Fix benchmarking CI and benchmark Shared and Private storage modes

[christiangnrd]

No description provided.

[github-actions]

Metal Benchmarks

Benchmark suite	Current: 192837b	Previous: 8652754	Ratio
private array/construct	27656.333333333332 ns
private array/broadcast	263166.5 ns
private array/random/randn/Float32	522750 ns
private array/random/randn!/Float32	408291.5 ns
private array/random/rand!/Int64	380312.5 ns
private array/random/rand!/Float32	340687.5 ns
private array/random/rand/Int64	709250 ns
private array/random/rand/Float32	374479.5 ns
private array/copyto!/gpu_to_gpu	226000 ns
private array/copyto!/cpu_to_gpu	254916 ns
private array/copyto!/gpu_to_cpu	254625 ns
private array/accumulate/1d	7988500 ns
private array/accumulate/2d	1008583 ns
private array/iteration/findall/int	7984978.5 ns
private array/iteration/findall/bool	7978083.5 ns
private array/iteration/findfirst/int	1173333 ns
private array/iteration/findfirst/bool	1125375 ns
private array/iteration/scalar	1533125 ns
private array/iteration/logical	8185750 ns
private array/iteration/findmin/1d	1090542 ns
private array/iteration/findmin/2d	880292 ns
private array/reductions/reduce/1d	484459 ns
private array/reductions/reduce/2d	503375 ns
private array/reductions/mapreduce/1d	486667 ns
private array/reductions/mapreduce/2d	500625 ns
private array/permutedims/4d	686416.5 ns
private array/permutedims/2d	671833 ns
private array/permutedims/3d	858541 ns
private array/copy	343520.5 ns
latency/precompile	4419564583 ns	4401680834 ns	1.00
latency/ttfp	6722446521 ns	6678542687 ns	1.01
latency/import	723983584 ns	721498042 ns	1.00
integration/metaldevrt	564500 ns	708167 ns	0.80
integration/byval/slices=1	1594708 ns	1530625 ns	1.04
integration/byval/slices=3	11814354.5 ns	11010542 ns	1.07
integration/byval/reference	1597166 ns	1585084 ns	1.01
integration/byval/slices=2	2683500 ns	2472708 ns	1.09
kernel/indexing	244937 ns	454333 ns	0.54
kernel/indexing_checked	251416 ns	455667 ns	0.55
kernel/launch	58361 ns	8459 ns	6.90
metal/synchronization/stream	14270.5 ns	14667 ns	0.97
metal/synchronization/context	14833 ns	14916 ns	0.99
shared array/construct	24441.699999999997 ns
shared array/broadcast	263125 ns
shared array/random/randn/Float32	532958 ns
shared array/random/randn!/Float32	376062.5 ns
shared array/random/rand!/Int64	384229 ns
shared array/random/rand!/Float32	346334 ns
shared array/random/rand/Int64	663792 ns
shared array/random/rand/Float32	461000 ns
shared array/copyto!/gpu_to_gpu	276250 ns
shared array/copyto!/cpu_to_gpu	87458 ns
shared array/copyto!/gpu_to_cpu	82542 ns
shared array/accumulate/1d	7988875 ns
shared array/accumulate/2d	1011584 ns
shared array/iteration/findall/int	7978042 ns
shared array/iteration/findall/bool	7979895.5 ns
shared array/iteration/findfirst/int	977250 ns
shared array/iteration/findfirst/bool	936750 ns
shared array/iteration/scalar	153792 ns
shared array/iteration/logical	7999667 ns
shared array/iteration/findmin/1d	893583.5 ns
shared array/iteration/findmin/2d	893500 ns
shared array/reductions/reduce/1d	372854 ns
shared array/reductions/reduce/2d	511375 ns
shared array/reductions/mapreduce/1d	381042 ns
shared array/reductions/mapreduce/2d	506208 ns
shared array/permutedims/4d	672625 ns
shared array/permutedims/2d	676541.5 ns
shared array/permutedims/3d	874292 ns
shared array/copy	514083 ns

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