Use global phase intrinsic for R with PauliI to make application of phase more consistent
#1450
Assignees
Labels
bug
Something isn't working
Comments
|
See #1317 (comment) and related discussion for a bunch more context. |
swernli
added a commit
that referenced
this issue
May 1, 2024
This change introduces a new, stdlib-only intrinsic for applying global phase during simulation to make phase behavior more consisten and fixes #1450. This intrinsic only applies the global phase during simulation; the new `GlobalPhase` intrinsic is a no-op for QIR generation, circuit generation, and resource estimation. The change also includes updates to the stdlib and associated tests to account for the corrected global phase behavior. This also fixes a bug in `DumpRegister` related to iteration through the state vector that ensures the "base" amplitude used for separating the state is consistently chosen, which avoids corner cases that could introduce a phase in the displayed output that wasn't present in the underlying state.
github-merge-queue bot
pushed a commit
that referenced
this issue
May 6, 2024
This change introduces a new, stdlib-only intrinsic for applying global phase during simulation to make phase behavior more consisten and fixes #1450. This intrinsic only applies the global phase during simulation; the new `GlobalPhase` intrinsic is a no-op for QIR generation, circuit generation, and resource estimation. The change also includes updates to the stdlib and associated tests to account for the corrected global phase behavior. This also fixes a bug in `DumpRegister` related to iteration through the state vector that ensures the "base" amplitude used for separating the state is consistently chosen, which avoids corner cases that could introduce a phase in the displayed output that wasn't present in the underlying state.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The current decompositions treat
RwithPauliIas identity, which can create confusion when comparing to mathematical behavior where Ri is expected to introduce a global phase. Instead, we should introduce a global phase intrinsic that is used during simulation to apply expected global phase changes and incorporate that into the behavior of theR1gate as well. This global phase intrinsic should be treated as a no-op during compilation for hardware, so it is not a__quantum__qis__intrinsic but rather a simulation specific intrinsic.The text was updated successfully, but these errors were encountered: