Move MTL tests and add a few

[christiangnrd]

No description provided.

[github-actions]

Metal Benchmarks

Benchmark suite	Current: 54aca63	Previous: e6abcc6	Ratio
private array/construct	27125 ns	26616.071428571428 ns	1.02
private array/broadcast	460000 ns	460792 ns	1.00
private array/random/randn/Float32	815249.5 ns	769959 ns	1.06
private array/random/randn!/Float32	654687.5 ns	646500 ns	1.01
private array/random/rand!/Int64	562667 ns	546000 ns	1.03
private array/random/rand!/Float32	600916.5 ns	586458 ns	1.02
private array/random/rand/Int64	775708 ns	777000 ns	1.00
private array/random/rand/Float32	595396 ns	585875 ns	1.02
private array/copyto!/gpu_to_gpu	655959 ns	659812 ns	0.99
private array/copyto!/cpu_to_gpu	825958 ns	823645.5 ns	1.00
private array/copyto!/gpu_to_cpu	647958 ns	624479.5 ns	1.04
private array/accumulate/1d	1322333.5 ns	1332291 ns	0.99
private array/accumulate/2d	1375083 ns	1376666 ns	1.00
private array/iteration/findall/int	2032875 ns	2040417 ns	1.00
private array/iteration/findall/bool	1771792 ns	1822812.5 ns	0.97
private array/iteration/findfirst/int	1689458 ns	1681646 ns	1.00
private array/iteration/findfirst/bool	1665062.5 ns	1650958 ns	1.01
private array/iteration/scalar	3791750 ns	3843750 ns	0.99
private array/iteration/logical	3149583 ns	3170916.5 ns	0.99
private array/iteration/findmin/1d	1754229 ns	1762708 ns	1.00
private array/iteration/findmin/2d	1341750 ns	1330854 ns	1.01
private array/reductions/reduce/1d	1028375 ns	1018354 ns	1.01
private array/reductions/reduce/2d	655792 ns	657208 ns	1.00
private array/reductions/mapreduce/1d	1026625 ns	1025958 ns	1.00
private array/reductions/mapreduce/2d	659250 ns	650333 ns	1.01
private array/permutedims/4d	2708834 ns	2706375 ns	1.00
private array/permutedims/2d	995334 ns	989125 ns	1.01
private array/permutedims/3d	1569167 ns	1562396 ns	1.00
private array/copy	518209 ns	599500 ns	0.86
latency/precompile	5140507458 ns	5119209208 ns	1.00
latency/ttfp	6632038000.5 ns	6736074937.5 ns	0.98
latency/import	1154911458.5 ns	1150978708 ns	1.00
integration/metaldevrt	725104.5 ns	712813 ns	1.02
integration/byval/slices=1	1559208 ns	1640250 ns	0.95
integration/byval/slices=3	8801000 ns	10578833.5 ns	0.83
integration/byval/reference	1587145.5 ns	1533458 ns	1.04
integration/byval/slices=2	2543125 ns	2647313 ns	0.96
kernel/indexing	448417 ns	493833 ns	0.91
kernel/indexing_checked	454145.5 ns	484521 ns	0.94
kernel/launch	8000 ns	9889 ns	0.81
metal/synchronization/stream	14250 ns	14792 ns	0.96
metal/synchronization/context	14791 ns	15041 ns	0.98
shared array/construct	25488.14285714286 ns	26827.428571428572 ns	0.95
shared array/broadcast	469916 ns	453250 ns	1.04
shared array/random/randn/Float32	849208 ns	762499.5 ns	1.11
shared array/random/randn!/Float32	652750 ns	644042 ns	1.01
shared array/random/rand!/Int64	562458 ns	538500 ns	1.04
shared array/random/rand!/Float32	601395.5 ns	570792 ns	1.05
shared array/random/rand/Int64	751354.5 ns	769625 ns	0.98
shared array/random/rand/Float32	614083 ns	645625 ns	0.95
shared array/copyto!/gpu_to_gpu	86917 ns	89416 ns	0.97
shared array/copyto!/cpu_to_gpu	87583 ns	86458 ns	1.01
shared array/copyto!/gpu_to_cpu	76916 ns	77292 ns	1.00
shared array/accumulate/1d	1333875 ns	1335958 ns	1.00
shared array/accumulate/2d	1383958 ns	1384187 ns	1.00
shared array/iteration/findall/int	1792792 ns	1755958.5 ns	1.02
shared array/iteration/findall/bool	1583792 ns	1576938 ns	1.00
shared array/iteration/findfirst/int	1385500 ns	1393500 ns	0.99
shared array/iteration/findfirst/bool	1359834 ns	1352667 ns	1.01
shared array/iteration/scalar	155583 ns	154625 ns	1.01
shared array/iteration/logical	2949583 ns	2967917 ns	0.99
shared array/iteration/findmin/1d	1447292 ns	1454875 ns	0.99
shared array/iteration/findmin/2d	1360250 ns	1360000 ns	1.00
shared array/reductions/reduce/1d	724875 ns	715000 ns	1.01
shared array/reductions/reduce/2d	662833.5 ns	660917 ns	1.00
shared array/reductions/mapreduce/1d	736125 ns	733708.5 ns	1.00
shared array/reductions/mapreduce/2d	664000 ns	658333.5 ns	1.01
shared array/permutedims/4d	3078708.5 ns	2704395.5 ns	1.14
shared array/permutedims/2d	997479 ns	998416 ns	1.00
shared array/permutedims/3d	1571916 ns	1578750.5 ns	1.00
shared array/copy	243833 ns	248708 ns	0.98

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