[Do not merge] Test KernelIntrinsics

[christiangnrd]

Not a draft to also run benchmarks

[github-actions]

Your PR requires formatting changes to meet the project's style guidelines.
Please consider running Runic (git runic main) to apply these changes.

Click here to view the suggested changes.

diff --git a/src/MetalKernels.jl b/src/MetalKernels.jl
index 4e856194..7573c5e1 100644
--- a/src/MetalKernels.jl
+++ b/src/MetalKernels.jl
@@ -136,26 +136,26 @@ end
 
 KI.argconvert(::MetalBackend, arg) = mtlconvert(arg)
 
-function KI.kernel_function(::MetalBackend, f::F, tt::TT=Tuple{}; name=nothing, kwargs...) where {F,TT}
+function KI.kernel_function(::MetalBackend, f::F, tt::TT = Tuple{}; name = nothing, kwargs...) where {F, TT}
     kern = mtlfunction(f, tt; name, kwargs...)
-    KI.Kernel{MetalBackend, typeof(kern)}(MetalBackend(), kern)
+    return KI.Kernel{MetalBackend, typeof(kern)}(MetalBackend(), kern)
 end
 
-function (obj::KI.Kernel{MetalBackend})(args...; numworkgroups=1, workgroupsize=1)
+function (obj::KI.Kernel{MetalBackend})(args...; numworkgroups = 1, workgroupsize = 1)
     KI.check_launch_args(numworkgroups, workgroupsize)
 
-    obj.kern(args...; threads=workgroupsize, groups=numworkgroups)
+    return obj.kern(args...; threads = workgroupsize, groups = numworkgroups)
 end
 
 
-function KI.kernel_max_work_group_size(kikern::KI.Kernel{<:MetalBackend}; max_work_items::Int=typemax(Int))::Int
-    Int(min(kikern.kern.pipeline.maxTotalThreadsPerThreadgroup, max_work_items))
+function KI.kernel_max_work_group_size(kikern::KI.Kernel{<:MetalBackend}; max_work_items::Int = typemax(Int))::Int
+    return Int(min(kikern.kern.pipeline.maxTotalThreadsPerThreadgroup, max_work_items))
 end
 function KI.max_work_group_size(::MetalBackend)::Int
-    Int(device().maxThreadsPerThreadgroup.width)
+    return Int(device().maxThreadsPerThreadgroup.width)
 end
 function KI.multiprocessor_count(::MetalBackend)::Int
-    Metal.num_gpu_cores()
+    return Metal.num_gpu_cores()
 end
 
 
diff --git a/src/broadcast.jl b/src/broadcast.jl
index 5d107ec2..3455fad2 100644
--- a/src/broadcast.jl
+++ b/src/broadcast.jl
@@ -66,9 +66,9 @@ end
     if _broadcast_shapes[Is] > BROADCAST_SPECIALIZATION_THRESHOLD
         ## COV_EXCL_START
         function broadcast_cartesian_static(dest, bc, Is)
-             i = KI.get_global_id().x
-             stride = KI.get_global_size().x
-             while 1 <= i <= length(dest)
+            i = KI.get_global_id().x
+            stride = KI.get_global_size().x
+            while 1 <= i <= length(dest)
                 I = @inbounds Is[i]
                 @inbounds dest[I] = bc[I]
                 i += stride
@@ -91,13 +91,13 @@ end
        (isa(IndexStyle(dest), IndexLinear) && isa(IndexStyle(bc), IndexLinear))
         ## COV_EXCL_START
         function broadcast_linear(dest, bc)
-             i = KI.get_global_id().x
-             stride = KI.get_global_size().x
-             while 1 <= i <= length(dest)
-                 @inbounds dest[i] = bc[i]
-                 i += stride
-             end
-             return
+            i = KI.get_global_id().x
+            stride = KI.get_global_size().x
+            while 1 <= i <= length(dest)
+                @inbounds dest[i] = bc[i]
+                i += stride
+            end
+            return
         end
         ## COV_EXCL_STOP
 
@@ -168,9 +168,9 @@ end
     else
         ## COV_EXCL_START
         function broadcast_cartesian(dest, bc)
-             i = KI.get_global_id().x
-             stride = KI.get_global_size().x
-             while 1 <= i <= length(dest)
+            i = KI.get_global_id().x
+            stride = KI.get_global_size().x
+            while 1 <= i <= length(dest)
                 I = @inbounds CartesianIndices(dest)[i]
                 @inbounds dest[I] = bc[I]
                 i += stride
diff --git a/src/device/random.jl b/src/device/random.jl
index 12b053a2..edc999cd 100644
--- a/src/device/random.jl
+++ b/src/device/random.jl
@@ -89,8 +89,8 @@ end
         @inbounds global_random_counters()[simdgroupId]
     elseif field === :ctr2
         globalId = KI.get_global_id().x +
-                   (KI.get_global_id().y - 1i32) * KI.get_global_size().x +
-                   (KI.get_global_id().z - 1i32) * KI.get_global_size().x * KI.get_global_size().y
+            (KI.get_global_id().y - 1i32) * KI.get_global_size().x +
+            (KI.get_global_id().z - 1i32) * KI.get_global_size().x * KI.get_global_size().y
         globalId % UInt32
     end::UInt32
 end
diff --git a/src/mapreduce.jl b/src/mapreduce.jl
index 7be5ef43..a737e8d0 100644
--- a/src/mapreduce.jl
+++ b/src/mapreduce.jl
@@ -224,7 +224,8 @@ function GPUArrays.mapreducedim!(f::F, op::OP, R::WrappedMtlArray{T},
     # we might not be able to launch all those threads to reduce each slice in one go.
     # that's why each threads also loops across their inputs, processing multiple values
     # so that we can span the entire reduction dimension using a single item group.
-    kernel = KI.@kernel backend launch = false partial_mapreduce_device(f, op, init, Val(maxthreads), Val(Rreduce), Val(Rother),
+    kernel = KI.@kernel backend launch = false partial_mapreduce_device(
+        f, op, init, Val(maxthreads), Val(Rreduce), Val(Rother),
                                                           Val(UInt64(length(Rother))), Val(grain), Val(shuffle), R, A)
 
