Julia 1.11 Unsupported #1358
Comments
|
We need an update to the new GPUCompiler version |
wsmoses
changed the title
|
I'm getting the following error here ERROR: LoadError: UndefVarError: `PassBuilder` not defined in `Enzyme.Compiler`
Stacktrace:
[1] macro expansion
@ ~/.julia/packages/LLVM/5DlHM/src/base.jl:96 [inlined]
[2] (::Enzyme.Compiler.var"#prop_julia_addr#28416"{LLVM.TargetMachine})(f::LLVM.Function)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Pljwm/src/compiler/optimize.jl:75
[3] function_pass_callback(ptr::Ptr{Nothing}, data::Ptr{Nothing})
@ LLVM ~/.julia/packages/LLVM/5DlHM/src/pass.jl:49
[4] LLVMRunPassManager
@ ~/.julia/packages/LLVM/5DlHM/lib/16/libLLVM.jl:3351 [inlined]
[5] run!
@ ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:39 [inlined]
[6] (::Enzyme.Compiler.var"#28512#28513"{LLVM.Module, LLVM.TargetMachine})(pm::LLVM.ModulePassManager)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Pljwm/src/compiler/optimize.jl:2029
[7] LLVM.ModulePassManager(::Enzyme.Compiler.var"#28512#28513"{LLVM.Module, LLVM.TargetMachine}; kwargs::@Kwargs{})
@ LLVM ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:33
[8] ModulePassManager
@ ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:30 [inlined]
[9] optimize!
@ ~/.julia/packages/Enzyme/Pljwm/src/compiler/optimize.jl:1951 [inlined]
[10] codegen(output::Symbol, job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}; libraries::Bool, deferred_codegen::Bool, optimize::Bool, toplevel::Bool, strip::Bool, validate::Bool, only_entry::Bool, parent_job::Nothing)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:5787
[11] codegen
@ ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:5194 [inlined]
[12] _thunk(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, postopt::Bool)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:6682
[13] _thunk
@ ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:6682 [inlined]
[14] cached_compilation
@ ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:6720 [inlined]
[15] (::Enzyme.Compiler.var"#28633#28634"{Active, FFIABI, Const{typeof(loss_function)}, Enzyme.API.DEM_ReverseModeCombined, (false, false, false, false, false, false), true, false, Tuple{Const{Lux.Chain{@NamedTuple{layer_1::Lux.Dense{true, typeof(tanh_fast), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_2::Lux.Dense{true, typeof(identity), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_3::WrappedFunction{:direct_call, typeof(softmax)}}, Nothing}}, Const{Matrix{Float32}}, Const{OneHotMatrix{UInt32, Vector{UInt32}}}, Duplicated{@NamedTuple{layer_1::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_2::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_3::@NamedTuple{}}}, Const{@NamedTuple{layer_1::@NamedTuple{}, layer_2::@NamedTuple{}, layer_3::@NamedTuple{}}}}, 0x00000000000068d4, 1, Core.MethodInstance})(ctx::LLVM.Context)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:6795
[16] JuliaContext(f::Enzyme.Compiler.var"#28633#28634"{Active, FFIABI, Const{typeof(loss_function)}, Enzyme.API.DEM_ReverseModeCombined, (false, false, false, false, false, false), true, false, Tuple{Const{Lux.Chain{@NamedTuple{layer_1::Lux.Dense{true, typeof(tanh_fast), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_2::Lux.Dense{true, typeof(identity), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_3::WrappedFunction{:direct_call, typeof(softmax)}}, Nothing}}, Const{Matrix{Float32}}, Const{OneHotMatrix{UInt32, Vector{UInt32}}}, Duplicated{@NamedTuple{layer_1::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_2::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_3::@NamedTuple{}}}, Const{@NamedTuple{layer_1::@NamedTuple{}, layer_2::@NamedTuple{}, layer_3::@NamedTuple{}}}}, 0x00000000000068d4, 1, Core.MethodInstance}; kwargs::@Kwargs{})
@ GPUCompiler ~/.julia/packages/GPUCompiler/Y4hSX/src/driver.jl:52
[17] JuliaContext
@ ~/.julia/packages/GPUCompiler/Y4hSX/src/driver.jl:42 [inlined]
[18] thunkbase(mi::Core.MethodInstance, ::Val{0x00000000000068d4}, ::Type{Const{typeof(loss_function)}}, ::Type{Active}, tt::Type{Tuple{Const{Lux.Chain{@NamedTuple{layer_1::Lux.Dense{true, typeof(tanh_fast), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_2::Lux.Dense{true, typeof(identity), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_3::WrappedFunction{:direct_call, typeof(softmax)}}, Nothing}}, Const{Matrix{Float32}}, Const{OneHotMatrix{UInt32, Vector{UInt32}}}, Duplicated{@NamedTuple{layer_1::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_2::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_3::@NamedTuple{}}}, Const{@NamedTuple{layer_1::@NamedTuple{}, layer_2::@NamedTuple{}, layer_3::@NamedTuple{}}}}}, ::Val{Enzyme.API.DEM_ReverseModeCombined}, ::Val{1}, ::Val{(false, false, false, false, false, false)}, ::Val{true}, ::Val{false}, ::Type{FFIABI})
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:6740
[19] #s2021#28635
@ ~/.julia/packages/Enzyme/Pljwm/src/compiler.jl:6826 [inlined]
[20] var"#s2021#28635"(FA::Any, A::Any, TT::Any, Mode::Any, ModifiedBetween::Any, width::Any, ReturnPrimal::Any, ShadowInit::Any, World::Any, ABI::Any, ::Any, ::Any, ::Any, ::Any, tt::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any)
@ Enzyme.Compiler ./none:0
[21] (::Core.GeneratedFunctionStub)(::UInt64, ::LineNumberNode, ::Any, ::Vararg{Any})
@ Core ./boot.jl:709
[22] autodiff
@ ~/.julia/packages/Enzyme/Pljwm/src/Enzyme.jl:309 [inlined]
[23] autodiff
@ ~/.julia/packages/Enzyme/Pljwm/src/Enzyme.jl:326 [inlined]
[24] gradient_loss_function(model::Lux.Chain{@NamedTuple{layer_1::Lux.Dense{true, typeof(tanh_fast), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_2::Lux.Dense{true, typeof(identity), typeof(glorot_uniform), typeof(WeightInitializers.zeros32)}, layer_3::WrappedFunction{:direct_call, typeof(softmax)}}, Nothing}, x::Matrix{Float32}, y::OneHotMatrix{UInt32, Vector{UInt32}}, ps::@NamedTuple{layer_1::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_2::@NamedTuple{weight::Matrix{Float32}, bias::Matrix{Float32}}, layer_3::@NamedTuple{}}, st::@NamedTuple{layer_1::@NamedTuple{}, layer_2::@NamedTuple{}, layer_3::@NamedTuple{}})
@ Main ~/work/Reactant.jl/Reactant.jl/test/nn_lux.jl:65
[25] top-level scope
@ ~/work/Reactant.jl/Reactant.jl/test/nn_lux.jl:78
[26] include(fname::String)
@ Main ./sysimg.jl:38
[27] top-level scope
@ ~/work/Reactant.jl/Reactant.jl/test/runtests.jl:49
[28] include(fname::String)
@ Main ./sysimg.jl:38
[29] top-level scope
@ none:6
in expression starting at /home/runner/work/Reactant.jl/Reactant.jl/test/nn_lux.jl:78
in expression starting at /home/runner/work/Reactant.jl/Reactant.jl/test/runtests.jl:49
Package Reactant errored during testing |
|
Is this on main? |
I am getting this on |
|
What is your Manifest? Did you re-resolve. This means you ended up with an old version of GPUCompiler most likely. |
|
I think Enzyme has an old GPUCompiler in its compat, which we should likely remove |
|
With using Enzyme
x = rand(Float32, 32);
Enzyme.gradient(Reverse, sum, x)ErrorClosest candidates are:
cpu_features!(::LLVM.Module)
@ GPUCompiler ~/.julia/packages/GPUCompiler/89rev/src/optim.jl:290
Stacktrace:
[1] (::Enzyme.Compiler.var"#28298#28299"{LLVM.Module, LLVM.TargetMachine})(pm::LLVM.ModulePassManager)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler/optimize.jl:1965
[2] LLVM.ModulePassManager(::Enzyme.Compiler.var"#28298#28299"{LLVM.Module, LLVM.TargetMachine}; kwargs::@Kwargs{})
@ LLVM ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:33
[3] ModulePassManager
@ ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:30
[4] optimize!(mod::LLVM.Module, tm::LLVM.TargetMachine)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler/optimize.jl:1951
[5] codegen(output::Symbol, job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}; libraries::Bool, deferred_codegen::Bool, optimize::Bool, toplevel::Bool, strip::Bool, validate::Bool, only_entry::Bool, parent_job::Nothing)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:5807
[6] _thunk(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, postopt::Bool)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6710
[7] cached_compilation
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6748 [inlined]
[8] thunkbase(ctx::LLVM.Context, mi::Core.MethodInstance, ::Val{0x0000000000006819}, ::Type{Const{…}}, ::Type{Active}, tt::Type{Tuple{…}}, ::Val{Enzyme.API.DEM_ReverseModeCombined}, ::Val{1}, ::Val{(false, false)}, ::Val{false}, ::Val{false}, ::Type{FFIABI})
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6821
[9] #s2021#28415
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6873 [inlined]
[10] var"#s2021#28415"(FA::Any, A::Any, TT::Any, Mode::Any, ModifiedBetween::Any, width::Any, ReturnPrimal::Any, ShadowInit::Any, World::Any, ABI::Any, ::Any, ::Any, ::Any, ::Any, tt::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any)
@ Enzyme.Compiler ./none:0
[11] (::Core.GeneratedFunctionStub)(::UInt64, ::LineNumberNode, ::Any, ::Vararg{Any})
@ Core ./boot.jl:709
[12] autodiff
@ ~/.julia/packages/Enzyme/YDcYf/src/Enzyme.jl:309 [inlined]
[13] autodiff
@ ~/.julia/packages/Enzyme/YDcYf/src/Enzyme.jl:326 [inlined]
[14] gradient(rm::ReverseMode{false, FFIABI, false}, f::typeof(sum), x::Vector{Float32})
@ Enzyme ~/.julia/packages/Enzyme/YDcYf/src/Enzyme.jl:1027
[15] top-level scope
@ REPL[3]:1
[16] top-level scope
@ none:1
Some type information was truncated. Use `show(err)` to see complete types.Manifest.toml# This file is machine-generated - editing it directly is not advised
julia_version = "1.11.0-rc1"
manifest_format = "2.0"
