Adapt to LLVMDowngrader_jll 0.11 and GPUCompiler 2.8

[maleadt]

Bumps the compat bounds to LLVMDowngrader_jll 0.11 (JuliaPackaging/Yggdrasil#14758, built against LLVM 23) and GPUCompiler 2.8 (JuliaGPU/GPUCompiler.jl#931). The downgrader C API is unchanged, so this is compat-only.

[github-actions]

AMDGPU.jl Benchmarks

Details

Benchmark suite	Current: a5715fe	Previous: 63c6c40	Ratio
amdgpu/synchronization/context/device	552.5 ns	582.5 ns	0.95
amdgpu/synchronization/stream/blocking	242.5 ns	235 ns	1.03
amdgpu/synchronization/stream/nonblocking	317.5 ns	342.5 ns	0.93
applications/bitonic_sort	1123402 ns	1135780.75 ns	0.99
applications/convolution	105069.25 ns	105389 ns	1.00
applications/floyd_warshall	8872969 ns	8837117.5 ns	1.00
applications/histogram	813468 ns	803238.75 ns	1.01
applications/prefix_sum	232134.25 ns	231228.25 ns	1.00
array/accumulate/Float32/1d	79026 ns	79176 ns	1.00
array/accumulate/Float32/dims=1	275636.25 ns	270856.25 ns	1.02
array/accumulate/Float32/dims=1L	80396 ns	82303.75 ns	0.98
array/accumulate/Float32/dims=2	70473.5 ns	70921 ns	0.99
array/accumulate/Float32/dims=2L	2754078.25 ns	2753465.25 ns	1.00
array/accumulate/Int64/1d	66423.25 ns	77441 ns	0.86
array/accumulate/Int64/dims=1	241750.5 ns	242275.75 ns	1.00
array/accumulate/Int64/dims=1L	84071 ns	83723.5 ns	1.00
array/accumulate/Int64/dims=2	83831 ns	85516.25 ns	0.98
array/accumulate/Int64/dims=2L	2890907.5 ns	2887214.75 ns	1.00
array/broadcast	72871 ns	72326 ns	1.01
array/construct	2535 ns	2327.5 ns	1.09
array/copy	37630.5 ns	33308 ns	1.13
array/copyto!/cpu_to_gpu	110621.25 ns	111544 ns	0.99
array/copyto!/gpu_to_cpu	111288.75 ns	112239 ns	0.99
array/copyto!/gpu_to_gpu	57608.25 ns	53915.75 ns	1.07
array/iteration/findall/bool	133251.75 ns	134521.75 ns	0.99
array/iteration/findall/int	142029.25 ns	148569.5 ns	0.96
array/iteration/findfirst/bool	182454.75 ns	182537.5 ns	1.00
array/iteration/findfirst/int	142182 ns	140624.25 ns	1.01
array/iteration/findmin/1d	106089 ns	118906.5 ns	0.89
array/iteration/findmin/2d	107011.25 ns	109151.5 ns	0.98
array/iteration/logical	240828.25 ns	238428.25 ns	1.01
array/iteration/scalar	294819 ns	292276.25 ns	1.01
array/permutedims/2d	71163.5 ns	70986 ns	1.00
array/permutedims/3d	58660.75 ns	70406 ns	0.83
array/permutedims/4d	73286 ns	60245.75 ns	1.22
array/random/rand/Float32	45980.5 ns	45383 ns	1.01
array/random/rand/Int64	53388.25 ns	53550.75 ns	1.00
array/random/rand!/Float32	64843.25 ns	64251 ns	1.01
array/random/rand!/Int64	72350.75 ns	71186 ns	1.02
array/random/randn/Float32	78948.5 ns	79301 ns	1.00
array/random/randn!/Float32	81088.5 ns	80246.25 ns	1.01
array/reductions/mapreduce/Float32/1d	91298.75 ns	94648.75 ns	0.96
array/reductions/mapreduce/Float32/dims=1	79958.5 ns	86616.25 ns	0.92
array/reductions/mapreduce/Float32/dims=1L	838178.5 ns	835469 ns	1.00
array/reductions/mapreduce/Float32/dims=2	81486.25 ns	82718.75 ns	0.99
array/reductions/mapreduce/Float32/dims=2L	137251.75 ns	138384.5 ns	0.99
array/reductions/mapreduce/Int64/1d	91048.75 ns	94203.75 ns	0.97
array/reductions/mapreduce/Int64/dims=1	79873.5 ns	86458.75 ns	0.92
array/reductions/mapreduce/Int64/dims=1L	838553.5 ns	841119 ns	1.00
array/reductions/mapreduce/Int64/dims=2	81236 ns	83068.5 ns	0.98
array/reductions/mapreduce/Int64/dims=2L	139326.75 ns	138761.75 ns	1.00
array/reductions/reduce/Float32/1d	91071.25 ns	94551.5 ns	0.96
array/reductions/reduce/Float32/dims=1	79966 ns	86223.75 ns	0.93
array/reductions/reduce/Float32/dims=1L	838870.75 ns	841294 ns	1.00
array/reductions/reduce/Float32/dims=2	80641.25 ns	82688.5 ns	0.98
array/reductions/reduce/Float32/dims=2L	138321.75 ns	137874.25 ns	1.00
array/reductions/reduce/Int64/1d	91186.25 ns	94288.75 ns	0.97
array/reductions/reduce/Int64/dims=1	89918.75 ns	85981.25 ns	1.05
array/reductions/reduce/Int64/dims=1L	840936 ns	840701.25 ns	1.00
array/reductions/reduce/Int64/dims=2	80818.5 ns	83251.25 ns	0.97
array/reductions/reduce/Int64/dims=2L	139501.75 ns	138074.5 ns	1.01
array/reverse/1d	36653 ns	43553 ns	0.84
array/reverse/1dL	72510.75 ns	73793.5 ns	0.98
array/reverse/1dL_inplace	78601 ns	78213.5 ns	1.00
array/reverse/1d_inplace	59550.75 ns	41343 ns	1.44
array/reverse/2d	49355.75 ns	49133 ns	1.00
array/reverse/2dL	81271 ns	84381 ns	0.96
array/reverse/2dL_inplace	89656.25 ns	89121 ns	1.01
array/reverse/2d_inplace	61480.75 ns	61473.25 ns	1.00
array/sorting/1d	330829.25 ns	335059.75 ns	0.99
gemm/tiled	1882594.25 ns	1903771.5 ns	0.99
gemm/tiled_unbounded	1859131.5 ns	1922259 ns	0.97
integration/byval/reference	39591 ns	39361 ns	1.01
integration/byval/slices=1	41291 ns	41280 ns	1.00
integration/byval/slices=2	155423 ns	158962 ns	0.98
integration/byval/slices=3	237574 ns	237943 ns	1.00
integration/volumerhs	4906710 ns	4900479 ns	1.00
kernel/indexing	56153.25 ns	55808.25 ns	1.01
kernel/indexing_checked	57553.25 ns	38333 ns	1.50
kernel/launch	1297.5 ns	1337.5 ns	0.97
kernel/rand	99923.75 ns	99514 ns	1.00
latency/import	1754761451 ns	1748092483 ns	1.00
latency/precompile	39908224005 ns	39846434991 ns	1.00
latency/ttfp	2329367497 ns	2322353398 ns	1.00
stencil/diffusion3d	1619181.25 ns	1620301.5 ns	1.00
stencil/diffusion3d_checked	1655336.5 ns	1658012.25 ns	1.00

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

[christiangnrd]

Close #1080