     # how many threads do we want?
@@ -260,7 +261,8 @@ function GPUArrays.mapreducedim!(f::F, op::OP, R::WrappedMtlArray{T},
         # we can cover the dimensions to reduce using a single group
         kernel(f, op, init, Val(maxthreads), Val(Rreduce), Val(Rother),
                Val(UInt64(length(Rother))), Val(grain), Val(shuffle), R, A;
-               workgroupsize = threads, numworkgroups = groups)
+            workgroupsize = threads, numworkgroups = groups
+        )
     else
         # temporary empty array whose type will match the final partial array
 	    partial = similar(R, ntuple(_ -> 0, Val(ndims(R)+1)))
@@ -287,7 +289,8 @@ function GPUArrays.mapreducedim!(f::F, op::OP, R::WrappedMtlArray{T},
         partial_kernel(f, op, init, Val(threads), Val(Rreduce),
                         Val(Rother), Val(UInt64(length(Rother))),
                         Val(grain), Val(shuffle), partial, A;
-                        numworkgroups = partial_groups, workgroupsize = partial_threads)
+            numworkgroups = partial_groups, workgroupsize = partial_threads
+        )
 
         GPUArrays.mapreducedim!(identity, op, R, partial; init=init)
     end
diff --git a/test/kernelabstractions.jl b/test/kernelabstractions.jl
index cda5b249..339fcbc8 100644
--- a/test/kernelabstractions.jl
+++ b/test/kernelabstractions.jl
@@ -7,6 +7,6 @@ Testsuite.testsuite(MetalBackend, "Metal", Metal, MtlArray, Metal.MtlDeviceArray
     "Convert",           # depends on https://github.com/JuliaGPU/Metal.jl/issues/69
     "SpecialFunctions",  # gamma and erfc not currently supported on Metal.jl
     "sparse",            # not supported yet
-    "CPU synchronization",
-    "fallback test: callable types",
+            "CPU synchronization",
+            "fallback test: callable types",
 ]))
diff --git a/test/runtests.jl b/test/runtests.jl
index 32b45c8c..14fcfb93 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -1,6 +1,6 @@
 @static if VERSION < v"1.11" && get(ENV, "BUILDKITE_PIPELINE_NAME", "Metal.jl") == "Metal.jl"
     using Pkg
-    Pkg.add(url="https://github.com/JuliaGPU/KernelAbstractions.jl", rev="main")
+    Pkg.add(url = "https://github.com/JuliaGPU/KernelAbstractions.jl", rev = "main")
 end
 
 using Metal

[github-actions]

