add fastmath flag

[vchuravy]

No description provided.

[github-actions]

CUDA.jl Benchmarks

Benchmark suite	Current: 6c0345c	Previous: 57e06f9	Ratio
latency/precompile	45661499477.5 ns	45680313076.5 ns	1.00
latency/ttfp	6644709519 ns	6618514052 ns	1.00
latency/import	3165546495.5 ns	3159190068.5 ns	1.00
integration/volumerhs	9623672.5 ns	9615523 ns	1.00
integration/byval/slices=1	147339 ns	147042 ns	1.00
integration/byval/slices=3	425644 ns	425857 ns	1.00
integration/byval/reference	145349.5 ns	145233.5 ns	1.00
integration/byval/slices=2	286663 ns	286656.5 ns	1.00
integration/cudadevrt	103791 ns	103621 ns	1.00
kernel/indexing	14663 ns	14335 ns	1.02
kernel/indexing_checked	15282 ns	15041 ns	1.02
kernel/occupancy	674.2594936708861 ns	672.879746835443 ns	1.00
kernel/launch	2313.6666666666665 ns	2180 ns	1.06
kernel/rand	17361 ns	15098.5 ns	1.15
array/reverse/1d	20002 ns	20194 ns	0.99
array/reverse/2d	25387 ns	24780 ns	1.02
array/reverse/1d_inplace	11562 ns	10893 ns	1.06
array/reverse/2d_inplace	13306 ns	13342 ns	1.00
array/copy	20937 ns	21420 ns	0.98
array/iteration/findall/int	159683 ns	161487 ns	0.99
array/iteration/findall/bool	139460.5 ns	141184 ns	0.99
array/iteration/findfirst/int	154725 ns	155481.5 ns	1.00
array/iteration/findfirst/bool	155802 ns	156277 ns	1.00
array/iteration/scalar	73427 ns	74379 ns	0.99
array/iteration/logical	218733.5 ns	221595 ns	0.99
array/iteration/findmin/1d	42194 ns	42443 ns	0.99
array/iteration/findmin/2d	95047 ns	95456.5 ns	1.00
array/reductions/reduce/1d	43278.5 ns	37041 ns	1.17
array/reductions/reduce/2d	48985.5 ns	51842 ns	0.94
array/reductions/mapreduce/1d	41489 ns	35370 ns	1.17
array/reductions/mapreduce/2d	42242.5 ns	41509 ns	1.02
array/broadcast	21528 ns	21339.5 ns	1.01
array/copyto!/gpu_to_gpu	13762 ns	13807 ns	1.00
array/copyto!/cpu_to_gpu	209543 ns	211351 ns	0.99
array/copyto!/gpu_to_cpu	243591 ns	245693 ns	0.99
array/accumulate/1d	109500 ns	110163 ns	0.99
array/accumulate/2d	82231.5 ns	81026 ns	1.01
array/construct	1259.6 ns	1272.2 ns	0.99
array/random/randn/Float32	45832.5 ns	45868 ns	1.00
array/random/randn!/Float32	26472 ns	27161 ns	0.97
array/random/rand!/Int64	27426 ns	27291 ns	1.00
array/random/rand!/Float32	8798.666666666666 ns	8809.666666666666 ns	1.00
array/random/rand/Int64	30177 ns	30432 ns	0.99
array/random/rand/Float32	13263 ns	13379 ns	0.99
array/permutedims/4d	61834 ns	62223 ns	0.99
array/permutedims/2d	56175 ns	56419 ns	1.00
array/permutedims/3d	57143.5 ns	57054.5 ns	1.00
array/sorting/1d	2777303 ns	2767302 ns	1.00
array/sorting/by	3368494.5 ns	3355332 ns	1.00
array/sorting/2d	1085912 ns	1083060 ns	1.00
cuda/synchronization/stream/auto	1026.6 ns	1096.1 ns	0.94
cuda/synchronization/stream/nonblocking	6544.6 ns	6533.2 ns	1.00
cuda/synchronization/stream/blocking	800.8686868686868 ns	865.6755319148936 ns	0.93
cuda/synchronization/context/auto	1166.1 ns	1242.7 ns	0.94
cuda/synchronization/context/nonblocking	6772.4 ns	6746.8 ns	1.00
cuda/synchronization/context/blocking	948.25 ns	983.9 ns	0.96

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