Miles
Miles
Thanks for that helpful clarification. After thinking about it further, I believe that definition of `tol` is actually the most sensible one. Our code inside C++ ITensor happened to make...
Thanks, this looks like an interesting PR but some aspects of it aren't obvious to me. Could you please describe the changes at a high level? Also I made some...
Thanks - sounds good. We can consider having additional dependencies such as Thrust but by default these should be turned off or not required if the CUDA backend is not...
Hi, thanks for contributing this and sorry it's been so long. Part of the issue is that while this functionality is nice to have, I wasn't sure about maintaining it...
For a test case, if one removes `randomize(T);` from this code synopsis, an error results: http://itensor.org/docs.cgi?page=classes/combiner&vers=cppv2
Thanks for pointing this out. I was chatting with a colleague about your comment and he confirmed what you are saying about the enforceability. I was already planning to weaken...
I think we should have this feature, and it's great that you've already nearly implemented it. There is a question of whether we should implement it similar to a combiner,...
Hi Marten, Thanks for catching the bug. I think defining addition and subtraction with diag-sparse tensors and dense tensors would be a reasonable thing to have. I had not defined...
Hi Matt, What do you think of this: in the current version 2, we could start making this behavior issue a deprecation warning: "ITensor has complex storage: use .cplx to...
That's a very good point!