project_hash = "9d9865fdd982a60cc61fa85e3720d1f44ecb386c"
[[deps.ArgTools]]
uuid = "0dad84c5-d112-42e6-8d28-ef12dabb789f"
version = "1.1.2"
[[deps.Artifacts]]
uuid = "56f22d72-fd6d-98f1-02f0-08ddc0907c33"
version = "1.11.0"
[[deps.Base64]]
uuid = "2a0f44e3-6c83-55bd-87e4-b1978d98bd5f"
version = "1.11.0"
[[deps.CEnum]]
git-tree-sha1 = "389ad5c84de1ae7cf0e28e381131c98ea87d54fc"
uuid = "fa961155-64e5-5f13-b03f-caf6b980ea82"
version = "0.5.0"
[[deps.CompilerSupportLibraries_jll]]
deps = ["Artifacts", "Libdl"]
uuid = "e66e0078-7015-5450-92f7-15fbd957f2ae"
version = "1.1.1+0"
[[deps.Dates]]
deps = ["Printf"]
uuid = "ade2ca70-3891-5945-98fb-dc099432e06a"
version = "1.11.0"
[[deps.Downloads]]
deps = ["ArgTools", "FileWatching", "LibCURL", "NetworkOptions"]
uuid = "f43a241f-c20a-4ad4-852c-f6b1247861c6"
version = "1.6.0"
[[deps.Enzyme]]
deps = ["CEnum", "EnzymeCore", "Enzyme_jll", "GPUCompiler", "LLVM", "Libdl", "LinearAlgebra", "ObjectFile", "Preferences", "Printf", "Random"]
git-tree-sha1 = "b11d3c1aa7166ef05331ee762e7e1108722af436"
repo-rev = "main"
repo-url = "https://github.com/EnzymeAD/Enzyme.jl.git"
uuid = "7da242da-08ed-463a-9acd-ee780be4f1d9"
version = "0.12.24"
[deps.Enzyme.extensions]
EnzymeChainRulesCoreExt = "ChainRulesCore"
EnzymeLogExpFunctionsExt = "LogExpFunctions"
EnzymeSpecialFunctionsExt = "SpecialFunctions"
EnzymeStaticArraysExt = "StaticArrays"
[deps.Enzyme.weakdeps]
ChainRulesCore = "d360d2e6-b24c-11e9-a2a3-2a2ae2dbcce4"
LogExpFunctions = "2ab3a3ac-af41-5b50-aa03-7779005ae688"
SpecialFunctions = "276daf66-3868-5448-9aa4-cd146d93841b"
StaticArrays = "90137ffa-7385-5640-81b9-e52037218182"
[[deps.EnzymeCore]]
git-tree-sha1 = "d445df66dd8761a4c27df950db89c6a3a0629fe7"
uuid = "f151be2c-9106-41f4-ab19-57ee4f262869"
version = "0.7.7"
[deps.EnzymeCore.extensions]
AdaptExt = "Adapt"
[deps.EnzymeCore.weakdeps]
Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"
[[deps.Enzyme_jll]]
deps = ["Artifacts", "JLLWrappers", "LazyArtifacts", "Libdl", "TOML"]
git-tree-sha1 = "a509ac21f4f44df224bf7cb6901fa4236d015b5e"
repo-rev = "main"
repo-url = "https://github.com/JuliaBinaryWrappers/Enzyme_jll.jl.git"
uuid = "7cc45869-7501-5eee-bdea-0790c847d4ef"
version = "0.0.136+0"
[[deps.ExprTools]]
git-tree-sha1 = "27415f162e6028e81c72b82ef756bf321213b6ec"
uuid = "e2ba6199-217a-4e67-a87a-7c52f15ade04"
version = "0.1.10"
[[deps.FileWatching]]
uuid = "7b1f6079-737a-58dc-b8bc-7a2ca5c1b5ee"
version = "1.11.0"
[[deps.GPUCompiler]]
deps = ["ExprTools", "InteractiveUtils", "LLVM", "Libdl", "Logging", "PrecompileTools", "Preferences", "Scratch", "Serialization", "TOML", "TimerOutputs", "UUIDs"]
git-tree-sha1 = "36e1cbe62869fe2a95958d0d3fcd5dad47cac6fd"
repo-rev = "master"
repo-url = "https://github.com/JuliaGPU/GPUCompiler.jl.git"
uuid = "61eb1bfa-7361-4325-ad38-22787b887f55"
version = "0.26.7"
[[deps.InteractiveUtils]]
deps = ["Markdown"]
uuid = "b77e0a4c-d291-57a0-90e8-8db25a27a240"
version = "1.11.0"
[[deps.JLLWrappers]]
deps = ["Artifacts", "Preferences"]
git-tree-sha1 = "7e5d6779a1e09a36db2a7b6cff50942a0a7d0fca"
uuid = "692b3bcd-3c85-4b1f-b108-f13ce0eb3210"
version = "1.5.0"
[[deps.LLVM]]
deps = ["CEnum", "LLVMExtra_jll", "Libdl", "Preferences", "Printf", "Requires", "Unicode"]
git-tree-sha1 = "020abd49586480c1be84f57da0017b5d3db73f7c"
uuid = "929cbde3-209d-540e-8aea-75f648917ca0"
version = "8.0.0"
[deps.LLVM.extensions]
BFloat16sExt = "BFloat16s"
[deps.LLVM.weakdeps]
BFloat16s = "ab4f0b2a-ad5b-11e8-123f-65d77653426b"
[[deps.LLVMExtra_jll]]
deps = ["Artifacts", "JLLWrappers", "LazyArtifacts", "Libdl", "TOML"]
git-tree-sha1 = "c2636c264861edc6d305e6b4d528f09566d24c5e"
uuid = "dad2f222-ce93-54a1-a47d-0025e8a3acab"
version = "0.0.30+0"
[[deps.LazyArtifacts]]
deps = ["Artifacts", "Pkg"]
uuid = "4af54fe1-eca0-43a8-85a7-787d91b784e3"
version = "1.11.0"
[[deps.LibCURL]]
deps = ["LibCURL_jll", "MozillaCACerts_jll"]
uuid = "b27032c2-a3e7-50c8-80cd-2d36dbcbfd21"
version = "0.6.4"
[[deps.LibCURL_jll]]
deps = ["Artifacts", "LibSSH2_jll", "Libdl", "MbedTLS_jll", "Zlib_jll", "nghttp2_jll"]
uuid = "deac9b47-8bc7-5906-a0fe-35ac56dc84c0"
version = "8.6.0+0"
[[deps.LibGit2]]
deps = ["Base64", "LibGit2_jll", "NetworkOptions", "Printf", "SHA"]
uuid = "76f85450-5226-5b5a-8eaa-529ad045b433"
version = "1.11.0"
[[deps.LibGit2_jll]]
deps = ["Artifacts", "LibSSH2_jll", "Libdl", "MbedTLS_jll"]
uuid = "e37daf67-58a4-590a-8e99-b0245dd2ffc5"
version = "1.7.2+0"
[[deps.LibSSH2_jll]]
deps = ["Artifacts", "Libdl", "MbedTLS_jll"]
uuid = "29816b5a-b9ab-546f-933c-edad1886dfa8"
version = "1.11.0+1"
[[deps.Libdl]]
uuid = "8f399da3-3557-5675-b5ff-fb832c97cbdb"
version = "1.11.0"
[[deps.LinearAlgebra]]
deps = ["Libdl", "OpenBLAS_jll", "libblastrampoline_jll"]
uuid = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
version = "1.11.0"
[[deps.Logging]]
uuid = "56ddb016-857b-54e1-b83d-db4d58db5568"
version = "1.11.0"
[[deps.Markdown]]
deps = ["Base64"]
uuid = "d6f4376e-aef5-505a-96c1-9c027394607a"
version = "1.11.0"
[[deps.MbedTLS_jll]]
deps = ["Artifacts", "Libdl"]
uuid = "c8ffd9c3-330d-5841-b78e-0817d7145fa1"
version = "2.28.6+0"
[[deps.MozillaCACerts_jll]]
uuid = "14a3606d-f60d-562e-9121-12d972cd8159"
version = "2023.12.12"
[[deps.NetworkOptions]]
uuid = "ca575930-c2e3-43a9-ace4-1e988b2c1908"
version = "1.2.0"
[[deps.ObjectFile]]
deps = ["Reexport", "StructIO"]
git-tree-sha1 = "195e0a19842f678dd3473ceafbe9d82dfacc583c"
uuid = "d8793406-e978-5875-9003-1fc021f44a92"
version = "0.4.1"
[[deps.OpenBLAS_jll]]
deps = ["Artifacts", "CompilerSupportLibraries_jll", "Libdl"]
uuid = "4536629a-c528-5b80-bd46-f80d51c5b363"
version = "0.3.27+1"
[[deps.Pkg]]
deps = ["Artifacts", "Dates", "Downloads", "FileWatching", "LibGit2", "Libdl", "Logging", "Markdown", "Printf", "Random", "SHA", "TOML", "Tar", "UUIDs", "p7zip_jll"]
uuid = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
version = "1.11.0"
[deps.Pkg.extensions]
REPLExt = "REPL"
[deps.Pkg.weakdeps]
REPL = "3fa0cd96-eef1-5676-8a61-b3b8758bbffb"
[[deps.PrecompileTools]]
deps = ["Preferences"]
git-tree-sha1 = "5aa36f7049a63a1528fe8f7c3f2113413ffd4e1f"
uuid = "aea7be01-6a6a-4083-8856-8a6e6704d82a"
version = "1.2.1"
[[deps.Preferences]]
deps = ["TOML"]
git-tree-sha1 = "9306f6085165d270f7e3db02af26a400d580f5c6"
uuid = "21216c6a-2e73-6563-6e65-726566657250"
version = "1.4.3"
[[deps.Printf]]
deps = ["Unicode"]
uuid = "de0858da-6303-5e67-8744-51eddeeeb8d7"
version = "1.11.0"
[[deps.Random]]
deps = ["SHA"]
uuid = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
version = "1.11.0"
[[deps.Reexport]]
git-tree-sha1 = "45e428421666073eab6f2da5c9d310d99bb12f9b"
uuid = "189a3867-3050-52da-a836-e630ba90ab69"
version = "1.2.2"
[[deps.Requires]]
deps = ["UUIDs"]
git-tree-sha1 = "838a3a4188e2ded87a4f9f184b4b0d78a1e91cb7"
uuid = "ae029012-a4dd-5104-9daa-d747884805df"
version = "1.3.0"
[[deps.SHA]]
uuid = "ea8e919c-243c-51af-8825-aaa63cd721ce"
version = "0.7.0"
[[deps.Scratch]]
deps = ["Dates"]
git-tree-sha1 = "3bac05bc7e74a75fd9cba4295cde4045d9fe2386"
uuid = "6c6a2e73-6563-6170-7368-637461726353"
version = "1.2.1"
[[deps.Serialization]]
uuid = "9e88b42a-f829-5b0c-bbe9-9e923198166b"
version = "1.11.0"
[[deps.StructIO]]
deps = ["Test"]
git-tree-sha1 = "010dc73c7146869c042b49adcdb6bf528c12e859"
uuid = "53d494c1-5632-5724-8f4c-31dff12d585f"
version = "0.3.0"
[[deps.TOML]]
deps = ["Dates"]
uuid = "fa267f1f-6049-4f14-aa54-33bafae1ed76"
version = "1.0.3"
[[deps.Tar]]
deps = ["ArgTools", "SHA"]
uuid = "a4e569a6-e804-4fa4-b0f3-eef7a1d5b13e"
version = "1.10.0"
[[deps.Test]]
deps = ["InteractiveUtils", "Logging", "Random", "Serialization"]
uuid = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
version = "1.11.0"
[[deps.TimerOutputs]]
deps = ["ExprTools", "Printf"]
git-tree-sha1 = "5a13ae8a41237cff5ecf34f73eb1b8f42fff6531"
uuid = "a759f4b9-e2f1-59dc-863e-4aeb61b1ea8f"
version = "0.5.24"
[[deps.UUIDs]]
deps = ["Random", "SHA"]
uuid = "cf7118a7-6976-5b1a-9a39-7adc72f591a4"
version = "1.11.0"
[[deps.Unicode]]
uuid = "4ec0a83e-493e-50e2-b9ac-8f72acf5a8f5"
version = "1.11.0"
[[deps.Zlib_jll]]
deps = ["Libdl"]
uuid = "83775a58-1f1d-513f-b197-d71354ab007a"
version = "1.2.13+1"
[[deps.libblastrampoline_jll]]
deps = ["Artifacts", "Libdl"]
uuid = "8e850b90-86db-534c-a0d3-1478176c7d93"
version = "5.8.0+1"
[[deps.nghttp2_jll]]
deps = ["Artifacts", "Libdl"]
uuid = "8e850ede-7688-5339-a07c-302acd2aaf8d"
version = "1.59.0+0"
[[deps.p7zip_jll]]
deps = ["Artifacts", "Libdl"]
uuid = "3f19e933-33d8-53b3-aaab-bd5110c3b7a0"
version = "17.4.0+2"Installing the latest release of GPUCompiler I get a different error: ErrorERROR: UndefVarError: `PassBuilder` not defined in `Enzyme.Compiler`
Suggestion: check for spelling errors or missing imports.
Stacktrace:
[1] macro expansion
@ ~/.julia/packages/LLVM/5DlHM/src/base.jl:96 [inlined]
[2] (::Enzyme.Compiler.var"#prop_julia_addr#28202"{LLVM.TargetMachine})(f::LLVM.Function)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler/optimize.jl:75
[3] function_pass_callback(ptr::Ptr{Nothing}, data::Ptr{Nothing})
@ LLVM ~/.julia/packages/LLVM/5DlHM/src/pass.jl:49
[4] LLVMRunPassManager
@ ~/.julia/packages/LLVM/5DlHM/lib/16/libLLVM.jl:3351 [inlined]
[5] run!
@ ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:39 [inlined]
[6] (::Enzyme.Compiler.var"#28298#28299"{LLVM.Module, LLVM.TargetMachine})(pm::LLVM.ModulePassManager)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler/optimize.jl:2029
[7] LLVM.ModulePassManager(::Enzyme.Compiler.var"#28298#28299"{LLVM.Module, LLVM.TargetMachine}; kwargs::@Kwargs{})
@ LLVM ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:33
[8] ModulePassManager
@ ~/.julia/packages/LLVM/5DlHM/src/passmanager.jl:30 [inlined]
[9] optimize!
