Lukas Burgholzer

Just rebased this PR on top of main and cleaned up the commits a little bit. This should be a solid start once there is more time to look into...

Ok. Further progress here. To keep the history of this PR, I just squashed the previous commits and added another commit that reverts the respective changes. It turns out, that...

Adding some interesting articles on C++ ABI compatibility here: - https://uwekorn.com/2019/09/15/how-we-build-apache-arrows-manylinux-wheels.html - https://discuss.python.org/t/how-to-set-glibcxx-use-cxx11-abi-for-manylinux2014-and-manylinux2010-wheels/10551

The Python errors on Windows are interesting as I can't quite make sense of them. This needs some further investigation. It has something to do with the `mqt-core-python` target. So...

Alright. The C++ tests working under Windows were simply because no shared libraries were built there. The reason for the failure in the Python module was symbol visibility, which is...

Closing in favor of #662

It turns out that this issue is not limited to the matrix DD case and its not solely the matrix normalization that is at fault. A similar phenomenon also happens...

Ok. Just noticed, that this might, in fact, be only due to the matrix normalization. The following circuit works smoothly ```qasm OPENQASM 2.0; include "qelib1.inc"; qreg q[81]; qreg k[1]; h...

As for why the error already shows at $82$ qubits: $$\frac{1}{\sqrt{2}}^{81} - \frac{1}{\sqrt{2}}^{82} = 1.884e^{-13} < 2.274e^{-13} = 2^{-42} = \varepsilon$$ Thus, the lookup of the top edge weight for...

Closing for now. Please feal free to reopen if you plan to continue working on this.