don't return reshaped array from mapreducedim!

[simeonschaub]

Same as JuliaGPU/CUDA.jl#3219, discovered in JuliaGPU/GPUArrays.jl#754.

[github-actions]

AMDGPU.jl Benchmarks

Details

Benchmark suite	Current: 74626e8	Previous: f39609f	Ratio
amdgpu/synchronization/context/device	572.5 ns	577.5 ns	0.99
amdgpu/synchronization/stream/blocking	237.5 ns	242.5 ns	0.98
amdgpu/synchronization/stream/nonblocking	327.5 ns	327.5 ns	1
array/accumulate/Float32/1d	71748.5 ns	72208.5 ns	0.99
array/accumulate/Float32/dims=1	251706.25 ns	273376.5 ns	0.92
array/accumulate/Float32/dims=1L	77708.75 ns	79091 ns	0.98
array/accumulate/Float32/dims=2	71488.5 ns	73971.25 ns	0.97
array/accumulate/Float32/dims=2L	2621822.75 ns	2625672 ns	1.00
array/accumulate/Int64/1d	75803.5 ns	78243.75 ns	0.97
array/accumulate/Int64/dims=1	245016.25 ns	245298.5 ns	1.00
array/accumulate/Int64/dims=1L	83688.75 ns	83871 ns	1.00
array/accumulate/Int64/dims=2	81073.5 ns	80931.25 ns	1.00
array/accumulate/Int64/dims=2L	2866528.75 ns	2865473 ns	1.00
array/broadcast	73581.25 ns	73611 ns	1.00
array/construct	2202.5 ns	2300.25 ns	0.96
array/copy	36890.75 ns	35548 ns	1.04
array/copyto!/cpu_to_gpu	112201.75 ns	112059 ns	1.00
array/copyto!/gpu_to_cpu	105589 ns	104761.5 ns	1.01
array/copyto!/gpu_to_gpu	35578 ns	49765.75 ns	0.71
array/iteration/findall/bool	136204.25 ns	141354.5 ns	0.96
array/iteration/findall/int	149397 ns	156922.25 ns	0.95
array/iteration/findfirst/bool	120581.5 ns	124271.75 ns	0.97
array/iteration/findfirst/int	123874.25 ns	124834.25 ns	0.99
array/iteration/findmin/1d	109854.25 ns	114066.75 ns	0.96
array/iteration/findmin/2d	96911.5 ns	108581.5 ns	0.89
array/iteration/logical	246273.75 ns	256408.5 ns	0.96
array/iteration/scalar	295689.25 ns	296556.5 ns	1.00
array/permutedims/2d	71793.5 ns	71528.5 ns	1.00
array/permutedims/3d	70966 ns	71013.5 ns	1.00
array/permutedims/4d	74128.75 ns	74041 ns	1.00
array/random/rand/Float32	45115.5 ns	45140.5 ns	1.00
array/random/rand/Int64	54613.25 ns	55005.75 ns	0.99
array/random/rand!/Float32	65173.5 ns	65316 ns	1.00
array/random/rand!/Int64	70856 ns	73568.5 ns	0.96
array/random/randn/Float32	76241 ns	77641.25 ns	0.98
array/random/randn!/Float32	81081.25 ns	75446.25 ns	1.07
array/reductions/mapreduce/Float32/1d	95931.5 ns	88421.25 ns	1.08
array/reductions/mapreduce/Float32/dims=1	84986.25 ns	84406.25 ns	1.01
array/reductions/mapreduce/Float32/dims=1L	782346.25 ns	782181 ns	1.00
array/reductions/mapreduce/Float32/dims=2	84338.75 ns	83961.25 ns	1.00
array/reductions/mapreduce/Float32/dims=2L	143072 ns	142397 ns	1.00
array/reductions/mapreduce/Int64/1d	96124 ns	88046.25 ns	1.09
array/reductions/mapreduce/Int64/dims=1	84561.25 ns	84006 ns	1.01
array/reductions/mapreduce/Int64/dims=1L	787096.25 ns	791836.25 ns	0.99
array/reductions/mapreduce/Int64/dims=2	83973.75 ns	84366 ns	1.00
array/reductions/mapreduce/Int64/dims=2L	143324.5 ns	143969.5 ns	1.00
array/reductions/reduce/Float32/1d	88228.75 ns	88223.75 ns	1.00
array/reductions/reduce/Float32/dims=1	84628.75 ns	83906.25 ns	1.01
array/reductions/reduce/Float32/dims=1L	782741.25 ns	786496.25 ns	1.00
array/reductions/reduce/Float32/dims=2	84058.75 ns	84028.75 ns	1.00
array/reductions/reduce/Float32/dims=2L	142632 ns	142229.5 ns	1.00
array/reductions/reduce/Int64/1d	88468.75 ns	88981.25 ns	0.99
array/reductions/reduce/Int64/dims=1	83926.25 ns	84336.25 ns	1.00
array/reductions/reduce/Int64/dims=1L	790876.5 ns	789723.75 ns	1.00
array/reductions/reduce/Int64/dims=2	83943.75 ns	84358.75 ns	1.00
array/reductions/reduce/Int64/dims=2L	143332 ns	143122 ns	1.00
array/reverse/1d	43338 ns	43295.5 ns	1.00
array/reverse/1dL	71173.5 ns	71186 ns	1.00
array/reverse/1dL_inplace	78806.25 ns	70888.5 ns	1.11
array/reverse/1d_inplace	59848.5 ns	59876 ns	1.00
array/reverse/2d	49208.25 ns	48905.75 ns	1.01
array/reverse/2dL	79771.25 ns	79303.5 ns	1.01
array/reverse/2dL_inplace	91249 ns	91751.25 ns	0.99
array/reverse/2d_inplace	62396 ns	62295.75 ns	1.00
array/sorting/1d	337515 ns	341022.25 ns	0.99
integration/byval/reference	39410 ns	39651 ns	0.99
integration/byval/slices=1	40501 ns	40960 ns	0.99
integration/byval/slices=2	140342 ns	152982 ns	0.92
integration/byval/slices=3	236574 ns	247524 ns	0.96
integration/volumerhs	5051703 ns	5045271 ns	1.00
kernel/indexing	57478.5 ns	47473.25 ns	1.21
kernel/indexing_checked	58885.75 ns	58823.25 ns	1.00
kernel/launch	1552.5 ns	1552.5 ns	1
kernel/rand	98108.75 ns	98511.5 ns	1.00
latency/import	1720869481 ns	1720954000 ns	1.00
latency/precompile	37413313176 ns	37438172352 ns	1.00
latency/ttfp	5739159769 ns	5748382815 ns	1.00

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

[luraess]

Thanks!