update GPUToolbox to v0.2

[vchuravy]

No description provided.

[maleadt]

Why not simply import everything? We control that package, after all.

[github-actions]

CUDA.jl Benchmarks

Benchmark suite	Current: 9161a0e	Previous: c75b56f	Ratio
latency/precompile	46241783821 ns	46181620127.5 ns	1.00
latency/ttfp	7104757954 ns	7068629094 ns	1.01
latency/import	3740569375 ns	3724762175 ns	1.00
integration/volumerhs	9624153.5 ns	9624216.5 ns	1.00
integration/byval/slices=1	146935 ns	146894 ns	1.00
integration/byval/slices=3	425223 ns	425137.5 ns	1.00
integration/byval/reference	144909 ns	144952 ns	1.00
integration/byval/slices=2	286077 ns	285936 ns	1.00
integration/cudadevrt	103282 ns	103412 ns	1.00
kernel/indexing	14211 ns	14099 ns	1.01
kernel/indexing_checked	14757 ns	14674 ns	1.01
kernel/occupancy	671.8037974683544 ns	701.1379310344828 ns	0.96
kernel/launch	2059.7 ns	2179.6666666666665 ns	0.94
kernel/rand	14590 ns	14749 ns	0.99
array/reverse/1d	19644 ns	19776 ns	0.99
array/reverse/2d	25320 ns	24908 ns	1.02
array/reverse/1d_inplace	10356.5 ns	10219 ns	1.01
array/reverse/2d_inplace	12076 ns	11910 ns	1.01
array/copy	21226 ns	21311 ns	1.00
array/iteration/findall/int	159303 ns	158209 ns	1.01
array/iteration/findall/bool	139699 ns	139123 ns	1.00
array/iteration/findfirst/int	154653 ns	153168 ns	1.01
array/iteration/findfirst/bool	155261 ns	154631 ns	1.00
array/iteration/scalar	72166 ns	71886 ns	1.00
array/iteration/logical	216022.5 ns	213254 ns	1.01
array/iteration/findmin/1d	41704 ns	40786 ns	1.02
array/iteration/findmin/2d	94285 ns	93428 ns	1.01
array/reductions/reduce/1d	40190.5 ns	35669 ns	1.13
array/reductions/reduce/2d	51223.5 ns	40477 ns	1.27
array/reductions/mapreduce/1d	38542 ns	33443 ns	1.15
array/reductions/mapreduce/2d	51598 ns	40694.5 ns	1.27
array/broadcast	20887 ns	20825 ns	1.00
array/copyto!/gpu_to_gpu	11687 ns	13806 ns	0.85
array/copyto!/cpu_to_gpu	210260 ns	208873 ns	1.01
array/copyto!/gpu_to_cpu	245325 ns	242948 ns	1.01
array/accumulate/1d	109209 ns	108924 ns	1.00
array/accumulate/2d	79725.5 ns	80034 ns	1.00
array/construct	1312.6 ns	1297.3 ns	1.01
array/random/randn/Float32	44061 ns	43906.5 ns	1.00
array/random/randn!/Float32	26846 ns	26669 ns	1.01
array/random/rand!/Int64	27252 ns	27027 ns	1.01
array/random/rand!/Float32	8717.333333333334 ns	8863 ns	0.98
array/random/rand/Int64	33643 ns	30048.5 ns	1.12
array/random/rand/Float32	13102 ns	13342 ns	0.98
array/permutedims/4d	61149.5 ns	60675.5 ns	1.01
array/permutedims/2d	55648.5 ns	55115.5 ns	1.01
array/permutedims/3d	56588 ns	55700 ns	1.02
array/sorting/1d	2766862 ns	2777689 ns	1.00
array/sorting/by	3354947 ns	3368739 ns	1.00
array/sorting/2d	1085516.5 ns	1084912 ns	1.00
cuda/synchronization/stream/auto	1039.5 ns	1013.5384615384615 ns	1.03
cuda/synchronization/stream/nonblocking	6487.8 ns	6485.2 ns	1.00
cuda/synchronization/stream/blocking	840.1782178217821 ns	826 ns	1.02
cuda/synchronization/context/auto	1161.2 ns	1197.3 ns	0.97
cuda/synchronization/context/nonblocking	6706.8 ns	6768.6 ns	0.99
cuda/synchronization/context/blocking	931.1578947368421 ns	946.4193548387096 ns	0.98

This comment was automatically generated by workflow using github-action-benchmark.