Add some more tests for DEFEXPI functionality #501
Conversation
| @@ -984,20 +984,22 @@ If ENSURE-VALID is T, then a memory reference such as 'foo[0]' will result in an | |||
| :arguments qubit-list)))) | |||
|
|
|||
| (defun pauli-term->matrix (term arguments parameters parameter-names) | |||
| (let* ((prefactor-fn | |||
There was a problem hiding this comment.
I was wondering where this commit snuck off to.
ecpeterson
force-pushed
the
tests/defexpi
branch
from
8b4f729
to
798c87c
Compare
|
The only suggested test not added is an explicit test involving |
|
Suggestion: test against |
| (let* ((cpp (compiler-hook (parse-quil prog-text) chip)) | ||
| (m (parsed-program-to-logical-matrix cpp))) | ||
| (is (quil::double= 1d0 (abs (magicl:ref m 0 0)))) | ||
| (loop :for j :below (magicl:matrix-rows m) |
There was a problem hiding this comment.
This is checking identityp up to phase.
| (declare (ignorable ,@parameter-names)) | ||
| ,(delayed-expression-expression (pauli-term-prefactor term))))) | ||
| (apply | ||
| (compile nil `(lambda ,parameter-names |
There was a problem hiding this comment.
Definitely not for this commit, but we might instead simply interpret the expression instead of calling the compiler.
i think this is a good suggestion. Could @jmbr or @ecpeterson make a "good-first-issue" for this? |
|
I will. |
cf. #498 (review)