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler/optimize.jl:1951 [inlined]
[10] codegen(output::Symbol, job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}; libraries::Bool, deferred_codegen::Bool, optimize::Bool, toplevel::Bool, strip::Bool, validate::Bool, only_entry::Bool, parent_job::Nothing)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:5807
[11] codegen
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:5208 [inlined]
[12] _thunk(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, postopt::Bool)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6710
[13] _thunk
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6710 [inlined]
[14] cached_compilation
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6748 [inlined]
[15] thunkbase(ctx::LLVM.Context, mi::Core.MethodInstance, ::Val{0x0000000000006819}, ::Type{Const{…}}, ::Type{Active}, tt::Type{Tuple{…}}, ::Val{Enzyme.API.DEM_ReverseModeCombined}, ::Val{1}, ::Val{(false, false)}, ::Val{false}, ::Val{false}, ::Type{FFIABI})
@ Enzyme.Compiler ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6821
[16] #s2021#28415
@ ~/.julia/packages/Enzyme/YDcYf/src/compiler.jl:6873 [inlined]
[17] var"#s2021#28415"(FA::Any, A::Any, TT::Any, Mode::Any, ModifiedBetween::Any, width::Any, ReturnPrimal::Any, ShadowInit::Any, World::Any, ABI::Any, ::Any, ::Any, ::Any, ::Any, tt::Any, ::Any, ::Any, ::Any, ::Any, ::Any, ::Any)
@ Enzyme.Compiler ./none:0
[18] (::Core.GeneratedFunctionStub)(::UInt64, ::LineNumberNode, ::Any, ::Vararg{Any})
@ Core ./boot.jl:709
[19] autodiff
@ ~/.julia/packages/Enzyme/YDcYf/src/Enzyme.jl:309 [inlined]
[20] autodiff
@ ~/.julia/packages/Enzyme/YDcYf/src/Enzyme.jl:326 [inlined]
[21] gradient(rm::ReverseMode{false, FFIABI, false}, f::typeof(sum), x::Vector{Float32})
@ Enzyme ~/.julia/packages/Enzyme/YDcYf/src/Enzyme.jl:1027
[22] top-level scope
@ REPL[3]:1
[23] top-level scope
@ none:1
Some type information was truncated. Use `show(err)` to see complete types. |
|
That said enzyme definitely has some older GPUCompiler in compat which shouldn't be there. Currently installing AMDGPU (which installs GPUCompiler 0.26.5) causes enzyme to not precompile https://buildkite.com/julialang/luxlib-dot-jl/builds/835#0190d6c1-141c-4f6b-ab1d-eec8c2e4f7bc/317-648 |
Same here with the latest release candidate of Julia 1.11: System info: |
|
Present issue here is that LLVM.jl dropped support for API's which we need to support 1.11 x/ref maleadt/LLVM.jl#435 |
|
I was wondering if there are any updates on 1.11 support, as the release draws nearer? |
|
Various codes work now and no precompilation failures, but not all. Specifically support for the new gc_loaded intrinsic needs to be added, but I don't understand the semantics of it yet and need help from @gbaraldi and or @vtjnash to add. If you understand the meaning of it well enough to explain it and/or support it, be my guest! But since it's a GC related thing and I don't want to accidentally cause segfaults, it remains as an error atm. |
|
I don't know the first thing about that, just wanted to check if I could re-activate DI tests for Enzyme on v1.11. Guess I'll give it a try! |
|
With 1.11 released, I am seeing quite a few GC failures https://buildkite.com/julialang/lux-dot-jl/builds/4429#019269e2-95b3-4504-8f1e-48deeadeeaee/322-701. I am assuming this is the unsupported feature you mentioned? |
|
Here's a simple MWE distilled from #1951, in case it helps: using Enzyme
f(x) = sum(x .* transpose(x))
Enzyme.gradient(Enzyme.Reverse, f, [1.0])julia> Enzyme.gradient(Enzyme.Reverse, f, [1.0])
ERROR: Enzyme compilation failed.
Current scope:
; Function Attrs: mustprogress willreturn
define "enzyme_type"="{[-1]:Float@double}" double @preprocess_julia_f_47698({} addrspace(10)* noundef nonnull align 8 dereferenceable(24) "enzyme_type"="{[-1]:Pointer, [-1,0]:Pointer, [-1,8]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double, [-1,16]:Integer, [-1,17]:Integer, [-1,18]:Integer, [-1,19]:Integer, [-1,20]:Integer, [-1,21]:Integer, [-1,22]:Integer, [-1,23]:Integer}" "enzymejl_parmtype"="126279284194864" "enzymejl_parmtype_ref"="2" %0) local_unnamed_addr #22 !dbg !459 {
top:
%1 = alloca [2 x i64], align 8
%2 = alloca [1 x {} addrspace(10)*], align 8
%3 = alloca [1 x i64], align 8
%4 = alloca [1 x i64], align 8
%5 = alloca [1 x i64], align 8
%6 = alloca [1 x i64], align 8
%7 = alloca [1 x i64], align 8
%pgcstack = call {}*** @julia.get_pgcstack() #27
%current_task1227 = getelementptr inbounds {}**, {}*** %pgcstack, i64 -14
%8 = bitcast {}*** %current_task1227 to {}*
%ptls_field228 = getelementptr inbounds {}**, {}*** %pgcstack, i64 2
%9 = bitcast {}*** %ptls_field228 to i64***
%ptls_load229230 = load i64**, i64*** %9, align 8, !tbaa !18
%10 = getelementptr inbounds i64*, i64** %ptls_load229230, i64 2
%safepoint = load i64*, i64** %10, align 8, !tbaa !22
fence syncscope("singlethread") seq_cst
call void @julia.safepoint(i64* %safepoint) #27, !dbg !460
fence syncscope("singlethread") seq_cst
%11 = addrspacecast {} addrspace(10)* %0 to {} addrspace(11)*, !dbg !461
%12 = bitcast {} addrspace(10)* %0 to i8 addrspace(10)*, !dbg !461
%13 = addrspacecast i8 addrspace(10)* %12 to i8 addrspace(11)*, !dbg !461
%14 = getelementptr inbounds i8, i8 addrspace(11)* %13, i64 16, !dbg !461
%15 = bitcast i8 addrspace(11)* %14 to i64 addrspace(11)*, !dbg !461
%16 = load i64, i64 addrspace(11)* %15, align 8, !dbg !461, !tbaa !25, !alias.scope !67, !noalias !68
%17 = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 %16, i64 %16) #27, !dbg !466
%18 = extractvalue { i64, i1 } %17, 0, !dbg !466
%19 = extractvalue { i64, i1 } %17, 1, !dbg !466
%20 = icmp ugt i64 %16, 9223372036854775806, !dbg !476
%21 = or i1 %20, %19, !dbg !476
br i1 %21, label %L49, label %L53, !dbg !477
L49: ; preds = %top
%22 = call noalias nonnull align 8 dereferenceable(8) "enzyme_inactive" "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 8, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279338028512 to {}*) to {} addrspace(10)*)) #28, !dbg !477
%23 = bitcast {} addrspace(10)* %22 to [1 x {} addrspace(10)*] addrspace(10)*, !dbg !477, !enzyme_inactive !0
%24 = getelementptr [1 x {} addrspace(10)*], [1 x {} addrspace(10)*] addrspace(10)* %23, i64 0, i64 0, !dbg !477
store {} addrspace(10)* addrspacecast ({}* inttoptr (i64 126279346398928 to {}*) to {} addrspace(10)*), {} addrspace(10)* addrspace(10)* %24, align 8, !dbg !477, !tbaa !41, !alias.scope !45, !noalias !478
%25 = addrspacecast {} addrspace(10)* %22 to {} addrspace(12)*, !dbg !477, !enzyme_inactive !0
call void @ijl_throw({} addrspace(12)* %25) #29, !dbg !477
unreachable, !dbg !477
L53: ; preds = %top
%.not = icmp eq i64 %18, 0, !dbg !481
br i1 %.not, label %L55, label %L57, !dbg !481
L55: ; preds = %L53
%26 = load atomic {} addrspace(10)*, {} addrspace(10)** inttoptr (i64 126279284195088 to {} addrspace(10)**) unordered, align 16, !dbg !483, !tbaa !22, !alias.scope !39, !noalias !40
%.not231 = icmp eq {} addrspace(10)* %26, null, !dbg !483
br i1 %.not231, label %fail, label %L59, !dbg !483
L57: ; preds = %L53
%27 = call "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @jl_alloc_genericmemory({} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284195056 to {}*) to {} addrspace(10)*), i64 %18) #27, !dbg !484
br label %L59, !dbg !484
L59: ; preds = %L57, %L55
%value_phi8 = phi {} addrspace(10)* [ %27, %L57 ], [ %26, %L55 ]
%28 = bitcast {} addrspace(10)* %value_phi8 to { i64, {} addrspace(10)** } addrspace(10)*, !dbg !485
%29 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %28 to { i64, {} addrspace(10)** } addrspace(11)*, !dbg !485
%30 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %29, i64 0, i32 1, !dbg !485
%31 = bitcast {} addrspace(10)** addrspace(11)* %30 to i8* addrspace(11)*, !dbg !485
%32 = load i8*, i8* addrspace(11)* %31, align 8, !dbg !485, !tbaa !25, !alias.scope !96, !noalias !97, !nonnull !0
%33 = call noalias nonnull align 8 dereferenceable(32) "enzyme_type"="{[-1]:Pointer, [-1,0]:Pointer, [-1,8]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double, [-1,16]:Integer, [-1,17]:Integer, [-1,18]:Integer, [-1,19]:Integer, [-1,20]:Integer, [-1,21]:Integer, [-1,22]:Integer, [-1,23]:Integer, [-1,24]:Integer, [-1,25]:Integer, [-1,26]:Integer, [-1,27]:Integer, [-1,28]:Integer, [-1,29]:Integer, [-1,30]:Integer, [-1,31]:Integer}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 32, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279294399424 to {}*) to {} addrspace(10)*)) #28, !dbg !468
%34 = bitcast {} addrspace(10)* %33 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !468
%35 = addrspacecast { i8*, {} addrspace(10)* } addrspace(10)* %34 to { i8*, {} addrspace(10)* } addrspace(11)*, !dbg !468
%.repack = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %35, i64 0, i32 0, !dbg !468
store i8* %32, i8* addrspace(11)* %.repack, align 8, !dbg !468, !tbaa !98, !alias.scope !101, !noalias !486
%.repack232 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %35, i64 0, i32 1, !dbg !468
store {} addrspace(10)* %value_phi8, {} addrspace(10)* addrspace(11)* %.repack232, align 8, !dbg !468, !tbaa !98, !alias.scope !101, !noalias !486
%36 = bitcast {} addrspace(10)* %33 to i8 addrspace(10)*, !dbg !468
%37 = addrspacecast i8 addrspace(10)* %36 to i8 addrspace(11)*, !dbg !468
%38 = getelementptr inbounds i8, i8 addrspace(11)* %37, i64 16, !dbg !468
%.sroa.0167.0..sroa_idx = bitcast i8 addrspace(11)* %38 to i64 addrspace(11)*, !dbg !468
store i64 %16, i64 addrspace(11)* %.sroa.0167.0..sroa_idx, align 8, !dbg !468, !tbaa !25, !alias.scope !103, !noalias !487
%.sroa.2168.0..sroa_idx169 = getelementptr inbounds i8, i8 addrspace(11)* %37, i64 24, !dbg !468
%39 = bitcast i8 addrspace(11)* %.sroa.2168.0..sroa_idx169 to i64 addrspace(11)*, !dbg !468
store i64 %16, i64 addrspace(11)* %39, align 8, !dbg !468, !tbaa !25, !alias.scope !103, !noalias !487
%40 = addrspacecast {} addrspace(10)* %value_phi8 to {} addrspace(11)*, !dbg !488
%41 = bitcast {} addrspace(10)* %value_phi8 to i64 addrspace(10)*, !dbg !498
%42 = addrspacecast i64 addrspace(10)* %41 to i64 addrspace(11)*, !dbg !498
%43 = load i64, i64 addrspace(11)* %42, align 8, !dbg !498, !tbaa !126, !alias.scope !101, !noalias !128
%44 = icmp eq i64 %43, 0, !dbg !498
%45 = icmp eq i64 %16, 0
%or.cond = select i1 %44, i1 true, i1 %45, !dbg !490
%46 = bitcast i8 addrspace(10)* %12 to i64 addrspace(10)*, !dbg !490
br i1 %or.cond, label %L152, label %L102, !dbg !490
L102: ; preds = %L59
%47 = load {} addrspace(10)**, {} addrspace(10)** addrspace(11)* %30, align 8, !dbg !499, !tbaa !132, !alias.scope !101, !noalias !128, !nonnull !0
%48 = bitcast {} addrspace(10)* %value_phi8 to {} addrspace(10)* addrspace(10)*, !dbg !499
%49 = addrspacecast {} addrspace(10)* addrspace(10)* %48 to {} addrspace(10)* addrspace(11)*, !dbg !499
%50 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %49, i64 2, !dbg !499
%51 = addrspacecast {} addrspace(10)** %47 to {} addrspace(10)* addrspace(11)*, !dbg !499
%.not234 = icmp eq {} addrspace(10)* addrspace(11)* %50, %51, !dbg !499
br i1 %.not234, label %guard_exit, label %guard_pass, !dbg !499