Metal Benchmarks

Details

Benchmark suite	Current: 865af1a	Previous: d1a8cc2	Ratio
latency/precompile	30229762937 ns	26336492125 ns	1.15
latency/ttfp	2505249124.5 ns	2442892125 ns	1.03
latency/import	1574101104 ns	1495421459 ns	1.05
integration/metaldevrt	886458.5 ns	518291 ns	1.71
integration/byval/slices=1	1573250 ns	1190833 ns	1.32
integration/byval/slices=3	9320833.5 ns	8102291 ns	1.15
integration/byval/reference	1574125 ns	1189208 ns	1.32
integration/byval/slices=2	2657834 ns	2114958 ns	1.26
kernel/indexing	658250 ns	260375 ns	2.53
kernel/indexing_checked	506416.5 ns	272167 ns	1.86
kernel/launch	13459 ns	12500 ns	1.08
kernel/rand	575291 ns	292458 ns	1.97
array/construct	6958 ns	6709 ns	1.04
array/broadcast	607959 ns	284000 ns	2.14
array/random/randn/Float32	1022000 ns	493667 ns	2.07
array/random/randn!/Float32	755250 ns	418250 ns	1.81
array/random/rand!/Int64	547459 ns	314416.5 ns	1.74
array/random/rand!/Float32	586583 ns	275792 ns	2.13
array/random/rand/Int64	803813 ns	462958 ns	1.74
array/random/rand/Float32	638104 ns	349625 ns	1.83
array/accumulate/Int64/1d	1335374.5 ns	970583 ns	1.38
array/accumulate/Int64/dims=1	1979584 ns	1042125 ns	1.90
array/accumulate/Int64/dims=2	2305000 ns	1363688 ns	1.69
array/accumulate/Int64/dims=1L	11904000 ns	9571750 ns	1.24
array/accumulate/Int64/dims=2L	9977625 ns	7945646 ns	1.26
array/accumulate/Float32/1d	1159875 ns	808292 ns	1.43
array/accumulate/Float32/dims=1	1789563 ns	904895.5 ns	1.98
array/accumulate/Float32/dims=2	2071542 ns	1184584 ns	1.75
array/accumulate/Float32/dims=1L	10015875 ns	8567583 ns	1.17
array/accumulate/Float32/dims=2L	8151125 ns	4391750 ns	1.86
array/reductions/reduce/Int64/1d	1352333 ns	744625 ns	1.82
array/reductions/reduce/Int64/dims=1	1141395.5 ns	685541 ns	1.66
array/reductions/reduce/Int64/dims=2	1342541 ns	722666 ns	1.86
array/reductions/reduce/Int64/dims=1L	2051229 ns	1164834 ns	1.76
array/reductions/reduce/Int64/dims=2L	4290209 ns	2224833 ns	1.93
array/reductions/reduce/Float32/1d	1062542 ns	558917 ns	1.90
array/reductions/reduce/Float32/dims=1	853834 ns	400208 ns	2.13
array/reductions/reduce/Float32/dims=2	895250 ns	430792 ns	2.08
array/reductions/reduce/Float32/dims=1L	1355375 ns	675854.5 ns	2.01
array/reductions/reduce/Float32/dims=2L	1905833.5 ns	1244709 ns	1.53
array/reductions/mapreduce/Int64/1d	1353000 ns	749417 ns	1.81
array/reductions/mapreduce/Int64/dims=1	1095417 ns	687167 ns	1.59
array/reductions/mapreduce/Int64/dims=2	1331625 ns	723333 ns	1.84
array/reductions/mapreduce/Int64/dims=1L	2071542 ns	1138000 ns	1.82
array/reductions/mapreduce/Int64/dims=2L	4315458 ns	1854834 ns	2.33
array/reductions/mapreduce/Float32/1d	1065896 ns	561291 ns	1.90
array/reductions/mapreduce/Float32/dims=1	863895.5 ns	400834 ns	2.16
array/reductions/mapreduce/Float32/dims=2	910354.5 ns	431458 ns	2.11
array/reductions/mapreduce/Float32/dims=1L	1345937.5 ns	690812.5 ns	1.95
array/reductions/mapreduce/Float32/dims=2L	1888062 ns	1251208 ns	1.51
array/private/copyto!/gpu_to_gpu	636084 ns	231917 ns	2.74
array/private/copyto!/cpu_to_gpu	801750 ns	255125 ns	3.14
array/private/copyto!/gpu_to_cpu	803958.5 ns	254833 ns	3.15
array/private/iteration/findall/int	1666354 ns	1211667 ns	1.38
array/private/iteration/findall/bool	1487104 ns	1071937.5 ns	1.39
array/private/iteration/findfirst/int	2136437 ns	1192604.5 ns	1.79
array/private/iteration/findfirst/bool	2116354.5 ns	1183042 ns	1.79
array/private/iteration/scalar	5696459 ns	1706917 ns	3.34
array/private/iteration/logical	2725771 ns	1644459 ns	1.66
array/private/iteration/findmin/1d	2597833.5 ns	1415000 ns	1.84
array/private/iteration/findmin/2d	1874312.5 ns	1220583 ns	1.54
array/private/copy	571958 ns	326542 ns	1.75
array/shared/copyto!/gpu_to_gpu	84541 ns	78584 ns	1.08
array/shared/copyto!/cpu_to_gpu	79916 ns	79729.5 ns	1.00
array/shared/copyto!/gpu_to_cpu	82375 ns	77083 ns	1.07
array/shared/iteration/findall/int	1666375 ns	1213417 ns	1.37
array/shared/iteration/findall/bool	1507416.5 ns	1077291 ns	1.40
array/shared/iteration/findfirst/int	1703958 ns	1003500 ns	1.70
array/shared/iteration/findfirst/bool	1721333 ns	995583 ns	1.73
array/shared/iteration/scalar	211125 ns	184375 ns	1.15
array/shared/iteration/logical	2653625.5 ns	1442938 ns	1.84
array/shared/iteration/findmin/1d	2217666.5 ns	1224583.5 ns	1.81
array/shared/iteration/findmin/2d	1871667 ns	1221208 ns	1.53
array/shared/copy	247000 ns	235833 ns	1.05
array/permutedims/4d	2670917 ns	1723000 ns	1.55
array/permutedims/2d	1167813 ns	558541 ns	2.09
array/permutedims/3d	1698042 ns	1119959 ns	1.52
metal/synchronization/stream	19500 ns	16750 ns	1.16
metal/synchronization/context	20417 ns	17541 ns	1.16

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