Edoardo Pasca
Edoardo Pasca
Let's actually get rid of this function and break backward compatibility, which we will actually break anyway with the new SIRF version.
@KrisThielemans any opinion?
[Now ](https://github.com/gadgetron/gadgetron/issues/925#issuecomment-763676235)the situation is even worse with Gadgetron requiring GCC 9.
Maybe it'd be good to build Gadgetron indipendently?
The error `cicc: not found` seems to point to an [issue with the installation](https://forums.developer.nvidia.com/t/when-i-arch-option-error-sh-cicc-command-not-found-takes-place/31753) of the SDK? Currently, by https://github.com/TomographicImaging/CCPi-Regularisation-Toolkit/pull/216 there's a refactoring of the CMake CUDA build in the...
I couldn't find any documentation about breaking a prange, did you?
The following code seems to suggest that `break` in a `prange` creates a multiple exit scenario: ```python import numba as nb import numpy as np @nb.njit(parallel=True) def break_prange(x, n): accumulator...
Thanks @samdporter, it looks like the `direct` method of `BlockOperator` is the culprit as it always returns a `BlockDataContainer`. https://github.com/TomographicImaging/CIL/blob/45e0b235ac79efc03dc5cb46b8ddda9d16f6331b/Wrappers/Python/cil/optimisation/operators/BlockOperator.py#L181 We could fix it as done in `adjoint` where in...
closed by #1802
> * As we do for the `TotalVariation` function, we have to extract the `domain` from a `DataContainer` and then use it for the GradientOperator. This is not correct. See...