L138: ; preds = %guard_exit15
%52 = bitcast {} addrspace(10)* %0 to i8* addrspace(10)*, !dbg !501
%53 = addrspacecast i8* addrspace(10)* %52 to i8* addrspace(11)*, !dbg !501
%54 = load i8*, i8* addrspace(11)* %53, align 8, !dbg !501, !tbaa !98, !alias.scope !101, !noalias !128
%55 = ptrtoint i8* %54 to i64, !dbg !504
%56 = load atomic {} addrspace(10)* ({} addrspace(10)*, i64, i64)*, {} addrspace(10)* ({} addrspace(10)*, i64, i64)** bitcast (void ()** @jlplt_jl_genericmemory_copy_slice_47719_got to {} addrspace(10)* ({} addrspace(10)*, i64, i64)**) unordered, align 8, !dbg !504
%57 = call nonnull "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* %56({} addrspace(10)* %253, i64 %55, i64 %16) #27, !dbg !504
%58 = bitcast {} addrspace(10)* %57 to { i64, {} addrspace(10)** } addrspace(10)*, !dbg !505
%59 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %58 to { i64, {} addrspace(10)** } addrspace(11)*, !dbg !505
%60 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %59, i64 0, i32 1, !dbg !505
%61 = bitcast {} addrspace(10)** addrspace(11)* %60 to i8* addrspace(11)*, !dbg !505
%62 = load i8*, i8* addrspace(11)* %61, align 8, !dbg !505, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%63 = load i64, i64 addrspace(11)* %15, align 8, !dbg !507, !tbaa !145, !alias.scope !101, !noalias !128
%64 = call noalias nonnull align 8 dereferenceable(24) "enzyme_type"="{[-1]:Pointer, [-1,0]:Pointer, [-1,8]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double, [-1,16]:Integer, [-1,17]:Integer, [-1,18]:Integer, [-1,19]:Integer, [-1,20]:Integer, [-1,21]:Integer, [-1,22]:Integer, [-1,23]:Integer}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 24, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194864 to {}*) to {} addrspace(10)*)) #28, !dbg !506
%65 = addrspacecast {} addrspace(10)* %64 to {} addrspace(11)*, !dbg !506
%66 = bitcast {} addrspace(10)* %64 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !506
%67 = addrspacecast { i8*, {} addrspace(10)* } addrspace(10)* %66 to { i8*, {} addrspace(10)* } addrspace(11)*, !dbg !506
%.repack240 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %67, i64 0, i32 0, !dbg !506
store i8* %62, i8* addrspace(11)* %.repack240, align 8, !dbg !506, !tbaa !98, !alias.scope !101, !noalias !486
%.repack241 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %67, i64 0, i32 1, !dbg !506
store {} addrspace(10)* %57, {} addrspace(10)* addrspace(11)* %.repack241, align 8, !dbg !506, !tbaa !98, !alias.scope !101, !noalias !486
%68 = bitcast {} addrspace(10)* %64 to i8 addrspace(10)*, !dbg !506
%69 = addrspacecast i8 addrspace(10)* %68 to i8 addrspace(11)*, !dbg !506
%70 = getelementptr inbounds i8, i8 addrspace(11)* %69, i64 16, !dbg !506
%71 = bitcast i8 addrspace(11)* %70 to i64 addrspace(11)*, !dbg !506
store i64 %63, i64 addrspace(11)* %71, align 8, !dbg !506, !tbaa !145, !alias.scope !101, !noalias !486
%.pre203 = load i64, i64 addrspace(11)* %42, align 8, !dbg !508, !tbaa !126, !alias.scope !101, !noalias !128
%72 = bitcast i8 addrspace(10)* %68 to i64 addrspace(10)*, !dbg !506
br label %L152, !dbg !506
L152: ; preds = %guard_exit15, %L138, %L59
%nodecayed..pre-phi = phi i64 addrspace(10)* [ %277, %guard_exit15 ], [ %46, %L59 ], [ %72, %L138 ], !dbg !515
%nodecayed..pre-phi214 = phi {} addrspace(10)* [ %0, %guard_exit15 ], [ %0, %L59 ], [ %64, %L138 ], !dbg !515
%73 = phi i64 [ %43, %guard_exit15 ], [ %43, %L59 ], [ %.pre203, %L138 ], !dbg !508
%74 = phi i64 [ %16, %guard_exit15 ], [ %16, %L59 ], [ %63, %L138 ], !dbg !515
%75 = addrspacecast i64 addrspace(10)* %nodecayed..pre-phi to i64 addrspace(11)*, !dbg !519
%76 = bitcast i64 addrspace(11)* %75 to i8 addrspace(11)*, !dbg !519
%77 = getelementptr i8, i8 addrspace(11)* %76, i64 16, !dbg !519
%78 = bitcast i8 addrspace(11)* %77 to i64 addrspace(11)*, !dbg !519
%79 = addrspacecast {} addrspace(10)* %nodecayed..pre-phi214 to {} addrspace(11)*, !dbg !519
%.not702 = icmp eq i64 %74, 1, !dbg !519
%80 = icmp eq i64 %73, 0, !dbg !508
%81 = icmp eq i64 %74, 0
%or.cond301 = select i1 %80, i1 true, i1 %81, !dbg !510
br i1 %or.cond301, label %L227, label %L172, !dbg !510
L172: ; preds = %L152
%82 = load {} addrspace(10)**, {} addrspace(10)** addrspace(11)* %30, align 8, !dbg !523, !tbaa !132, !alias.scope !101, !noalias !128, !nonnull !0
%83 = bitcast {} addrspace(10)* %value_phi8 to {} addrspace(10)* addrspace(10)*, !dbg !523
%84 = addrspacecast {} addrspace(10)* addrspace(10)* %83 to {} addrspace(10)* addrspace(11)*, !dbg !523
%85 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %84, i64 2, !dbg !523
%86 = addrspacecast {} addrspace(10)** %82 to {} addrspace(10)* addrspace(11)*, !dbg !523
%.not243 = icmp eq {} addrspace(10)* addrspace(11)* %85, %86, !dbg !523
br i1 %.not243, label %guard_exit23, label %guard_pass22, !dbg !523
L209: ; preds = %guard_exit28
%87 = bitcast {} addrspace(10)* %0 to i8* addrspace(10)*, !dbg !525
%88 = addrspacecast i8* addrspace(10)* %87 to i8* addrspace(11)*, !dbg !525
%89 = load i8*, i8* addrspace(11)* %88, align 8, !dbg !525, !tbaa !98, !alias.scope !101, !noalias !128
%90 = ptrtoint i8* %89 to i64, !dbg !529
%91 = load atomic {} addrspace(10)* ({} addrspace(10)*, i64, i64)*, {} addrspace(10)* ({} addrspace(10)*, i64, i64)** bitcast (void ()** @jlplt_jl_genericmemory_copy_slice_47719_got to {} addrspace(10)* ({} addrspace(10)*, i64, i64)**) unordered, align 8, !dbg !529
%92 = call nonnull "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* %91({} addrspace(10)* %294, i64 %90, i64 %74) #27, !dbg !529
%93 = bitcast {} addrspace(10)* %92 to { i64, {} addrspace(10)** } addrspace(10)*, !dbg !530
%94 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %93 to { i64, {} addrspace(10)** } addrspace(11)*, !dbg !530
%95 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %94, i64 0, i32 1, !dbg !530
%96 = bitcast {} addrspace(10)** addrspace(11)* %95 to i8* addrspace(11)*, !dbg !530
%97 = load i8*, i8* addrspace(11)* %96, align 8, !dbg !530, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%98 = load i64, i64 addrspace(11)* %15, align 8, !dbg !532, !tbaa !145, !alias.scope !101, !noalias !128
%99 = call noalias nonnull align 8 dereferenceable(24) "enzyme_type"="{[-1]:Pointer, [-1,0]:Pointer, [-1,8]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double, [-1,16]:Integer, [-1,17]:Integer, [-1,18]:Integer, [-1,19]:Integer, [-1,20]:Integer, [-1,21]:Integer, [-1,22]:Integer, [-1,23]:Integer}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 24, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194864 to {}*) to {} addrspace(10)*)) #28, !dbg !531
%100 = addrspacecast {} addrspace(10)* %99 to {} addrspace(11)*, !dbg !531
%101 = bitcast {} addrspace(10)* %99 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !531
%102 = addrspacecast { i8*, {} addrspace(10)* } addrspace(10)* %101 to { i8*, {} addrspace(10)* } addrspace(11)*, !dbg !531
%.repack249 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %102, i64 0, i32 0, !dbg !531
store i8* %97, i8* addrspace(11)* %.repack249, align 8, !dbg !531, !tbaa !98, !alias.scope !101, !noalias !486
%.repack250 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %102, i64 0, i32 1, !dbg !531
store {} addrspace(10)* %92, {} addrspace(10)* addrspace(11)* %.repack250, align 8, !dbg !531, !tbaa !98, !alias.scope !101, !noalias !486
%103 = bitcast {} addrspace(10)* %99 to i8 addrspace(10)*, !dbg !531
%104 = addrspacecast i8 addrspace(10)* %103 to i8 addrspace(11)*, !dbg !531
%105 = getelementptr inbounds i8, i8 addrspace(11)* %104, i64 16, !dbg !531
%106 = bitcast i8 addrspace(11)* %105 to i64 addrspace(11)*, !dbg !531
store i64 %98, i64 addrspace(11)* %106, align 8, !dbg !531, !tbaa !145, !alias.scope !101, !noalias !486
br label %L227, !dbg !533
L227: ; preds = %guard_exit28, %L209, %L152
%nodecayed..pre-phi217 = phi {} addrspace(10)* [ %99, %L209 ], [ %0, %L152 ], [ %0, %guard_exit28 ], !dbg !534
%107 = phi i64 [ %98, %L209 ], [ %74, %L152 ], [ %74, %guard_exit28 ], !dbg !534
%108 = addrspacecast {} addrspace(10)* %nodecayed..pre-phi217 to {} addrspace(11)*, !dbg !539
%.not703 = icmp eq i64 %107, 1, !dbg !539
br i1 %45, label %L451, label %L252.preheader, !dbg !544
L252.preheader: ; preds = %L227
%109 = bitcast {} addrspace(11)* %108 to { i8*, {} addrspace(10)* } addrspace(11)*
%110 = bitcast {} addrspace(11)* %108 to i8* addrspace(11)*
%111 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %109, i64 0, i32 1
%112 = bitcast i8* %32 to double*
%113 = bitcast i8* %32 to {} addrspace(10)**
%114 = call {} addrspace(10)* addrspace(13)* @julia.gc_loaded({} addrspace(10)* noundef %value_phi8, {} addrspace(10)** noundef %113) #27, !dbg !546
%115 = load i64, i64 addrspace(11)* %78, align 8, !tbaa !145, !alias.scope !101, !noalias !128
%.not257.peel.not = icmp eq i64 %115, 0
%exitcond629.peel.not = icmp eq i64 %16, 1
br i1 %.not257.peel.not, label %L316, label %L252.preheader.split, !dbg !547
L252.preheader.split: ; preds = %L252.preheader
%116 = bitcast {} addrspace(11)* %79 to { i8*, {} addrspace(10)* } addrspace(11)*
%117 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %116, i64 0, i32 1
%118 = bitcast {} addrspace(11)* %79 to i8* addrspace(11)*
%.pre913 = load i8*, i8* addrspace(11)* %118, align 8, !dbg !554, !tbaa !98, !alias.scope !101, !noalias !128
%.pre914 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %117, align 8, !dbg !554, !tbaa !98, !alias.scope !101, !noalias !128
%.phi.trans.insert = bitcast {} addrspace(10)* %.pre914 to i64 addrspace(10)*
%.phi.trans.insert915 = addrspacecast i64 addrspace(10)* %.phi.trans.insert to i64 addrspace(11)*
%.pre916 = load i64, i64 addrspace(11)* %.phi.trans.insert915, align 8, !dbg !554, !tbaa !126, !range !211, !alias.scope !101, !noalias !128
%.phi.trans.insert917 = bitcast {} addrspace(10)* %.pre914 to { i64, {} addrspace(10)** } addrspace(10)*
%.phi.trans.insert918 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %.phi.trans.insert917 to { i64, {} addrspace(10)** } addrspace(11)*
%.phi.trans.insert919 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %.phi.trans.insert918, i64 0, i32 1
%.phi.trans.insert920 = bitcast {} addrspace(10)** addrspace(11)* %.phi.trans.insert919 to i8* addrspace(11)*
%.pre921 = load i8*, i8* addrspace(11)* %.phi.trans.insert920, align 8, !dbg !554, !tbaa !22, !alias.scope !39, !noalias !40
br label %L252, !dbg !555
L252: ; preds = %L424, %L252.preheader.split
%iv = phi i64 [ %iv.next, %L424 ], [ 0, %L252.preheader.split ]
%iv.next = add nuw nsw i64 %iv, 1
%119 = add i64 %iv.next, -1
%120 = select i1 %.not703, i64 1, i64 %iv.next
%121 = add i64 %120, -1
%122 = mul i64 %119, %16
%123 = shl nuw nsw i64 %.pre916, 1, !dbg !554
%.not258.peel = icmp ne i64 %.pre916, 0, !dbg !554
%124 = bitcast i8* %.pre913 to double*, !dbg !554
%125 = ptrtoint i8* %.pre921 to i64, !dbg !554
%126 = ptrtoint i8* %.pre913 to i64, !dbg !554
%127 = sub i64 %126, %125, !dbg !554
%128 = shl nuw nsw i64 %.pre916, 3, !dbg !554
%129 = icmp ult i64 %127, %128, !dbg !554
%130 = and i1 %.not258.peel, %129, !dbg !554
br i1 %130, label %load.peel, label %oob.loopexit1, !dbg !554
