Qualtran icon indicating copy to clipboard operation
Qualtran copied to clipboard

Issues and improvements for chemistry algorithms

Open fdmalone opened this issue 2 years ago • 3 comments

This is a tracking issue for various known issues or limitations in the chemistry algorithms implementations.

THC

  • missing rotations primitive #386
  • cost differences: #390

SF

  • Missing implementations: #570
  • Load less data: #569
  • cost differences: #551

DF

  • Missing implementations: #571 #386
  • Load less data #569

First quantization

  • Missing decompositions: #572

Trotter

  • Missing arithmetic implementations: #485
  • Add QFT: #426
  • Fix arithmetic uncomputation in decompositions: #482

Other algorithms

  • For the second quantized algorithms there are also variants for periodic boundary conditions. Capturing at least the costs would be a great contribution.

fdmalone avatar Dec 07 '23 19:12 fdmalone

I think this should be a blocking issue for the 1.0 release @ncrubin @mpharrigan @tanujkhattar. At least finishing the implementations for the most important algorithms. People are going to want to use Qualtran to dive into these details.

wjhuggins avatar Dec 18 '24 20:12 wjhuggins

Contributors welcome

On Wed, Dec 18, 2024 at 12:21 PM William J. Huggins < @.***> wrote:

I think this should be a blocking issue for the 1.0 release @ncrubin https://github.com/ncrubin @mpharrigan https://github.com/mpharrigan @tanujkhattar https://github.com/tanujkhattar. At least finishing the implementations for the most important algorithms. People are going to want to use Qualtran to dive into these details.

— Reply to this email directly, view it on GitHub https://github.com/quantumlib/Qualtran/issues/573#issuecomment-2552195858, or unsubscribe https://github.com/notifications/unsubscribe-auth/AC4JSVHJECM2JWEHO6J6SS32GHKLRAVCNFSM6AAAAABT3PPVTWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKNJSGE4TKOBVHA . You are receiving this because you authored the thread.Message ID: @.***>

fdmalone avatar Dec 18 '24 22:12 fdmalone

But yeah, I would think that now it might not be too* much work to plug all the new primitives together, particularly for DF, and in particular the givens network (@ncrubin started this from the bottom up a while back in (#866), and I had an abandoned top down implementation #579)

(*) Famous last words

fdmalone avatar Dec 18 '24 22:12 fdmalone