Nicola

Julia Version 1.7.2 Commit bf53498635 (2022-02-06 15:21 UTC) Platform Info: OS: Windows (x86_64-w64-mingw32) CPU: Intel(R) Core(TM) i7-8550U CPU @ 1.80GHz WORD_SIZE: 64 LIBM: libopenlibm LLVM: libLLVM-12.0.1 (ORCJIT, skylake)

```julia ERROR: Enzyme compilation failed. Current scope: ; ModuleID = 'text' source_filename = "text" target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128-ni:10:11:12:13" target triple = "x86_64-w64-mingw32" ; Function Attrs: readnone declare {}*** @julia.get_pgcstack() local_unnamed_addr...

In my package called [FinancialToolbox ](https://github.com/rcalxrc08/FinancialToolbox.jl)there is a function called blsimpv. It is nothing else then the inverse of Black Scholes Formula (called blsprice in the package) with respect to...

Sorry I missed this, I think that is reasonable to raise an error in such situation since the problem is not the convergence of the method but the actual existence...

I will add a more appropriate check since the function to invert is monotone.

Would you mind to provide an example where it happens that the algorithm is throwing? I would like to switch to Roots.jl for Brent instead of using my own, but...

solved partially for Dual and HyperDual Numbers, no support for mixed derivatives yet. No support possible for complex in my knowledge

Support for dual is broken too unfortunately. When you have multiple dual inputs the output is wrong.

Added proper support for DualNumbers.jl. Support for the rest of AD packages is there but it isn't efficient. I was not able to avoid to run successive solvers.

This is stopping me from having a C^inf representation (wrt to everything) for implied volatility of more complex models.