load.peel: ; preds = %L252
%131 = bitcast i8* %.pre913 to {} addrspace(10)**, !dbg !554
%132 = call {} addrspace(10)* addrspace(13)* @julia.gc_loaded({} addrspace(10)* noundef %.pre914, {} addrspace(10)** noundef %131) #27, !dbg !554
%133 = bitcast {} addrspace(10)* addrspace(13)* %132 to double addrspace(13)*, !dbg !554
%134 = load double, double addrspace(13)* %133, align 8, !dbg !554, !tbaa !213, !alias.scope !45, !noalias !215
%.not262.peel = icmp ult i64 %121, %107, !dbg !556
br i1 %.not262.peel, label %L365.peel, label %L343, !dbg !561
L365.peel: ; preds = %load.peel
%135 = load i8*, i8* addrspace(11)* %110, align 8, !dbg !566, !tbaa !98, !alias.scope !101, !noalias !128
%136 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %111, align 8, !dbg !566, !tbaa !98, !alias.scope !101, !noalias !128, !dereferenceable_or_null !237, !align !238
%137 = bitcast {} addrspace(10)* %136 to i64 addrspace(10)*, !dbg !566
%138 = addrspacecast i64 addrspace(10)* %137 to i64 addrspace(11)*, !dbg !566
%139 = load i64, i64 addrspace(11)* %138, align 8, !dbg !566, !tbaa !126, !range !211, !alias.scope !101, !noalias !128
%140 = shl nuw nsw i64 %139, 1, !dbg !566
%141 = add i64 %139, %121, !dbg !566
%.not264.peel = icmp ult i64 %141, %140, !dbg !566
%142 = bitcast i8* %135 to double*, !dbg !566
%143 = getelementptr inbounds double, double* %142, i64 %121, !dbg !566
%144 = bitcast {} addrspace(10)* %136 to { i64, {} addrspace(10)** } addrspace(10)*, !dbg !566
%145 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %144 to { i64, {} addrspace(10)** } addrspace(11)*, !dbg !566
%146 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %145, i64 0, i32 1, !dbg !566
%147 = bitcast {} addrspace(10)** addrspace(11)* %146 to i8* addrspace(11)*, !dbg !566
%148 = load i8*, i8* addrspace(11)* %147, align 8, !dbg !566, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%149 = ptrtoint i8* %148 to i64, !dbg !566
%150 = ptrtoint double* %143 to i64, !dbg !566
%151 = sub i64 %150, %149, !dbg !566
%152 = shl nuw nsw i64 %139, 3, !dbg !566
%153 = icmp ult i64 %151, %152, !dbg !566
%154 = and i1 %.not264.peel, %153, !dbg !566
br i1 %154, label %idxend50.peel, label %oob48, !dbg !566
idxend50.peel: ; preds = %L365.peel
%155 = icmp eq i64 %139, 0, !dbg !566
br i1 %155, label %oob51, label %load52.peel, !dbg !566
load52.peel: ; preds = %idxend50.peel
%156 = bitcast i8* %135 to {} addrspace(10)**, !dbg !566
%157 = call {} addrspace(10)* addrspace(13)* @julia.gc_loaded({} addrspace(10)* noundef %136, {} addrspace(10)** noundef %156) #27, !dbg !566
%158 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(13)* %157, i64 %121, !dbg !566
%159 = bitcast {} addrspace(10)* addrspace(13)* %158 to double addrspace(13)*, !dbg !566
%160 = load double, double addrspace(13)* %159, align 8, !dbg !566, !tbaa !213, !alias.scope !45, !noalias !215
%161 = fmul double %134, %160, !dbg !569
%162 = load i64, i64 addrspace(11)* %42, align 8, !dbg !572, !tbaa !126, !range !211, !alias.scope !101, !noalias !128
%163 = shl nuw nsw i64 %162, 1, !dbg !572
%164 = add i64 %162, %122, !dbg !572
%.not268.peel = icmp ult i64 %164, %163, !dbg !572
%165 = getelementptr inbounds double, double* %112, i64 %122, !dbg !572
%166 = load i8*, i8* addrspace(11)* %31, align 8, !dbg !572, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%167 = ptrtoint i8* %166 to i64, !dbg !572
%168 = ptrtoint double* %165 to i64, !dbg !572
%169 = sub i64 %168, %167, !dbg !572
%170 = shl nuw nsw i64 %162, 3, !dbg !572
%171 = icmp ult i64 %169, %170, !dbg !572
%172 = and i1 %.not268.peel, %171, !dbg !572
br i1 %172, label %idxend55.peel, label %oob53.loopexit2, !dbg !572
idxend55.peel: ; preds = %load52.peel
%173 = icmp eq i64 %162, 0, !dbg !572
br i1 %173, label %oob56, label %load57.peel, !dbg !572
load57.peel: ; preds = %idxend55.peel
%174 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(13)* %114, i64 %122, !dbg !572
%175 = bitcast {} addrspace(10)* addrspace(13)* %174 to double addrspace(13)*, !dbg !572
store double %161, double addrspace(13)* %175, align 8, !dbg !572, !tbaa !213, !alias.scope !45, !noalias !478
br i1 %exitcond629.peel.not, label %L424, label %L304.preheader, !dbg !574
L304.preheader: ; preds = %load57.peel
br label %L304, !dbg !547
L304: ; preds = %L304.preheader, %load57
%iv3 = phi i64 [ 0, %L304.preheader ], [ %iv.next4, %load57 ]
%iv.next4 = add nuw nsw i64 %iv3, 1, !dbg !575
%176 = add nuw nsw i64 %iv.next4, 1, !dbg !575
%177 = select i1 %.not702, i64 1, i64 %176, !dbg !578
%178 = add nsw i64 %177, -1, !dbg !547
%.not257 = icmp ult i64 %178, %115, !dbg !547
br i1 %.not257, label %L319, label %L316.loopexit, !dbg !547
L316.loopexit: ; preds = %L304
br label %L316, !dbg !547
L316: ; preds = %L316.loopexit, %L252.preheader
%.lcssa531 = phi i64 [ 1, %L252.preheader ], [ %177, %L316.loopexit ], !dbg !578
%179 = getelementptr inbounds [1 x i64], [1 x i64]* %3, i64 0, i64 0, !dbg !547
store i64 %.lcssa531, i64* %179, align 8, !dbg !547, !tbaa !262, !alias.scope !264, !noalias !581
%180 = addrspacecast [1 x i64]* %3 to [1 x i64] addrspace(11)*, !dbg !547
call fastcc void @julia_throw_boundserror_47790({} addrspace(10)* nofree noundef nonnull align 8 dereferenceable(24) %nodecayed..pre-phi214, [1 x i64] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) %180) #29, !dbg !547
unreachable, !dbg !547
L319: ; preds = %L304
%181 = add i64 %178, %.pre916, !dbg !554
%.not258 = icmp ult i64 %181, %123, !dbg !554
%182 = getelementptr inbounds double, double* %124, i64 %178, !dbg !554
%183 = ptrtoint double* %182 to i64, !dbg !554
%184 = sub i64 %183, %125, !dbg !554
%185 = icmp ult i64 %184, %128, !dbg !554
%186 = and i1 %.not258, %185, !dbg !554
br i1 %186, label %load52, label %oob.loopexit, !dbg !554
L343: ; preds = %load.peel
%187 = getelementptr inbounds [2 x i64], [2 x i64]* %1, i64 0, i64 0
%188 = getelementptr inbounds [2 x i64], [2 x i64]* %1, i64 0, i64 1
store i64 1, i64* %187, align 8, !dbg !582, !tbaa !262, !alias.scope !264, !noalias !581
store i64 %120, i64* %188, align 8, !dbg !582, !tbaa !262, !alias.scope !264, !noalias !581
%value_phi37.fca.0.gep = getelementptr inbounds [1 x {} addrspace(10)*], [1 x {} addrspace(10)*]* %2, i64 0, i64 0, !dbg !561
store {} addrspace(10)* %nodecayed..pre-phi217, {} addrspace(10)** %value_phi37.fca.0.gep, align 8, !dbg !561, !noalias !583
%189 = addrspacecast [1 x {} addrspace(10)*]* %2 to [1 x {} addrspace(10)*] addrspace(11)*, !dbg !561
%190 = addrspacecast [2 x i64]* %1 to [2 x i64] addrspace(11)*, !dbg !561
call fastcc void @julia_throw_boundserror_47785([1 x {} addrspace(10)*] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) %189, [2 x i64] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(16) %190) #29, !dbg !561
unreachable, !dbg !561
L424.loopexit: ; preds = %load57
br label %L424, !dbg !584
L424: ; preds = %L424.loopexit, %load57.peel
%191 = add i64 %iv.next, 1, !dbg !584
%192 = icmp eq i64 %iv.next, %16, !dbg !588
br i1 %192, label %L451.loopexit, label %L252, !dbg !587
L451.loopexit: ; preds = %L424
br label %L451, !dbg !591
L451: ; preds = %L451.loopexit, %L227
switch i64 %18, label %L483 [
i64 0, label %L562
i64 1, label %L466
], !dbg !591
L466: ; preds = %L451
%193 = load i64, i64 addrspace(11)* %42, align 8, !dbg !601, !tbaa !126, !alias.scope !101, !noalias !128
%.not274 = icmp eq i64 %193, 0, !dbg !601
br i1 %.not274, label %L475, label %L478, !dbg !601
L475: ; preds = %L466
%194 = getelementptr inbounds [1 x i64], [1 x i64]* %4, i64 0, i64 0, !dbg !601
store i64 1, i64* %194, align 8, !dbg !601, !tbaa !262, !alias.scope !264, !noalias !581
%195 = addrspacecast [1 x i64]* %4 to [1 x i64] addrspace(11)*, !dbg !601
call fastcc void @julia_throw_boundserror_47773({} addrspace(10)* nofree noundef nonnull align 8 dereferenceable(32) %33, [1 x i64] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) %195) #29, !dbg !601
unreachable, !dbg !601
L478: ; preds = %L466
%.not275 = icmp sgt i64 %193, 0, !dbg !603
%196 = load i8*, i8* addrspace(11)* %31, align 8, !dbg !603, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%197 = ptrtoint i8* %196 to i64, !dbg !603
%198 = ptrtoint i8* %32 to i64, !dbg !603
%199 = sub i64 %198, %197, !dbg !603
%200 = shl nuw nsw i64 %193, 3, !dbg !603
%201 = icmp ult i64 %199, %200, !dbg !603
%202 = and i1 %.not275, %201, !dbg !603
br i1 %202, label %load73, label %oob69, !dbg !603
L483: ; preds = %L451
%203 = icmp sgt i64 %18, 15, !dbg !604
br i1 %203, label %L549, label %L487, !dbg !605
L487: ; preds = %L483
%204 = load i64, i64 addrspace(11)* %42, align 8, !dbg !606, !tbaa !126, !alias.scope !101, !noalias !128
%.not279 = icmp eq i64 %204, 0, !dbg !606
br i1 %.not279, label %L496, label %L499, !dbg !606
L496: ; preds = %L487
%205 = getelementptr inbounds [1 x i64], [1 x i64]* %7, i64 0, i64 0, !dbg !606
store i64 1, i64* %205, align 8, !dbg !606, !tbaa !262, !alias.scope !264, !noalias !581
%206 = addrspacecast [1 x i64]* %7 to [1 x i64] addrspace(11)*, !dbg !606
call fastcc void @julia_throw_boundserror_47773({} addrspace(10)* nofree noundef nonnull align 8 dereferenceable(32) %33, [1 x i64] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) %206) #29, !dbg !606
unreachable, !dbg !606
L499: ; preds = %L487
%207 = shl nuw i64 %204, 1, !dbg !608
%.not280 = icmp sgt i64 %204, 0, !dbg !608
%208 = bitcast i8* %32 to double*, !dbg !608
%209 = load i8*, i8* addrspace(11)* %31, align 8, !dbg !608, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%210 = ptrtoint i8* %209 to i64, !dbg !608
%211 = ptrtoint i8* %32 to i64, !dbg !608
%212 = sub i64 %211, %210, !dbg !608
%213 = shl nuw nsw i64 %204, 3, !dbg !608
%214 = icmp ult i64 %212, %213, !dbg !608
%215 = and i1 %.not280, %214, !dbg !608
br i1 %215, label %load79, label %oob75, !dbg !608
L514: ; preds = %load79
%216 = getelementptr inbounds [1 x i64], [1 x i64]* %6, i64 0, i64 0, !dbg !609
store i64 2, i64* %216, align 8, !dbg !609, !tbaa !262, !alias.scope !264, !noalias !581
%217 = addrspacecast [1 x i64]* %6 to [1 x i64] addrspace(11)*, !dbg !609
call fastcc void @julia_throw_boundserror_47773({} addrspace(10)* nofree noundef nonnull align 8 dereferenceable(32) %33, [1 x i64] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) %217) #29, !dbg !609
unreachable, !dbg !609
L517: ; preds = %load79
%218 = add nuw i64 %204, 1, !dbg !611
%.not284 = icmp ult i64 %218, %207, !dbg !611
%219 = getelementptr inbounds i8, i8* %32, i64 8, !dbg !611
%220 = ptrtoint i8* %219 to i64, !dbg !611
%221 = sub i64 %220, %210, !dbg !611
%222 = icmp ult i64 %221, %213, !dbg !611
%223 = and i1 %.not284, %222, !dbg !611
br i1 %223, label %load84, label %oob80, !dbg !611
L527: ; preds = %L527.preheader, %load91
%iv5 = phi i64 [ 0, %L527.preheader ], [ %iv.next6, %load91 ]
%value_phi85384 = phi double [ %372, %load91 ], [ %365, %L527.preheader ]
%224 = add nuw nsw i64 %iv5, 2, !dbg !612
%iv.next6 = add nuw nsw i64 %iv5, 1, !dbg !612
%225 = add nuw nsw i64 %224, 1, !dbg !612
%exitcond.not = icmp eq i64 %224, %204, !dbg !614
br i1 %exitcond.not, label %L539, label %L542, !dbg !614
L539: ; preds = %L527
%226 = getelementptr inbounds [1 x i64], [1 x i64]* %5, i64 0, i64 0, !dbg !614
