Fix atomic cmpexchg intrinsic signature

[christiangnrd]

The macOS 26 shader validation layer gets upset otherwise.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.56%. Comparing base (ae7ee24) to head (7ec5fbf).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #794   +/-   ##
=======================================
  Coverage   81.56%   81.56%           
=======================================
  Files          66       66           
  Lines        3141     3141           
=======================================
  Hits         2562     2562           
  Misses        579      579           

☔ View full report in Codecov by Sentry.
📢 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.

[github-actions]

Metal Benchmarks

Details

Benchmark suite	Current: 7ec5fbf	Previous: ae7ee24	Ratio
array/accumulate/Float32/1d	816458 ns	812000 ns	1.01
array/accumulate/Float32/dims=1	997667 ns	981104 ns	1.02
array/accumulate/Float32/dims=1L	10515187.5 ns	10016291 ns	1.05
array/accumulate/Float32/dims=2	1280542 ns	1244500 ns	1.03
array/accumulate/Float32/dims=2L	4841417 ns	6573125.5 ns	0.74
array/accumulate/Int64/1d	982083 ns	961959 ns	1.02
array/accumulate/Int64/dims=1	1113417 ns	1142458 ns	0.97
array/accumulate/Int64/dims=1L	12516417 ns	11897834 ns	1.05
array/accumulate/Int64/dims=2	1473479.5 ns	1458500 ns	1.01
array/accumulate/Int64/dims=2L	9146792 ns	9470625 ns	0.97
array/broadcast	334416.5 ns	363750 ns	0.92
array/construct	5625 ns	5791 ns	0.97
array/permutedims/2d	627875 ns	614166.5 ns	1.02
array/permutedims/3d	1129625 ns	1128000 ns	1.00
array/permutedims/4d	1358750 ns	1977209 ns	0.69
array/private/copy	358041.5 ns	422833.5 ns	0.85
array/private/copyto!/cpu_to_gpu	239667 ns	367208 ns	0.65
array/private/copyto!/gpu_to_cpu	235875 ns	364333.5 ns	0.65
array/private/copyto!/gpu_to_gpu	254479.5 ns	332458 ns	0.77
array/private/iteration/findall/bool	1129792 ns	1071333 ns	1.05
array/private/iteration/findall/int	1292000 ns	1244125 ns	1.04
array/private/iteration/findfirst/bool	1260625 ns	1326875 ns	0.95
array/private/iteration/findfirst/int	1305792 ns	1360792 ns	0.96
array/private/iteration/findmin/1d	1430250 ns	1453000 ns	0.98
array/private/iteration/findmin/2d	1183584 ns	1209959 ns	0.98
array/private/iteration/logical	1770417 ns	1627666 ns	1.09
array/private/iteration/scalar	1693500 ns	2603000 ns	0.65
array/random/rand/Float32	599792 ns	618416.5 ns	0.97
array/random/rand/Int64	654833 ns	679750 ns	0.96
array/random/rand!/Float32	526750 ns	552084 ns	0.95
array/random/rand!/Int64	482125 ns	504416 ns	0.96
array/random/randn/Float32	572166.5 ns	600000 ns	0.95
array/random/randn!/Float32	487000 ns	534333 ns	0.91
array/reductions/mapreduce/Float32/1d	496292 ns	734458.5 ns	0.68
array/reductions/mapreduce/Float32/dims=1	459458 ns	499667 ns	0.92
array/reductions/mapreduce/Float32/dims=1L	717937 ns	765000 ns	0.94
array/reductions/mapreduce/Float32/dims=2	461854.5 ns	500750 ns	0.92
array/reductions/mapreduce/Float32/dims=2L	1033104 ns	1333208 ns	0.77
array/reductions/mapreduce/Int64/1d	792375 ns	924083.5 ns	0.86
array/reductions/mapreduce/Int64/dims=1	771021 ns	789334 ns	0.98
array/reductions/mapreduce/Int64/dims=1L	1281209 ns	1667542 ns	0.77
array/reductions/mapreduce/Int64/dims=2	955416.5 ns	989083 ns	0.97
array/reductions/mapreduce/Int64/dims=2L	2261292 ns	2259083 ns	1.00
array/reductions/reduce/Float32/1d	492083 ns	727333 ns	0.68
array/reductions/reduce/Float32/dims=1	461000 ns	495416 ns	0.93
array/reductions/reduce/Float32/dims=1L	715770.5 ns	844021 ns	0.85
array/reductions/reduce/Float32/dims=2	461416.5 ns	494041 ns	0.93
array/reductions/reduce/Float32/dims=2L	1039167 ns	1347166.5 ns	0.77
array/reductions/reduce/Int64/1d	791750 ns	913917 ns	0.87
array/reductions/reduce/Int64/dims=1	766125 ns	775500 ns	0.99
array/reductions/reduce/Int64/dims=1L	1218833 ns	1723125 ns	0.71
array/reductions/reduce/Int64/dims=2	961041 ns	953708 ns	1.01
array/reductions/reduce/Int64/dims=2L	2261875 ns	2260834 ns	1.00
array/shared/copy	162750 ns	237020.5 ns	0.69
array/shared/copyto!/cpu_to_gpu	41542 ns	41125 ns	1.01
array/shared/copyto!/gpu_to_cpu	47833 ns	42042 ns	1.14
array/shared/copyto!/gpu_to_gpu	40209 ns	41000 ns	0.98
array/shared/iteration/findall/bool	1135166 ns	1075541 ns	1.06
array/shared/iteration/findall/int	1291708 ns	1246917 ns	1.04
array/shared/iteration/findfirst/bool	1040479.5 ns	1078458.5 ns	0.96
array/shared/iteration/findfirst/int	1087709 ns	1080625 ns	1.01
array/shared/iteration/findmin/1d	1207687.5 ns	1210417 ns	1.00
array/shared/iteration/findmin/2d	1184083 ns	1220500 ns	0.97
array/shared/iteration/logical	1624542 ns	1470292 ns	1.10
array/shared/iteration/scalar	5826.5 ns	5937.5 ns	0.98
integration/byval/reference	1173729.5 ns	1157812.5 ns	1.01
integration/byval/slices=1	1174667 ns	1157250 ns	1.02
integration/byval/slices=2	2118166.5 ns	2087500 ns	1.01
integration/byval/slices=3	20220250 ns	7967250 ns	2.54
integration/metaldevrt	457792 ns	463083 ns	0.99
kernel/indexing	316916.5 ns	354291 ns	0.89
kernel/indexing_checked	329583 ns	356375 ns	0.92
kernel/launch	13542 ns	13458 ns	1.01
kernel/rand	345875 ns	372625 ns	0.93
latency/import	1408823896 ns	1387370125 ns	1.02
latency/precompile	30185175187.5 ns	29863472292 ns	1.01
latency/ttfp	1682623583.5 ns	1656507125 ns	1.02
metal/synchronization/context	814.3203883495146 ns	835.6741573033707 ns	0.97
metal/synchronization/stream	429.02010050251255 ns	439.8131313131313 ns	0.98

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