Metal library parsing: using CodecBzip2 feature to ignore padding.

[maleadt]

Depends on JuliaIO/CodecBzip2.jl#45

[github-actions]

Metal Benchmarks

Benchmark suite	Current: 5859853	Previous: 1820957	Ratio
private array/construct	26416.714285714286 ns	26795.083333333336 ns	0.99
private array/broadcast	462875 ns	465417 ns	0.99
private array/random/randn/Float32	911083.5 ns	766458 ns	1.19
private array/random/randn!/Float32	595667 ns	658791 ns	0.90
private array/random/rand!/Int64	553875 ns	565416 ns	0.98
private array/random/rand!/Float32	555125 ns	585541 ns	0.95
private array/random/rand/Int64	920917 ns	794583 ns	1.16
private array/random/rand/Float32	831125 ns	603208 ns	1.38
private array/copyto!/gpu_to_gpu	568709 ns	661083 ns	0.86
private array/copyto!/cpu_to_gpu	702500 ns	821334 ns	0.86
private array/copyto!/gpu_to_cpu	639292 ns	769291 ns	0.83
private array/accumulate/1d	1453000 ns	1343959 ns	1.08
private array/accumulate/2d	1555188 ns	1393208.5 ns	1.12
private array/iteration/findall/int	2338229 ns	2103125 ns	1.11
private array/iteration/findall/bool	2110499.5 ns	1841854 ns	1.15
private array/iteration/findfirst/int	1846416.5 ns	1694416 ns	1.09
private array/iteration/findfirst/bool	1752917 ns	1667875 ns	1.05
private array/iteration/scalar	2750416 ns	3979292 ns	0.69
private array/iteration/logical	3595625 ns	3188666 ns	1.13
private array/iteration/findmin/1d	1936916 ns	1761458 ns	1.10
private array/iteration/findmin/2d	1444541 ns	1362292 ns	1.06
private array/reductions/reduce/1d	1016542 ns	1043000 ns	0.97
private array/reductions/reduce/2d	713250 ns	663833.5 ns	1.07
private array/reductions/mapreduce/1d	1004000 ns	1059834 ns	0.95
private array/reductions/mapreduce/2d	710083 ns	662417 ns	1.07
private array/permutedims/4d	2694750 ns	2564437 ns	1.05
private array/permutedims/2d	1105541.5 ns	1031750 ns	1.07
private array/permutedims/3d	1857249.5 ns	1593354.5 ns	1.17
private array/copy	871584 ns	591521 ns	1.47
latency/precompile	5924859583.5 ns	5763160083.5 ns	1.03
latency/ttfp	6778519146 ns	6663954270.5 ns	1.02
latency/import	1195181250 ns	1169291583 ns	1.02
integration/metaldevrt	771458 ns	714333 ns	1.08
integration/byval/slices=1	1685500 ns	1647416 ns	1.02
integration/byval/slices=3	20474041.5 ns	8868125 ns	2.31
integration/byval/reference	1674125 ns	1566291 ns	1.07
integration/byval/slices=2	2839250 ns	2727167 ns	1.04
kernel/indexing	465792 ns	457375 ns	1.02
kernel/indexing_checked	474625 ns	456500 ns	1.04
kernel/launch	8125 ns	37291.75 ns	0.22
metal/synchronization/stream	14750 ns	14916 ns	0.99
metal/synchronization/context	15250 ns	15042 ns	1.01
shared array/construct	26089.285714285714 ns	27152.833333333332 ns	0.96
shared array/broadcast	463167 ns	472750 ns	0.98
shared array/random/randn/Float32	915562.5 ns	791708.5 ns	1.16
shared array/random/randn!/Float32	604333 ns	659791.5 ns	0.92
shared array/random/rand!/Int64	550916 ns	552083.5 ns	1.00
shared array/random/rand!/Float32	555125 ns	595834 ns	0.93
shared array/random/rand/Int64	954000 ns	782875 ns	1.22
shared array/random/rand/Float32	829917 ns	578000 ns	1.44
shared array/copyto!/gpu_to_gpu	80666 ns	86667 ns	0.93
shared array/copyto!/cpu_to_gpu	82417 ns	89875 ns	0.92
shared array/copyto!/gpu_to_cpu	83000 ns	78083 ns	1.06
shared array/accumulate/1d	1485187.5 ns	1354292 ns	1.10
shared array/accumulate/2d	1541375 ns	1402292 ns	1.10
shared array/iteration/findall/int	2077667 ns	1845625 ns	1.13
shared array/iteration/findall/bool	1800834 ns	1610291 ns	1.12
shared array/iteration/findfirst/int	1530333 ns	1395270.5 ns	1.10
shared array/iteration/findfirst/bool	1467083 ns	1363083.5 ns	1.08
shared array/iteration/scalar	156375 ns	158584 ns	0.99
shared array/iteration/logical	3370375 ns	2985875 ns	1.13
shared array/iteration/findmin/1d	1621958.5 ns	1464479.5 ns	1.11
shared array/iteration/findmin/2d	1470479 ns	1373417 ns	1.07
shared array/reductions/reduce/1d	740041.5 ns	739562.5 ns	1.00
shared array/reductions/reduce/2d	717708 ns	671771 ns	1.07
shared array/reductions/mapreduce/1d	741187.5 ns	751584 ns	0.99
shared array/reductions/mapreduce/2d	732833 ns	668417 ns	1.10
shared array/permutedims/4d	2708250.5 ns	2553000 ns	1.06
shared array/permutedims/2d	1117208 ns	1035875 ns	1.08
shared array/permutedims/3d	1873479.5 ns	1593333 ns	1.18
shared array/copy	217833 ns	251771 ns	0.87

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