store i64 %225, i64* %226, align 8, !dbg !614, !tbaa !262, !alias.scope !264, !noalias !581
%227 = addrspacecast [1 x i64]* %5 to [1 x i64] addrspace(11)*, !dbg !614
call fastcc void @julia_throw_boundserror_47773({} addrspace(10)* nofree noundef nonnull align 8 dereferenceable(32) %33, [1 x i64] addrspace(11)* nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) %227) #29, !dbg !614
unreachable, !dbg !614
L542: ; preds = %L527
%228 = add i64 %224, %204, !dbg !615
%.not290 = icmp ult i64 %228, %207, !dbg !615
%229 = getelementptr inbounds double, double* %208, i64 %224, !dbg !615
%230 = ptrtoint double* %229 to i64, !dbg !615
%231 = sub i64 %230, %210, !dbg !615
%232 = icmp ult i64 %231, %213, !dbg !615
%233 = and i1 %.not290, %232, !dbg !615
br i1 %233, label %load91, label %oob87, !dbg !615
L549: ; preds = %L483
%234 = call fastcc double @julia_mapreduce_impl_47760({} addrspace(10)* noundef nonnull align 8 dereferenceable(32) %33, i64 noundef signext 1, i64 signext %18) #27, !dbg !616
br label %L562, !dbg !618
L562.loopexit: ; preds = %load91
br label %L562, !dbg !600
L562: ; preds = %L562.loopexit, %load84, %load73, %L549, %L451
%value_phi68 = phi double [ %350, %load73 ], [ %234, %L549 ], [ 0.000000e+00, %L451 ], [ %365, %load84 ], [ %372, %L562.loopexit ]
ret double %value_phi68, !dbg !600
fail: ; preds = %L55
%235 = load {}*, {}** @jl_undefref_exception, align 8, !dbg !483, !tbaa !22, !alias.scope !39, !noalias !40, !nonnull !0
%236 = addrspacecast {}* %235 to {} addrspace(12)*, !dbg !483
call void @ijl_throw({} addrspace(12)* %236) #29, !dbg !483
unreachable, !dbg !483
guard_pass: ; preds = %L102
%237 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %29, i64 1, !dbg !499
%238 = bitcast { i64, {} addrspace(10)** } addrspace(11)* %237 to {} addrspace(10)* addrspace(11)*, !dbg !499
%239 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %238, align 8, !dbg !499, !tbaa !22, !alias.scope !39, !noalias !40
%240 = icmp eq {} addrspace(10)* %239, null, !dbg !499
%241 = select i1 %240, {} addrspace(10)* %value_phi8, {} addrspace(10)* %239, !dbg !499
%.pre = addrspacecast {} addrspace(10)* %241 to {} addrspace(11)*
br label %guard_exit, !dbg !499
guard_exit: ; preds = %guard_pass, %L102
%nodecayed..pre223.pre-phi = phi {} addrspace(10)* [ %241, %guard_pass ], [ %value_phi8, %L102 ]
%242 = addrspacecast {} addrspace(10)* %nodecayed..pre223.pre-phi to {} addrspace(11)*, !dbg !499
%243 = call "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @julia.typeof({} addrspace(10)* nonnull %nodecayed..pre223.pre-phi) #30, !dbg !499
%244 = addrspacecast {} addrspace(10)* %243 to {} addrspace(11)*, !dbg !499
%245 = call nonnull {}* @julia.pointer_from_objref({} addrspace(11)* %244) #31, !dbg !499
%.not235 = icmp eq {}* %245, inttoptr (i64 126279284195056 to {}*), !dbg !499
%spec.select = select i1 %.not235, {} addrspace(11)* %242, {} addrspace(11)* %40, !dbg !499
%246 = bitcast {} addrspace(11)* %spec.select to i8 addrspace(11)*, !dbg !619
%247 = getelementptr inbounds i8, i8 addrspace(11)* %246, i64 8, !dbg !619
%248 = bitcast i8 addrspace(11)* %247 to i64 addrspace(11)*, !dbg !621
%249 = load i64, i64 addrspace(11)* %248, align 8, !dbg !621, !tbaa !324, !alias.scope !101, !noalias !128
%250 = bitcast {} addrspace(10)* %0 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !622
%251 = addrspacecast { i8*, {} addrspace(10)* } addrspace(10)* %250 to { i8*, {} addrspace(10)* } addrspace(11)*, !dbg !622
%252 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %251, i64 0, i32 1, !dbg !622
%253 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %252, align 8, !dbg !622, !tbaa !98, !alias.scope !101, !noalias !128, !dereferenceable_or_null !237, !align !238
%254 = bitcast {} addrspace(10)* %253 to { i64, {} addrspace(10)** } addrspace(10)*, !dbg !499
%255 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %254 to { i64, {} addrspace(10)** } addrspace(11)*, !dbg !499
%256 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %255, i64 0, i32 1, !dbg !499
%257 = load {} addrspace(10)**, {} addrspace(10)** addrspace(11)* %256, align 8, !dbg !499, !tbaa !132, !alias.scope !101, !noalias !128, !nonnull !0
%258 = bitcast {} addrspace(10)* %253 to {} addrspace(10)* addrspace(10)*, !dbg !499
%259 = addrspacecast {} addrspace(10)* addrspace(10)* %258 to {} addrspace(10)* addrspace(11)*, !dbg !499
%260 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %259, i64 2, !dbg !499
%261 = addrspacecast {} addrspace(10)** %257 to {} addrspace(10)* addrspace(11)*, !dbg !499
%.not237 = icmp eq {} addrspace(10)* addrspace(11)* %260, %261, !dbg !499
br i1 %.not237, label %guard_exit15, label %guard_pass14, !dbg !499
guard_pass14: ; preds = %guard_exit
%262 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %255, i64 1, !dbg !499
%263 = bitcast { i64, {} addrspace(10)** } addrspace(11)* %262 to {} addrspace(10)* addrspace(11)*, !dbg !499
%264 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %263, align 8, !dbg !499, !tbaa !22, !alias.scope !39, !noalias !40
%265 = icmp eq {} addrspace(10)* %264, null, !dbg !499
%266 = select i1 %265, {} addrspace(10)* %253, {} addrspace(10)* %264, !dbg !499
br label %guard_exit15, !dbg !499
guard_exit15: ; preds = %guard_pass14, %guard_exit
%267 = phi {} addrspace(10)* [ %253, %guard_exit ], [ %266, %guard_pass14 ], !dbg !499
%268 = call "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @julia.typeof({} addrspace(10)* %267) #30, !dbg !499
%269 = addrspacecast {} addrspace(10)* %268 to {} addrspace(11)*, !dbg !499
%270 = call nonnull {}* @julia.pointer_from_objref({} addrspace(11)* %269) #31, !dbg !499
%.not238 = icmp eq {}* %270, inttoptr (i64 126279284195056 to {}*), !dbg !499
%spec.select302 = select i1 %.not238, {} addrspace(10)* %267, {} addrspace(10)* %253, !dbg !499
%271 = bitcast {} addrspace(10)* %spec.select302 to i8 addrspace(10)*, !dbg !619
%272 = addrspacecast i8 addrspace(10)* %271 to i8 addrspace(11)*, !dbg !619
%273 = getelementptr inbounds i8, i8 addrspace(11)* %272, i64 8, !dbg !619
%274 = bitcast i8 addrspace(11)* %273 to i64 addrspace(11)*, !dbg !621
%275 = load i64, i64 addrspace(11)* %274, align 8, !dbg !621, !tbaa !324, !alias.scope !101, !noalias !128
%276 = icmp eq i64 %249, %275, !dbg !623
%277 = bitcast i8 addrspace(10)* %12 to i64 addrspace(10)*, !dbg !491
br i1 %276, label %L138, label %L152, !dbg !491
guard_pass22: ; preds = %L172
%278 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %29, i64 1, !dbg !523
%279 = bitcast { i64, {} addrspace(10)** } addrspace(11)* %278 to {} addrspace(10)* addrspace(11)*, !dbg !523
%280 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %279, align 8, !dbg !523, !tbaa !22, !alias.scope !39, !noalias !40
%281 = icmp eq {} addrspace(10)* %280, null, !dbg !523
%282 = select i1 %281, {} addrspace(10)* %value_phi8, {} addrspace(10)* %280, !dbg !523
%.pre701 = addrspacecast {} addrspace(10)* %282 to {} addrspace(11)*
br label %guard_exit23, !dbg !523
guard_exit23: ; preds = %guard_pass22, %L172
%nodecayed..pre.pre-phi = phi {} addrspace(10)* [ %282, %guard_pass22 ], [ %value_phi8, %L172 ]
%283 = addrspacecast {} addrspace(10)* %nodecayed..pre.pre-phi to {} addrspace(11)*, !dbg !523
%284 = call "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @julia.typeof({} addrspace(10)* nonnull %nodecayed..pre.pre-phi) #30, !dbg !523
%285 = addrspacecast {} addrspace(10)* %284 to {} addrspace(11)*, !dbg !523
%286 = call nonnull {}* @julia.pointer_from_objref({} addrspace(11)* %285) #31, !dbg !523
%.not244 = icmp eq {}* %286, inttoptr (i64 126279284195056 to {}*), !dbg !523
%spec.select303 = select i1 %.not244, {} addrspace(11)* %283, {} addrspace(11)* %40, !dbg !523
%287 = bitcast {} addrspace(11)* %spec.select303 to i8 addrspace(11)*, !dbg !626
%288 = getelementptr inbounds i8, i8 addrspace(11)* %287, i64 8, !dbg !626
%289 = bitcast i8 addrspace(11)* %288 to i64 addrspace(11)*, !dbg !628
%290 = load i64, i64 addrspace(11)* %289, align 8, !dbg !628, !tbaa !324, !alias.scope !101, !noalias !128
%291 = bitcast {} addrspace(10)* %0 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !629
%292 = addrspacecast { i8*, {} addrspace(10)* } addrspace(10)* %291 to { i8*, {} addrspace(10)* } addrspace(11)*, !dbg !629
%293 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(11)* %292, i64 0, i32 1, !dbg !629
%294 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %293, align 8, !dbg !629, !tbaa !98, !alias.scope !101, !noalias !128, !dereferenceable_or_null !237, !align !238
%295 = bitcast {} addrspace(10)* %294 to { i64, {} addrspace(10)** } addrspace(10)*, !dbg !632
%296 = addrspacecast { i64, {} addrspace(10)** } addrspace(10)* %295 to { i64, {} addrspace(10)** } addrspace(11)*, !dbg !632
%297 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %296, i64 0, i32 1, !dbg !632
%298 = load {} addrspace(10)**, {} addrspace(10)** addrspace(11)* %297, align 8, !dbg !632, !tbaa !132, !alias.scope !101, !noalias !128, !nonnull !0
%299 = bitcast {} addrspace(10)* %294 to {} addrspace(10)* addrspace(10)*, !dbg !632
%300 = addrspacecast {} addrspace(10)* addrspace(10)* %299 to {} addrspace(10)* addrspace(11)*, !dbg !632
%301 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %300, i64 2, !dbg !632
%302 = addrspacecast {} addrspace(10)** %298 to {} addrspace(10)* addrspace(11)*, !dbg !632
%.not246 = icmp eq {} addrspace(10)* addrspace(11)* %301, %302, !dbg !632
br i1 %.not246, label %guard_exit28, label %guard_pass27, !dbg !632
guard_pass27: ; preds = %guard_exit23
%303 = getelementptr inbounds { i64, {} addrspace(10)** }, { i64, {} addrspace(10)** } addrspace(11)* %296, i64 1, !dbg !632
%304 = bitcast { i64, {} addrspace(10)** } addrspace(11)* %303 to {} addrspace(10)* addrspace(11)*, !dbg !632
%305 = load {} addrspace(10)*, {} addrspace(10)* addrspace(11)* %304, align 8, !dbg !632, !tbaa !22, !alias.scope !39, !noalias !40
%306 = icmp eq {} addrspace(10)* %305, null, !dbg !632
%307 = select i1 %306, {} addrspace(10)* %294, {} addrspace(10)* %305, !dbg !632
br label %guard_exit28, !dbg !632
guard_exit28: ; preds = %guard_pass27, %guard_exit23
%308 = phi {} addrspace(10)* [ %294, %guard_exit23 ], [ %307, %guard_pass27 ], !dbg !632
%309 = call "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @julia.typeof({} addrspace(10)* %308) #30, !dbg !632
%310 = addrspacecast {} addrspace(10)* %309 to {} addrspace(11)*, !dbg !632
%311 = call nonnull {}* @julia.pointer_from_objref({} addrspace(11)* %310) #31, !dbg !632
%.not247 = icmp eq {}* %311, inttoptr (i64 126279284195056 to {}*), !dbg !632
%spec.select304 = select i1 %.not247, {} addrspace(10)* %308, {} addrspace(10)* %294, !dbg !632
%312 = bitcast {} addrspace(10)* %spec.select304 to i8 addrspace(10)*, !dbg !633
%313 = addrspacecast i8 addrspace(10)* %312 to i8 addrspace(11)*, !dbg !633
%314 = getelementptr inbounds i8, i8 addrspace(11)* %313, i64 8, !dbg !633
%315 = bitcast i8 addrspace(11)* %314 to i64 addrspace(11)*, !dbg !635
%316 = load i64, i64 addrspace(11)* %315, align 8, !dbg !635, !tbaa !324, !alias.scope !101, !noalias !128
%317 = icmp eq i64 %290, %316, !dbg !636
