Harden intrinsic compile-time validation

[maleadt]

No description provided.

[github-actions]

CUDA.jl Benchmarks

Details

Benchmark suite	Current: 7ff74bf	Previous: 069cdff	Ratio
array/accumulate/Float32/1d	97377 ns	98074 ns	0.99
array/accumulate/Float32/dims=1	70765 ns	74919 ns	0.94
array/accumulate/Float32/dims=1L	1598543 ns	1600002 ns	1.00
array/accumulate/Float32/dims=2	136147 ns	140367 ns	0.97
array/accumulate/Float32/dims=2L	657859 ns	659745 ns	1.00
array/accumulate/Int64/1d	116864 ns	117818 ns	0.99
array/accumulate/Int64/dims=1	74675 ns	79251 ns	0.94
array/accumulate/Int64/dims=1L	1715389 ns	1717606 ns	1.00
array/accumulate/Int64/dims=2	146884 ns	153507 ns	0.96
array/accumulate/Int64/dims=2L	986104 ns	986800 ns	1.00
array/broadcast	17420 ns	18060 ns	0.96
array/broadcast launch	8094.666666666667 ns
array/construct	888.7916666666666 ns	869.1052631578947 ns	1.02
array/copy	16188 ns	15967 ns	1.01
array/copyto!/cpu_to_gpu	206376 ns	208599 ns	0.99
array/copyto!/gpu_to_cpu	240029 ns	241280 ns	0.99
array/copyto!/gpu_to_gpu	8681.333333333334 ns	9209.333333333334 ns	0.94
array/iteration/findall/bool	130551 ns	132021 ns	0.99
array/iteration/findall/int	144064 ns	145286 ns	0.99
array/iteration/findfirst/bool	67599 ns	67274 ns	1.00
array/iteration/findfirst/int	68729 ns	68529 ns	1.00
array/iteration/findmin/1d	58915 ns	64309 ns	0.92
array/iteration/findmin/2d	98691 ns	99940 ns	0.99
array/iteration/logical	181155 ns	185815 ns	0.97
array/iteration/scalar	60535 ns	63136 ns	0.96
array/permutedims/2d	47344 ns	48431 ns	0.98
array/permutedims/3d	48641 ns	50026 ns	0.97
array/permutedims/4d	48921 ns	49849 ns	0.98
array/random/rand/Float32	11774 ns	11669 ns	1.01
array/random/rand/Int64	22116 ns	22883 ns	0.97
array/random/rand!/Float32	7589 ns	7731.75 ns	0.98
array/random/rand!/Int64	19663 ns	20210 ns	0.97
array/random/randn/Float32	32341 ns	32574 ns	0.99
array/random/randn!/Float32	24107 ns	23189 ns	1.04
array/reductions/mapreduce/Float32/1d	30985 ns	31371 ns	0.99
array/reductions/mapreduce/Float32/dims=1	36964 ns	37491 ns	0.99
array/reductions/mapreduce/Float32/dims=1L	49830 ns	49977 ns	1.00
array/reductions/mapreduce/Float32/dims=2	54080 ns	54508 ns	0.99
array/reductions/mapreduce/Float32/dims=2L	65782 ns	66329 ns	0.99
array/reductions/mapreduce/Int64/1d	37774 ns	37643 ns	1.00
array/reductions/mapreduce/Int64/dims=1	39752 ns	39961 ns	0.99
array/reductions/mapreduce/Int64/dims=1L	87904 ns	87917 ns	1.00
array/reductions/mapreduce/Int64/dims=2	56658 ns	57047 ns	0.99
array/reductions/mapreduce/Int64/dims=2L	82234 ns	82751 ns	0.99
array/reductions/reduce/Float32/1d	31216 ns	31482 ns	0.99
array/reductions/reduce/Float32/dims=1	37048 ns	37630 ns	0.98
array/reductions/reduce/Float32/dims=1L	49811 ns	50073 ns	0.99
array/reductions/reduce/Float32/dims=2	54199 ns	54518 ns	0.99
array/reductions/reduce/Float32/dims=2L	67403 ns	67963 ns	0.99
array/reductions/reduce/Int64/1d	37958 ns	38054 ns	1.00
array/reductions/reduce/Int64/dims=1	39663 ns	39983 ns	0.99
array/reductions/reduce/Int64/dims=1L	87774 ns	87988 ns	1.00
array/reductions/reduce/Int64/dims=2	56605 ns	56860 ns	1.00
array/reductions/reduce/Int64/dims=2L	82201 ns	82191 ns	1.00
array/reverse/1d	16240 ns	16098 ns	1.01
array/reverse/1dL	69003 ns	68844 ns	1.00
array/reverse/1dL_inplace	67229 ns	66918 ns	1.00
array/reverse/1d_inplace	10096 ns	9757 ns	1.03
array/reverse/2d	18984 ns	19108 ns	0.99
array/reverse/2dL	72288 ns	72561 ns	1.00
array/reverse/2dL_inplace	66818 ns	66785 ns	1.00
array/reverse/2d_inplace	9659 ns	9397 ns	1.03
array/sorting/1d	2639299 ns	2656939 ns	0.99
array/sorting/2d	1027118 ns	1038211 ns	0.99
array/sorting/by	3182201 ns	3192363 ns	1.00
cuda/synchronization/context/auto	1005.6 ns	1015.3 ns	0.99
cuda/synchronization/context/blocking	777.1979166666666 ns	784.6960784313726 ns	0.99
cuda/synchronization/context/nonblocking	5615.666666666667 ns	5655.166666666667 ns	0.99
cuda/synchronization/stream/auto	852.8032786885246 ns	857.3174603174604 ns	0.99
cuda/synchronization/stream/blocking	663.6792452830189 ns	654.7730061349694 ns	1.01
cuda/synchronization/stream/nonblocking	5493.5 ns	5491 ns	1.00
integration/byval/reference	147365 ns	147245 ns	1.00
integration/byval/slices=1	151324 ns	149267 ns	1.01
integration/byval/slices=2	294603 ns	292191 ns	1.01
integration/byval/slices=3	437511 ns	434868 ns	1.01
integration/cudadevrt	104557 ns	104357 ns	1.00
integration/volumerhs	9137644 ns	9315717 ns	0.98
kernel/indexing	12436 ns	12183 ns	1.02
kernel/indexing_checked	13280 ns	13078 ns	1.02
kernel/launch	1939.3 ns	2032.3333333333333 ns	0.95
kernel/occupancy	668.304347826087 ns	667.8466257668712 ns	1.00
kernel/rand	13531 ns	14903 ns	0.91
latency/import	4091768416 ns	4116528672 ns	0.99
latency/precompile	4794097643 ns	4814427609 ns	1.00
latency/ttfp	5137544706 ns	5198634459 ns	0.99

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

[kshyatt]

(Rebased to run this on top of the Crayons fix)