-
Found in https://github.com/SciML/LinearSolve.jl/pull/377, the Krylov methods have their own issue. MWE:
```julia
using Enzyme, ForwardDiff
using LinearSolve, LinearAlgebra, Test
n = 4
A = ra…
-
Gridap is a great example of code that can only be differentiated by Enzyme. The following is a snippet that can be used as a test case. It currently fails. This isn't minimal at all but it would be g…
-
Hi there, I have just recently learned about the Enzyme project. I have read through some of the presentations and thought that it could be a very promising tool for my use case. One of the bits that …
-
**Description**
The code shown below will crash at runtime with a SIGABRT / double free.
**Steps to reproduce**
Paste the code below into a single file, compile, and run:
```swift
import _Dif…
-
I just checked again the model in https://github.com/TuringLang/Turing.jl/pull/1887#issue-1389132949 on that branch, and it once again (after being fixed in #457) segfaults after emitting warnings. Be…
-
My group is getting started trying out Enzyme, and I noticed some unintuitive behavior on square roots: the derivative of `sqrt` at zero seems to be specially computed as zero. `__enzyme_autodiff(my_s…
-
An example of a **Recursive** Neural Network on Enzyme (state-of-the-art LLVM Auto Diff).
```
using Enzyme
abstract type Node end
struct Leaf first
return y, n.x
end
struct Branch f…
-
Hi,
I think there has been a regression on the issue reported in https://github.com/EnzymeAD/Enzyme.jl/issues/647 on differentiating Bessel functions.
I tried both forward-mode and reverse-mode AD…
-
On the following reimplementation of `dot`, the defined reverse-mode rule works fine for real inputs but raises an `ERROR: LLVM error: function failed verification (4)` for complex inputs:
```julia…
-
With Enzyme v0.10.15 on Julia 1.8.5, I get
```julia
No augmented forward pass found for Faddeeva_erfcx_re
declare double @Faddeeva_erfcx_re(double) local_unnamed_addr
```
on a function.
If thi…