br i1 %317, label %L209, label %L227, !dbg !511
oob.loopexit: ; preds = %L319
br label %oob, !dbg !554
oob.loopexit1: ; preds = %L252
br label %oob, !dbg !554
oob: ; preds = %oob.loopexit1, %oob.loopexit
%.lcssa532 = phi i64 [ %177, %oob.loopexit ], [ 1, %oob.loopexit1 ], !dbg !578
%318 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !554
%319 = bitcast {} addrspace(10)* %318 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !554
%.repack259 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %319, i64 0, i32 0, !dbg !554
store i8* %.pre913, i8* addrspace(10)* %.repack259, align 8, !dbg !554, !tbaa !41, !alias.scope !45, !noalias !478
%.repack260 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %319, i64 0, i32 1, !dbg !554
store {} addrspace(10)* %.pre914, {} addrspace(10)* addrspace(10)* %.repack260, align 8, !dbg !554, !tbaa !41, !alias.scope !45, !noalias !478
%320 = addrspacecast {} addrspace(10)* %318 to {} addrspace(12)*, !dbg !554
call void @ijl_bounds_error_int({} addrspace(12)* %320, i64 %.lcssa532) #29, !dbg !554
unreachable, !dbg !554
oob48: ; preds = %L365.peel
%321 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !566
%322 = bitcast {} addrspace(10)* %321 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !566
%.repack265 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %322, i64 0, i32 0, !dbg !566
store i8* %135, i8* addrspace(10)* %.repack265, align 8, !dbg !566, !tbaa !41, !alias.scope !45, !noalias !478
%.repack266 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %322, i64 0, i32 1, !dbg !566
store {} addrspace(10)* %136, {} addrspace(10)* addrspace(10)* %.repack266, align 8, !dbg !566, !tbaa !41, !alias.scope !45, !noalias !478
%323 = addrspacecast {} addrspace(10)* %321 to {} addrspace(12)*, !dbg !566
call void @ijl_bounds_error_int({} addrspace(12)* %323, i64 %120) #29, !dbg !566
unreachable, !dbg !566
oob51: ; preds = %idxend50.peel
%324 = addrspacecast {} addrspace(10)* %136 to {} addrspace(12)*, !dbg !566
call void @ijl_bounds_error_int({} addrspace(12)* noundef %324, i64 noundef 1) #29, !dbg !566
unreachable, !dbg !566
load52: ; preds = %L319
%325 = add i64 %iv.next4, %122, !dbg !572
%326 = add i64 %325, %162, !dbg !572
%.not268 = icmp ult i64 %326, %163, !dbg !572
%327 = getelementptr inbounds double, double* %112, i64 %325, !dbg !572
%328 = ptrtoint double* %327 to i64, !dbg !572
%329 = sub i64 %328, %167, !dbg !572
%330 = icmp ult i64 %329, %170, !dbg !572
%331 = and i1 %.not268, %330, !dbg !572
br i1 %331, label %load57, label %oob53.loopexit, !dbg !572
oob53.loopexit: ; preds = %load52
br label %oob53, !dbg !639
oob53.loopexit2: ; preds = %load52.peel
br label %oob53, !dbg !639
oob53: ; preds = %oob53.loopexit2, %oob53.loopexit
%.lcssa504 = phi i64 [ %176, %oob53.loopexit ], [ 1, %oob53.loopexit2 ], !dbg !575
%332 = add i64 %.lcssa504, %122, !dbg !639
%333 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !572
%334 = bitcast {} addrspace(10)* %333 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !572
%.repack269 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %334, i64 0, i32 0, !dbg !572
store i8* %32, i8* addrspace(10)* %.repack269, align 8, !dbg !572, !tbaa !41, !alias.scope !45, !noalias !478
%.repack270 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %334, i64 0, i32 1, !dbg !572
store {} addrspace(10)* %value_phi8, {} addrspace(10)* addrspace(10)* %.repack270, align 8, !dbg !572, !tbaa !41, !alias.scope !45, !noalias !478
%335 = addrspacecast {} addrspace(10)* %333 to {} addrspace(12)*, !dbg !572
call void @ijl_bounds_error_int({} addrspace(12)* %335, i64 %332) #29, !dbg !572
unreachable, !dbg !572
oob56: ; preds = %idxend55.peel
%336 = addrspacecast {} addrspace(10)* %value_phi8 to {} addrspace(12)*, !dbg !572
call void @ijl_bounds_error_int({} addrspace(12)* %336, i64 noundef 1) #29, !dbg !572
unreachable, !dbg !572
load57: ; preds = %load52
%337 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(13)* %132, i64 %178, !dbg !554
%338 = bitcast {} addrspace(10)* addrspace(13)* %337 to double addrspace(13)*, !dbg !554
%339 = load double, double addrspace(13)* %338, align 8, !dbg !554, !tbaa !213, !alias.scope !45, !noalias !215
%340 = load double, double addrspace(13)* %159, align 8, !dbg !566, !tbaa !213, !alias.scope !45, !noalias !215
%341 = fmul double %339, %340, !dbg !569
%342 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(13)* %114, i64 %325, !dbg !572
%343 = bitcast {} addrspace(10)* addrspace(13)* %342 to double addrspace(13)*, !dbg !572
store double %341, double addrspace(13)* %343, align 8, !dbg !572, !tbaa !213, !alias.scope !45, !noalias !478
%exitcond629.not = icmp eq i64 %176, %16, !dbg !645
br i1 %exitcond629.not, label %L424.loopexit, label %L304, !dbg !574, !llvm.loop !646
oob69: ; preds = %L478
%344 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !603
%345 = bitcast {} addrspace(10)* %344 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !603
%.repack276 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %345, i64 0, i32 0, !dbg !603
store i8* %32, i8* addrspace(10)* %.repack276, align 8, !dbg !603, !tbaa !41, !alias.scope !45, !noalias !478
%.repack277 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %345, i64 0, i32 1, !dbg !603
store {} addrspace(10)* %value_phi8, {} addrspace(10)* addrspace(10)* %.repack277, align 8, !dbg !603, !tbaa !41, !alias.scope !45, !noalias !478
%346 = addrspacecast {} addrspace(10)* %344 to {} addrspace(12)*, !dbg !603
call void @ijl_bounds_error_int({} addrspace(12)* %346, i64 noundef 1) #29, !dbg !603
unreachable, !dbg !603
load73: ; preds = %L478
%347 = bitcast i8* %32 to {} addrspace(10)**, !dbg !603
%348 = call {} addrspace(10)* addrspace(13)* @julia.gc_loaded({} addrspace(10)* noundef %value_phi8, {} addrspace(10)** noundef %347) #27, !dbg !603
%349 = bitcast {} addrspace(10)* addrspace(13)* %348 to double addrspace(13)*, !dbg !603
%350 = load double, double addrspace(13)* %349, align 8, !dbg !603, !tbaa !213, !alias.scope !45, !noalias !215
br label %L562, !dbg !618
oob75: ; preds = %L499
%351 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !608
%352 = bitcast {} addrspace(10)* %351 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !608
%.repack281 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %352, i64 0, i32 0, !dbg !608
store i8* %32, i8* addrspace(10)* %.repack281, align 8, !dbg !608, !tbaa !41, !alias.scope !45, !noalias !478
%.repack282 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %352, i64 0, i32 1, !dbg !608
store {} addrspace(10)* %value_phi8, {} addrspace(10)* addrspace(10)* %.repack282, align 8, !dbg !608, !tbaa !41, !alias.scope !45, !noalias !478
%353 = addrspacecast {} addrspace(10)* %351 to {} addrspace(12)*, !dbg !608
call void @ijl_bounds_error_int({} addrspace(12)* %353, i64 noundef 1) #29, !dbg !608
unreachable, !dbg !608
load79: ; preds = %L499
%354 = bitcast i8* %32 to {} addrspace(10)**, !dbg !608
%355 = call {} addrspace(10)* addrspace(13)* @julia.gc_loaded({} addrspace(10)* noundef %value_phi8, {} addrspace(10)** noundef %354) #27, !dbg !608
%356 = bitcast {} addrspace(10)* addrspace(13)* %355 to double addrspace(13)*, !dbg !608
%357 = load double, double addrspace(13)* %356, align 8, !dbg !608, !tbaa !213, !alias.scope !45, !noalias !215
%358 = icmp ult i64 %204, 2, !dbg !609
br i1 %358, label %L514, label %L517, !dbg !609
oob80: ; preds = %L517
%359 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !611
%360 = bitcast {} addrspace(10)* %359 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !611
%.repack285 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %360, i64 0, i32 0, !dbg !611
store i8* %32, i8* addrspace(10)* %.repack285, align 8, !dbg !611, !tbaa !41, !alias.scope !45, !noalias !478
%.repack286 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %360, i64 0, i32 1, !dbg !611
store {} addrspace(10)* %value_phi8, {} addrspace(10)* addrspace(10)* %.repack286, align 8, !dbg !611, !tbaa !41, !alias.scope !45, !noalias !478
%361 = addrspacecast {} addrspace(10)* %359 to {} addrspace(12)*, !dbg !611
call void @ijl_bounds_error_int({} addrspace(12)* %361, i64 noundef 2) #29, !dbg !611
unreachable, !dbg !611
load84: ; preds = %L517
%362 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(13)* %355, i64 1, !dbg !611
%363 = bitcast {} addrspace(10)* addrspace(13)* %362 to double addrspace(13)*, !dbg !611
%364 = load double, double addrspace(13)* %363, align 8, !dbg !611, !tbaa !213, !alias.scope !45, !noalias !215
%365 = fadd double %357, %364, !dbg !647
%.not288383 = icmp sgt i64 %18, 2, !dbg !650
br i1 %.not288383, label %L527.preheader, label %L562, !dbg !651
L527.preheader: ; preds = %load84
br label %L527, !dbg !614
oob87: ; preds = %L542
%366 = call noalias nonnull align 8 dereferenceable(16) "enzyme_type"="{[-1]:Pointer, [-1,-1]:Pointer, [-1,8,0]:Integer, [-1,8,1]:Integer, [-1,8,2]:Integer, [-1,8,3]:Integer, [-1,8,4]:Integer, [-1,8,5]:Integer, [-1,8,6]:Integer, [-1,8,7]:Integer, [-1,8,8]:Pointer, [-1,8,8,-1]:Float@double}" {} addrspace(10)* @julia.gc_alloc_obj({}* nonnull %8, i64 noundef 16, {} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 126279284194960 to {}*) to {} addrspace(10)*)) #28, !dbg !615
%367 = bitcast {} addrspace(10)* %366 to { i8*, {} addrspace(10)* } addrspace(10)*, !dbg !615
%.repack291 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %367, i64 0, i32 0, !dbg !615
store i8* %32, i8* addrspace(10)* %.repack291, align 8, !dbg !615, !tbaa !41, !alias.scope !45, !noalias !478
%.repack292 = getelementptr inbounds { i8*, {} addrspace(10)* }, { i8*, {} addrspace(10)* } addrspace(10)* %367, i64 0, i32 1, !dbg !615
store {} addrspace(10)* %value_phi8, {} addrspace(10)* addrspace(10)* %.repack292, align 8, !dbg !615, !tbaa !41, !alias.scope !45, !noalias !478
%368 = addrspacecast {} addrspace(10)* %366 to {} addrspace(12)*, !dbg !615
call void @ijl_bounds_error_int({} addrspace(12)* %368, i64 %225) #29, !dbg !615
unreachable, !dbg !615
load91: ; preds = %L542
%369 = getelementptr inbounds {} addrspace(10)*, {} addrspace(10)* addrspace(13)* %355, i64 %224, !dbg !615
%370 = bitcast {} addrspace(10)* addrspace(13)* %369 to double addrspace(13)*, !dbg !615
%371 = load double, double addrspace(13)* %370, align 8, !dbg !615, !tbaa !213, !alias.scope !45, !noalias !215
%372 = fadd double %value_phi85384, %371, !dbg !652
%exitcond628.not = icmp eq i64 %225, %18, !dbg !650
br i1 %exitcond628.not, label %L562.loopexit, label %L527, !dbg !651
}
Did not have return index set when differentiating function
call %57 = call nonnull "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* %56({} addrspace(10)* %253, i64 %55, i64 %16) #27, !dbg !130
augmentcall %_augmented63 = call { i8*, {} addrspace(10)* } %80({} addrspace(10)* %279, {} addrspace(10)* %"'il_phi28", i64 %75, i64 %"'ipc62", i64 %18, i64 %18), !dbg !164
Stacktrace:
[1] copy
@ ./array.jl:350
[2] unaliascopy
@ ./abstractarray.jl:1516
[3] unalias
@ ./abstractarray.jl:1500
[4] broadcast_unalias
@ ./broadcast.jl:941
[5] preprocess
@ ./broadcast.jl:948
[6] preprocess_args
@ ./broadcast.jl:950
[7] preprocess
@ ./broadcast.jl:947
[8] copyto!
