Valentin Churavy
Valentin Churavy
Ok `0.10.3` will return an error here instead of crashing
@wsmoses - https://github.com/EnzymeAD/Enzyme.jl/blob/9f456aa8f691ac2fa1e624cefa110c4b53a1e6d2/test/runtests.jl#L1019-L1025 - https://github.com/EnzymeAD/Enzyme.jl/blob/9f456aa8f691ac2fa1e624cefa110c4b53a1e6d2/test/runtests.jl#L575 Both also hit this case were they hit a null-pointer, but our tests indicated it was correct.
Box is dead, long live sret
@wsmoses no idea what this was needed for. So I rescued the code to a branch xD
Fails with 0.9.1 locally with: ``` double, [0,56]:Pointer, [0,56,0]:Float@double} %21 = fmul double %14, %18, !dbg !40 julia: /workspace/srcdir/Enzyme/enzyme/Enzyme/GradientUtils.cpp:4246: llvm::Value* GradientUtils::invertPointerM(llvm::Value*, llvm::IRBuilder&, bool): Assertion `arg->getType()->isIntOrIntVectorTy()' failed. signal (6): Aborted in...
We will need forward mode variants for `enzyme_code_llvm`.
@wsmoses this is fixed by #408 but prints a ridiculously large type-tree.
@michel2323 can you add a test that checks that the values are what you expect?
@wsmoses was your comment in https://github.com/EnzymeAD/Enzyme.jl/issues/347#issuecomment-1212435054 w.r.t this issue?
On https://github.com/EnzymeAD/Enzyme.jl/pull/408 I see: ``` vchuravy@odin ~/s/Enzyme (vc/anonymous)> julia --project test.jl dx = [3.0, 2.0] ERROR: LoadError: TypeError: signal (11): Segmentation fault in expression starting at none:0 ijl_subtype_env at /cache/build/default-amdci5-5/julialang/julia-release-1-dot-8/src/subtype.c:1857...