KernelAbstractions 0.10

[christiangnrd]

Not a draft to also run benchmarks

[github-actions]

Metal Benchmarks

Details

Benchmark suite	Current: 2d2a094	Previous: 7c5b7d4	Ratio
array/accumulate/Float32/1d	404167 ns	405125 ns	1.00
array/accumulate/Float32/dims=1	378125 ns	380709 ns	0.99
array/accumulate/Float32/dims=1L	8801292 ns	8813708 ns	1.00
array/accumulate/Float32/dims=2	439792 ns	449833 ns	0.98
array/accumulate/Float32/dims=2L	2633125 ns	2672417 ns	0.99
array/accumulate/Int64/1d	914792 ns	857042 ns	1.07
array/accumulate/Int64/dims=1	849584 ns	934334 ns	0.91
array/accumulate/Int64/dims=1L	9677125 ns	9530500 ns	1.02
array/accumulate/Int64/dims=2	1390584 ns	1253708 ns	1.11
array/accumulate/Int64/dims=2L	6523083 ns	6539791 ns	1.00
array/broadcast	210792 ns	186583 ns	1.13
array/construct	2166 ns	2208 ns	0.98
array/permutedims/2d	672167 ns	545333 ns	1.23
array/permutedims/3d	1237916 ns	1066167 ns	1.16
array/permutedims/4d	1427709 ns	1252917 ns	1.14
array/private/copy	239500 ns	235208 ns	1.02
array/private/copyto!/cpu_to_gpu	195500 ns	199791 ns	0.98
array/private/copyto!/gpu_to_cpu	196250 ns	199709 ns	0.98
array/private/copyto!/gpu_to_gpu	197584 ns	196458 ns	1.01
array/private/iteration/findall/bool	1124333 ns	1075000 ns	1.05
array/private/iteration/findall/int	1281500 ns	1238042 ns	1.04
array/private/iteration/findfirst/bool	1182208 ns	1101750 ns	1.07
array/private/iteration/findfirst/int	1165625 ns	1119041 ns	1.04
array/private/iteration/findmin/1d	1295500 ns	1239875 ns	1.04
array/private/iteration/findmin/2d	1091417 ns	1086292 ns	1.00
array/private/iteration/logical	1808459 ns	1752250 ns	1.03
array/private/iteration/scalar	1175500 ns	1229458 ns	0.96
array/random/rand/Float32	429500 ns	430042 ns	1.00
array/random/rand/Int64	519875 ns	526708 ns	0.99
array/random/rand!/Float32	320458 ns	370666 ns	0.86
array/random/rand!/Int64	316584 ns	401208 ns	0.79
array/random/randn/Float32	388458 ns	388208 ns	1.00
array/random/randn!/Float32	334958 ns	322500 ns	1.04
array/reductions/mapreduce/Float32/1d	400875 ns	410500 ns	0.98
array/reductions/mapreduce/Float32/dims=1	331584 ns	297000 ns	1.12
array/reductions/mapreduce/Float32/dims=1L	625709 ns	635084 ns	0.99
array/reductions/mapreduce/Float32/dims=2	370750 ns	332959 ns	1.11
array/reductions/mapreduce/Float32/dims=2L	1033833 ns	1023375 ns	1.01
array/reductions/mapreduce/Int64/1d	577292 ns	586875 ns	0.98
array/reductions/mapreduce/Int64/dims=1	597625 ns	620583 ns	0.96
array/reductions/mapreduce/Int64/dims=1L	1040250 ns	1044000 ns	1.00
array/reductions/mapreduce/Int64/dims=2	725625 ns	541875 ns	1.34
array/reductions/mapreduce/Int64/dims=2L	2257500 ns	2194542 ns	1.03
array/reductions/reduce/Float32/1d	406250 ns	405708 ns	1.00
array/reductions/reduce/Float32/dims=1	333542 ns	283625 ns	1.18
array/reductions/reduce/Float32/dims=1L	630416 ns	638291 ns	0.99
array/reductions/reduce/Float32/dims=2	226709 ns	226208 ns	1.00
array/reductions/reduce/Float32/dims=2L	457708 ns	470917 ns	0.97
array/reductions/reduce/Int64/1d	575834 ns	543209 ns	1.06
array/reductions/reduce/Int64/dims=1	601542 ns	633083 ns	0.95
array/reductions/reduce/Int64/dims=1L	1038167 ns	1033417 ns	1.00
array/reductions/reduce/Int64/dims=2	244792 ns	233958 ns	1.05
array/reductions/reduce/Int64/dims=2L	652083 ns	662833 ns	0.98
array/shared/copy	135292 ns	135666 ns	1.00
array/shared/copyto!/cpu_to_gpu	37958 ns	37250 ns	1.02
array/shared/copyto!/gpu_to_cpu	38333 ns	37500 ns	1.02
array/shared/copyto!/gpu_to_gpu	38625 ns	38375 ns	1.01
array/shared/iteration/findall/bool	1150416 ns	1091042 ns	1.05
array/shared/iteration/findall/int	1293833 ns	1249125 ns	1.04
array/shared/iteration/findfirst/bool	855208 ns	902250 ns	0.95
array/shared/iteration/findfirst/int	872750 ns	977458 ns	0.89
array/shared/iteration/findmin/1d	1162584 ns	866167 ns	1.34
array/shared/iteration/findmin/2d	1091625 ns	791583 ns	1.38
array/shared/iteration/logical	1679250 ns	1582166 ns	1.06
array/shared/iteration/scalar	3593.75 ns	3890.625 ns	0.92
array/sorting/1d	1785375 ns	1624625 ns	1.10
array/sorting/2d	8309958 ns	8289250 ns	1.00
integration/byval/reference	1112209 ns	1110334 ns	1.00
integration/byval/slices=1	1113333 ns	1116542 ns	1.00
integration/byval/slices=2	2032250 ns	2020917 ns	1.01
integration/byval/slices=3	6877709 ns	9519250 ns	0.72
integration/metaldevrt	374875 ns	312708 ns	1.20
kernel/indexing	182917 ns	191000 ns	0.96
kernel/indexing_checked	351667 ns	366959 ns	0.96
kernel/launch	1895.9 ns	1850 ns	1.02
kernel/rand	375208 ns	377666 ns	0.99
latency/import	2125140000 ns	2163364250 ns	0.98
latency/precompile	54332746125 ns	42157958500 ns	1.29
latency/ttfp	2562425125 ns	2484808458 ns	1.03
metal/synchronization/context	644.2095808383234 ns	665.6139240506329 ns	0.97
metal/synchronization/stream	335.6090909090909 ns	338.27853881278537 ns	0.99

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

[codecov]

Codecov Report

❌ Patch coverage is 96.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 86.26%. Comparing base (265ec1d) to head (7f1d122).

Files with missing lines	Patch %	Lines
src/MetalKernels.jl	94.11%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #688   +/-   ##
=======================================
  Coverage   86.25%   86.26%           
=======================================
  Files          76       76           
  Lines        5296     5314   +18     
=======================================
+ Hits         4568     4584   +16     
- Misses        728      730    +2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[maleadt]

#994