@ ./broadcast.jl:964
[9] copyto!
@ ./broadcast.jl:920
[10] copy
@ ./broadcast.jl:892
[11] materialize
@ ./broadcast.jl:867
[12] f
@ ~/Work/GitHub/Julia/DifferentiationInterface.jl/DifferentiationInterface/test/playground.jl:16
Stacktrace:
[1] julia_error(cstr::Cstring, val::Ptr{…}, errtype::Enzyme.API.ErrorType, data::Ptr{…}, data2::Ptr{…}, B::Ptr{…})
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:2294
[2] EnzymeCreatePrimalAndGradient(logic::Enzyme.Logic, todiff::LLVM.Function, retType::Enzyme.API.CDIFFE_TYPE, constant_args::Vector{…}, TA::Enzyme.TypeAnalysis, returnValue::Bool, dretUsed::Bool, mode::Enzyme.API.CDerivativeMode, runtimeActivity::Bool, width::Int64, additionalArg::Ptr{…}, forceAnonymousTape::Bool, typeInfo::Enzyme.FnTypeInfo, uncacheable_args::Vector{…}, augmented::Ptr{…}, atomicAdd::Bool)
@ Enzyme.API ~/.julia/packages/Enzyme/Vjlrr/src/api.jl:253
[3] enzyme!(job::GPUCompiler.CompilerJob{…}, mod::LLVM.Module, primalf::LLVM.Function, TT::Type, mode::Enzyme.API.CDerivativeMode, width::Int64, parallel::Bool, actualRetType::Type, wrap::Bool, modifiedBetween::Tuple{…}, returnPrimal::Bool, expectedTapeType::Type, loweredArgs::Set{…}, boxedArgs::Set{…})
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:4706
[4] codegen(output::Symbol, job::GPUCompiler.CompilerJob{…}; libraries::Bool, deferred_codegen::Bool, optimize::Bool, toplevel::Bool, strip::Bool, validate::Bool, only_entry::Bool, parent_job::Nothing)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:7801
[5] codegen
@ ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:6638 [inlined]
[6] _thunk(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, postopt::Bool)
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:8909
[7] _thunk
@ ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:8909 [inlined]
[8] cached_compilation
@ ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:8950 [inlined]
[9] thunkbase(ctx::LLVM.Context, mi::Core.MethodInstance, ::Val{…}, ::Type{…}, ::Type{…}, tt::Type{…}, ::Val{…}, ::Val{…}, ::Val{…}, ::Val{…}, ::Val{…}, ::Type{…}, ::Val{…}, ::Val{…})
@ Enzyme.Compiler ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:9082
[10] #s2067#19118
@ ~/.julia/packages/Enzyme/Vjlrr/src/compiler.jl:9219 [inlined]
[11]
@ Enzyme.Compiler ./none:0
[12] (::Core.GeneratedFunctionStub)(::UInt64, ::LineNumberNode, ::Any, ::Vararg{Any})
@ Core ./boot.jl:707
[13] autodiff
@ ~/.julia/packages/Enzyme/Vjlrr/src/Enzyme.jl:473 [inlined]
[14] autodiff
@ ~/.julia/packages/Enzyme/Vjlrr/src/Enzyme.jl:512 [inlined]
[15] macro expansion
@ ~/.julia/packages/Enzyme/Vjlrr/src/Enzyme.jl:1705 [inlined]
[16] gradient(::EnzymeCore.ReverseMode{false, false, EnzymeCore.FFIABI, false, false}, ::typeof(f), ::Vector{Float64})
@ Enzyme ~/.julia/packages/Enzyme/Vjlrr/src/Enzyme.jl:1646
[17] top-level scope
@ ~/Work/GitHub/Julia/DifferentiationInterface.jl/DifferentiationInterface/test/playground.jl:18
Some type information was truncated. Use `show(err)` to see complete types. |
|
In case it helps, on Enzyme v0.13.10 the error has changed for the MWE above: julia> Enzyme.gradient(Enzyme.Reverse, f, [1.0])
ERROR:
No augmented forward pass found for jl_alloc_genericmemory
at context: %20 = call "enzyme_type"="{[-1]:Pointer}" {} addrspace(10)* @jl_alloc_genericmemory({} addrspace(10)* noundef addrspacecast ({}* inttoptr (i64 124954544733936 to {}*) to {} addrspace(10)*), i64 %11) #14, !dbg !81
Stacktrace:
[1] GenericMemory
@ ./boot.jl:516
[2] new_as_memoryref
@ ./boot.jl:535
[3] Array
@ ./boot.jl:582
[4] Array
@ ./boot.jl:592
[5] Array
@ ./boot.jl:599
[6] similar
@ ./abstractarray.jl:868
[7] similar
@ ./abstractarray.jl:867
[8] similar
@ ./broadcast.jl:224
[9] similar
@ ./broadcast.jl:223
[10] copy
@ ./broadcast.jl:892
[11] materialize
@ ./broadcast.jl:867
[12] f
@ ./REPL[4]:1
Stacktrace:
[1] GenericMemory
@ ./boot.jl:516 [inlined]
[2] new_as_memoryref
@ ./boot.jl:535 [inlined]
[3] Array
@ ./boot.jl:582 [inlined]
[4] Array
@ ./boot.jl:592 [inlined]
[5] Array
@ ./boot.jl:599 [inlined]
[6] similar
@ ./abstractarray.jl:868 [inlined]
[7] similar
@ ./abstractarray.jl:867 [inlined]
[8] similar
@ ./broadcast.jl:224 [inlined]
[9] similar
@ ./broadcast.jl:223 [inlined]
[10] copy
@ ./broadcast.jl:892 [inlined]
[11] materialize
@ ./broadcast.jl:867 [inlined]
[12] f
@ ./REPL[4]:1 [inlined]
[13] diffejulia_f_14701wrap
@ ./REPL[4]:0
[14] macro expansion
@ ~/.julia/packages/Enzyme/RmraO/src/compiler.jl:8572 [inlined]
[15] enzyme_call
@ ~/.julia/packages/Enzyme/RmraO/src/compiler.jl:8138 [inlined]
[16] CombinedAdjointThunk
@ ~/.julia/packages/Enzyme/RmraO/src/compiler.jl:7911 [inlined]
[17] autodiff
@ ~/.julia/packages/Enzyme/RmraO/src/Enzyme.jl:491 [inlined]
[18] autodiff
@ ~/.julia/packages/Enzyme/RmraO/src/Enzyme.jl:512 [inlined]
[19] macro expansion
@ ~/.julia/packages/Enzyme/RmraO/src/Enzyme.jl:1719 [inlined]
[20] gradient(::ReverseMode{false, false, FFIABI, false, false}, ::typeof(f), ::Vector{Float64})
@ Enzyme ~/.julia/packages/Enzyme/RmraO/src/Enzyme.jl:1660
[21] top-level scope
@ REPL[5]:1 |
|
On Enzyme v0.13.11 this MWE works!!! Congrats! julia> Enzyme.gradient(Enzyme.Reverse, f, [1.0])
([2.0],) |
|
Slowly tracking the progress on 1.11 on my DI test suite, and now I'm encountering this bug. Do tell me if me creating MWEs here is annoying, I'll stop ^^ julia> using Enzyme
julia> f(x) = vcat(x, x)
f (generic function with 1 method)
julia> autodiff(Forward, f, Duplicated(rand(2), zeros(2)))
ERROR: Enzyme execution failed.
Enzyme: The original primal code hits this error condition, thus differentiating it does not make sense
Stacktrace:
[1] unsafe_copyto!
@ ./genericmemory.jl:117 [inlined]
[2] unsafe_copyto!
@ ./array.jl:284 [inlined]
[3] vcat
@ ./array.jl:2222
[4] f
@ ./REPL[8]:1 [inlined]
[5] fwddiffejulia_f_27066wrap
@ ./REPL[8]:0
[6] macro expansion
@ ~/.julia/packages/Enzyme/vgArw/src/compiler.jl:8136 [inlined]
[7] enzyme_call
@ ~/.julia/packages/Enzyme/vgArw/src/compiler.jl:7702 [inlined]
[8] ForwardModeThunk
@ ~/.julia/packages/Enzyme/vgArw/src/compiler.jl:7491 [inlined]
[9] autodiff
@ ~/.julia/packages/Enzyme/vgArw/src/Enzyme.jl:647 [inlined]
[10] autodiff
@ ~/.julia/packages/Enzyme/vgArw/src/Enzyme.jl:537 [inlined]
[11] autodiff(mode::ForwardMode{false, FFIABI, false, false}, f::typeof(f), args::Duplicated{Vector{Float64}})
@ Enzyme ~/.julia/packages/Enzyme/vgArw/src/Enzyme.jl:504
[12] top-level scope
@ REPL[9]:1 |
|
with a new patch release, i think your most recent error should be fixed |
|
Thanks! The DI test suite was able to progress a little further with Enzyme v0.13.12. Here is the first error message I get now (see CI log): Does it sound familiar enough to be fixed without an MWE? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I wanted to use split reverse mode in order to compute pullbacks with array outputs. The only doc I found is
https://enzymead.github.io/Enzyme.jl/stable/api/#EnzymeCore.autodiff_thunk-Union{Tuple{RABI},%20Tuple{ModifiedBetweenT},%20Tuple{Width},%20Tuple{ReturnShadow},%20Tuple{ReturnPrimal},%20Tuple{A},%20Tuple{FA},%20Tuple{EnzymeCore.ReverseModeSplit{ReturnPrimal,%20ReturnShadow,%20Width,%20ModifiedBetweenT,%20RABI},%20Type{FA},%20Type{A},%20Vararg{Any}}}%20where%20{FA%3C:Annotation,%20A%3C:Annotation,%20ReturnPrimal,%20ReturnShadow,%20Width,%20ModifiedBetweenT,%20RABI%3C:EnzymeCore.ABI}
so I tried it but it fails on Julia 1.11:
The text was updated successfully, but these errors were encountered: