ms	0.501 ± 0.024 ms	1.01 ± 0.073
saxpy/default/Float32/16384	0.0689 ± 0.035 ms	0.0697 ± 0.035 ms	0.989 ± 0.71
saxpy/default/Float32/2048	0.0859 ± 0.027 ms	0.0865 ± 0.026 ms	0.994 ± 0.43
saxpy/default/Float32/256	0.0874 ± 0.0099 ms	0.0873 ± 0.012 ms	1 ± 0.18
saxpy/default/Float32/262144	0.178 ± 0.031 ms	0.18 ± 0.032 ms	0.991 ± 0.25
saxpy/default/Float32/32768	0.0754 ± 0.031 ms	0.0764 ± 0.032 ms	0.986 ± 0.58
saxpy/default/Float32/4096	0.0836 ± 0.032 ms	0.0828 ± 0.03 ms	1.01 ± 0.53
saxpy/default/Float32/512	0.0866 ± 0.01 ms	0.0866 ± 0.012 ms	1 ± 0.18
saxpy/default/Float32/64	0.0871 ± 0.012 ms	0.0871 ± 0.012 ms	1 ± 0.2
saxpy/default/Float32/65536	0.0895 ± 0.031 ms	0.0924 ± 0.032 ms	0.968 ± 0.47
saxpy/default/Float64/1024	0.0864 ± 0.017 ms	0.0869 ± 0.013 ms	0.994 ± 0.25
saxpy/default/Float64/1048576	0.588 ± 0.078 ms	0.598 ± 0.1 ms	0.984 ± 0.21
saxpy/default/Float64/16384	0.0696 ± 0.031 ms	0.0684 ± 0.031 ms	1.02 ± 0.65
saxpy/default/Float64/2048	0.0845 ± 0.031 ms	0.0856 ± 0.029 ms	0.987 ± 0.49
saxpy/default/Float64/256	0.0867 ± 0.0096 ms	0.0868 ± 0.011 ms	0.999 ± 0.17
saxpy/default/Float64/262144	0.197 ± 0.037 ms	0.197 ± 0.039 ms	0.999 ± 0.27
saxpy/default/Float64/32768	0.0781 ± 0.029 ms	0.0796 ± 0.031 ms	0.981 ± 0.53
saxpy/default/Float64/4096	0.0805 ± 0.03 ms	0.0827 ± 0.029 ms	0.973 ± 0.49
saxpy/default/Float64/512	0.0867 ± 0.009 ms	0.0869 ± 0.0092 ms	0.998 ± 0.15
saxpy/default/Float64/64	0.0867 ± 0.013 ms	0.0875 ± 0.011 ms	0.99 ± 0.2
saxpy/default/Float64/65536	0.0952 ± 0.03 ms	0.0992 ± 0.032 ms	0.959 ± 0.43
saxpy/static workgroup=(1024,)/Float32/1024	0.0846 ± 0.014 ms	0.0838 ± 0.012 ms	1.01 ± 0.22
saxpy/static workgroup=(1024,)/Float32/1048576	0.43 ± 0.029 ms	0.434 ± 0.029 ms	0.99 ± 0.094
saxpy/static workgroup=(1024,)/Float32/16384	0.0662 ± 0.033 ms	0.0666 ± 0.034 ms	0.994 ± 0.71
saxpy/static workgroup=(1024,)/Float32/2048	0.0833 ± 0.028 ms	0.0839 ± 0.027 ms	0.993 ± 0.46
saxpy/static workgroup=(1024,)/Float32/256	0.0847 ± 0.012 ms	0.0846 ± 0.01 ms	1 ± 0.19
saxpy/static workgroup=(1024,)/Float32/262144	0.16 ± 0.033 ms	0.16 ± 0.033 ms	1 ± 0.29
saxpy/static workgroup=(1024,)/Float32/32768	0.0706 ± 0.029 ms	0.0702 ± 0.03 ms	1.01 ± 0.6
saxpy/static workgroup=(1024,)/Float32/4096	0.0803 ± 0.03 ms	0.0807 ± 0.029 ms	0.995 ± 0.52
saxpy/static workgroup=(1024,)/Float32/512	0.0846 ± 0.013 ms	0.0843 ± 0.01 ms	1 ± 0.19
saxpy/static workgroup=(1024,)/Float32/64	0.0847 ± 0.012 ms	0.0848 ± 0.012 ms	0.999 ± 0.2
saxpy/static workgroup=(1024,)/Float32/65536	0.084 ± 0.031 ms	0.0849 ± 0.032 ms	0.99 ± 0.52
saxpy/static workgroup=(1024,)/Float64/1024	0.0842 ± 0.014 ms	0.0839 ± 0.013 ms	1 ± 0.23
saxpy/static workgroup=(1024,)/Float64/1048576	0.549 ± 0.1 ms	0.556 ± 0.098 ms	0.986 ± 0.25
saxpy/static workgroup=(1024,)/Float64/16384	0.0658 ± 0.028 ms	0.0653 ± 0.03 ms	1.01 ± 0.64
saxpy/static workgroup=(1024,)/Float64/2048	0.0828 ± 0.03 ms	0.0817 ± 0.031 ms	1.01 ± 0.54
saxpy/static workgroup=(1024,)/Float64/256	0.0846 ± 0.0094 ms	0.0843 ± 0.01 ms	1 ± 0.16
saxpy/static workgroup=(1024,)/Float64/262144	0.187 ± 0.036 ms	0.195 ± 0.037 ms	0.959 ± 0.26
saxpy/static workgroup=(1024,)/Float64/32768	0.0769 ± 0.028 ms	0.0767 ± 0.028 ms	1 ± 0.52
saxpy/static workgroup=(1024,)/Float64/4096	0.0777 ± 0.03 ms	0.0783 ± 0.029 ms	0.992 ± 0.53
saxpy/static workgroup=(1024,)/Float64/512	0.0845 ± 0.01 ms	0.084 ± 0.012 ms	1.01 ± 0.19
saxpy/static workgroup=(1024,)/Float64/64	0.0846 ± 0.01 ms	0.0846 ± 0.01 ms	1 ± 0.17
saxpy/static workgroup=(1024,)/Float64/65536	0.0952 ± 0.029 ms	0.0947 ± 0.029 ms	1 ± 0.43
time_to_load	0.954 ± 0.033 s	0.937 ± 0.0048 s	1.02 ± 0.036

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).