Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

LLVM failed verification in custom batch rule #1075

Closed
ChrisRackauckas opened this issue Sep 24, 2023 · 2 comments · Fixed by SciML/LinearSolve.jl#405
Closed

LLVM failed verification in custom batch rule #1075

ChrisRackauckas opened this issue Sep 24, 2023 · 2 comments · Fixed by SciML/LinearSolve.jl#405

Comments

@ChrisRackauckas
Copy link
Contributor

MWE:

using Enzyme, ForwardDiff
using LinearSolve, LinearAlgebra, Test

n = 4
A = rand(n, n);
dA = zeros(n, n);
dA2 = zeros(n, n);
b1 = rand(n);
db1 = zeros(n);
db12 = zeros(n);

function fbatch(y, A, b1; alg = LUFactorization())
    prob = LinearProblem(A, b1)
    sol1 = solve(prob, alg)
    s1 = sol1.u
    y[1] = norm(s1)
    nothing
end

y = [0.0]
dy1 = [1.0]
dy2 = [1.0]
Enzyme.autodiff(Reverse, fbatch, BatchDuplicated(y, (dy1, dy2)), BatchDuplicated(copy(A), (dA, dA2)), BatchDuplicated(copy(b1), (db1, db12)))

dA2 = ForwardDiff.gradient(x->f(x,eltype(x).(b1)), copy(A))
db12 = ForwardDiff.gradient(x->f(eltype(x).(A),x), copy(b1))

@test_broken dA  dA_2
@test_broken dA2  dA_2
@test_broken db1  db1_2
@test_broken db12  db1_2

output3.txt

Bottom is:

ERROR: LLVM error: function failed verification (4)
Stacktrace:
  [1] handle_error(reason::Cstring)
    @ LLVM C:\Users\accou\.julia\packages\LLVM\lq6lJ\src\core\context.jl:134
  [2] EnzymeCreatePrimalAndGradient(logic::Enzyme.Logic, todiff::LLVM.Function, retType::Enzyme.API.CDIFFE_TYPE, constant_args::Vector{Enzyme.API.CDIFFE_TYPE}, TA::Enzyme.TypeAnalysis, returnValue::Bool, dretUsed::Bool, mode::Enzyme.API.CDerivativeMode, width::Int64, additionalArg::Ptr{Nothing}, forceAnonymousTape::Bool, typeInfo::Enzyme.FnTypeInfo, uncacheable_args::Vector{Bool}, augmented::Ptr{Nothing}, atomicAdd::Bool)
    @ Enzyme.API C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\api.jl:128
  [3] enzyme!(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, mod::LLVM.Module, primalf::LLVM.Function, TT::Type, mode::Enzyme.API.CDerivativeMode, width::Int64, parallel::Bool, actualRetType::Type, wrap::Bool, modifiedBetween::NTuple{4, Bool}, returnPrimal::Bool, jlrules::Vector{String}, expectedTapeType::Type)
    @ Enzyme.Compiler C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:7714
  [4] 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 C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:9274
  [5] codegen
    @ C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:8879 [inlined]
  [6] _thunk(job::GPUCompiler.CompilerJob{Enzyme.Compiler.EnzymeTarget, Enzyme.Compiler.EnzymeCompilerParams}, postopt::Bool)
    @ Enzyme.Compiler C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:9826
  [7] _thunk
    @ C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:9826 [inlined]
  [8] cached_compilation
    @ C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:9860 [inlined]
  [9] (::Enzyme.Compiler.var"#475#476"{DataType, DataType, DataType, Enzyme.API.CDerivativeMode, NTuple{4, Bool}, Int64, Bool, Bool, UInt64, DataType})(ctx::LLVM.Context)
    @ Enzyme.Compiler C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:9923
 [10] JuliaContext(f::Enzyme.Compiler.var"#475#476"{DataType, DataType, DataType, Enzyme.API.CDerivativeMode, NTuple{4, Bool}, Int64, Bool, Bool, UInt64, DataType})
    @ GPUCompiler C:\Users\accou\.julia\packages\GPUCompiler\2mJjc\src\driver.jl:47
 [11] #s292#474
    @ C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\compiler.jl:9878 [inlined]
 [12] var"#s292#474"(FA::Any, A::Any, TT::Any, Mode::Any, ModifiedBetween::Any, width::Any, ReturnPrimal::Any, ShadowInit::Any, World::Any, ABI::Any, ::Any, #unused#::Type, #unused#::Type, #unused#::Type, tt::Any, #unused#::Type, #unused#::Type, #unused#::Type, #unused#::Type, #unused#::Type, #unused#::Any)
    @ Enzyme.Compiler .\none:0
 [13] (::Core.GeneratedFunctionStub)(::Any, ::Vararg{Any})
    @ Core .\boot.jl:602
 [14] autodiff(::ReverseMode{false, FFIABI}, ::Const{typeof(fbatch)}, ::Type{Const{Nothing}}, ::BatchDuplicated{Vector{Float64}, 2}, ::Vararg{Any})
    @ Enzyme C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\Enzyme.jl:207
 [15] autodiff(::ReverseMode{false, FFIABI}, ::Const{typeof(fbatch)}, ::BatchDuplicated{Vector{Float64}, 2}, ::BatchDuplicated{Matrix{Float64}, 2}, ::Vararg{Any})
    @ Enzyme C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\Enzyme.jl:236
 [16] autodiff(::ReverseMode{false, FFIABI}, ::typeof(fbatch), ::BatchDuplicated{Vector{Float64}, 2}, ::BatchDuplicated{Matrix{Float64}, 2}, ::Vararg{Any})
    @ Enzyme C:\Users\accou\.julia\packages\Enzyme\VS5jo\src\Enzyme.jl:222
 [17] top-level scope
    @ c:\Users\accou\OneDrive\Computer\Desktop\test.jl:141
@ChrisRackauckas
Copy link
Contributor Author

Discovered as part of SciML/LinearSolve.jl#377.

@wsmoses
Copy link
Member

wsmoses commented Oct 27, 2023

Can you retry, I think this was fixed by #1083

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants