Update SelectTHC template to allow batching of Givens rotations #8682

ddhawan11 · 2025-11-20T16:52:31Z

Context:
SelectTHC template is modified to allow more flexibility to the user.

Description of the Change:
SelectTHC template is modified to take in another keyword argument, batched_rotations, that allows trade-off between T-gates and qubits by allowing batching of Givens rotations in the Select circuit.

Benefits:

Possible Drawbacks:

Related GitHub Issues:

[sc-103101]

github-actions · 2025-11-20T16:52:56Z

Hello. You may have forgotten to update the changelog!
Please edit doc/releases/changelog-dev.md with:

A one-to-two sentence description of the change. You may include a small working example for new features.
A link back to this PR.
Your name (or GitHub username) in the contributors section.

codecov · 2025-11-20T17:28:16Z

Codecov Report

❌ Patch coverage is 87.50000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 99.42%. Comparing base (10cbca2) to head (a715ead).
⚠️ Report is 1 commits behind head on qubitizeTHC_bugfix.

Files with missing lines	Patch %	Lines
pennylane/estimator/templates/select.py	87.50%	2 Missing ⚠️

Additional details and impacted files

@@                  Coverage Diff                   @@
##           qubitizeTHC_bugfix    #8682      +/-   ##
======================================================
- Coverage               99.43%   99.42%   -0.01%     
======================================================
  Files                     588      588              
  Lines                   62661    62653       -8     
======================================================
- Hits                    62307    62295      -12     
- Misses                    354      358       +4

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

soranjh

Thanks @dime10, please upgrade the explanation and name of the new argument.

pennylane/estimator/templates/select.py

soranjh · 2025-11-20T19:39:27Z

pennylane/estimator/templates/select.py

-        qrom_twobody = resource_rep(
+        # QROM to load rotation angles for 2-body integrals + 1-body integrals
+        qrom_full = resource_rep(


Why this is changed?

Just changed some comments and variable names to avoid ambiguity.

pennylane/estimator/templates/select.py

soranjh · 2025-11-20T19:41:54Z

pennylane/estimator/templates/select.py

-        # Adjoint of QROM for one body integrals Eq. 35 in arXiv:2011.03494
-        gate_list.append(GateCount(resource_rep(qre.Adjoint, {"base_cmpr_op": qrom_onebody})))
+        # Adjoint of QROM for two body integrals only Eq. 35 in arXiv:2011.03494
+        gate_list.append(GateCount(resource_rep(qre.Adjoint, {"base_cmpr_op": qrom_twobody})))


Please be careful with this changes. Currently, it is confusing why the changes are made.

I just changed some variable names for added clarity, this is not really changing the resources as the tests remain the same for the default version.

pennylane/estimator/templates/select.py

soranjh · 2025-11-20T19:44:57Z

tests/estimator/templates/test_estimator_select.py

+                29,
+                13,
+                1,
+                {"algo_wires": 138, "auxiliary_wires": 388, "toffoli_gates": 7493},


How can we trust these reference numbers?

I had to manually calculate them, there is not any reference which provides the exact numbers with batching. The reference that we cited provides a plot with degree of change with some degree of batching but no exact numbers.

tests/estimator/templates/test_estimator_select.py

pennylane/estimator/compact_hamiltonian.py

Jaybsoni

Left an initial review, I will read the paper and double check that the resource_decomp matches the results there as well in my next review.

Jaybsoni · 2025-11-27T21:02:28Z

pennylane/estimator/templates/select.py

+        if batched_rotations == num_orb - 1:
+            restore_qrom = False
+
+        num_givens_blocks = np.ceil((num_orb - 1) / batched_rotations)


This is what was causing the floats in the number of allocated qubits.

Suggested change

num_givens_blocks = np.ceil((num_orb - 1) / batched_rotations)

num_givens_blocks = int(np.ceil((num_orb - 1) / batched_rotations))

Jaybsoni · 2025-11-27T21:04:10Z

pennylane/estimator/templates/select.py

+     Total wires: 371.0
+       algorithmic wires: 58
+       allocated wires: 313.0
+         zero state: 313
+         any state: 0.0


See my other comment below that fixes the typo causing the floats.

Suggested change

Total wires: 371.0

algorithmic wires: 58

allocated wires: 313.0

zero state: 313

any state: 0.0

Total wires: 371

algorithmic wires: 58

allocated wires: 313

zero state: 313

any state: 0

Jaybsoni · 2025-11-27T21:05:32Z

pennylane/estimator/templates/select.py

-      'Hadamard': 6.406E+3
+       'Toffoli': 2.219E+3,
+       'CNOT': 1.058E+4,
+       'X': 268.0,


Suggested change

'X': 268.0,

'X': 268,

Jaybsoni · 2025-11-27T21:16:41Z

pennylane/estimator/templates/select.py

+     Total gates : 2.534E+4
+       'Toffoli': 2.633E+3,
+       'CNOT': 1.440E+4,
+       'X': 804.0,


Suggested change

'X': 804.0,

'X': 804,

pennylane/estimator/templates/select.py

Jaybsoni · 2025-11-27T21:20:42Z

pennylane/estimator/templates/select.py

+            used to trade-off extra wires for reduced circuit depth. Defaults to :code:`None`, which internally
+            determines the optimal depth.


Suggested change

used to trade-off extra wires for reduced circuit depth. Defaults to :code:`None`, which internally

determines the optimal depth.

used to trade-off extra wires for reduced circuit depth. Defaults to :code:`None`, which determines

the optimal depth.

Jaybsoni · 2025-11-27T21:28:00Z

pennylane/estimator/templates/select.py

    Resources:
        The resources are calculated based on Figure 5 in `arXiv:2011.03494 <https://arxiv.org/abs/2011.03494>`_



This should be updated to provide more information. For example I am not sure how the batched rotations are incorporated here? See the TrotterPauli template as an example.

pennylane/estimator/templates/select.py

Jaybsoni · 2025-11-27T21:36:58Z

pennylane/estimator/templates/select.py

        Resources:
-            The resources are calculated based on Figure 5 in `arXiv:2011.03494 <https://arxiv.org/abs/2011.03494>`_.
+            The resources are calculated based on Figure 5 in `arXiv:2011.03494 <https://arxiv.org/abs/2011.03494>`_ and
+            `arXiv:2501.06165 <https://arxiv.org/abs/2501.06165>`_.


Is there a specific section or figure I should be looking at? Fig 4?

Added figure number.

Jaybsoni · 2025-11-27T22:31:30Z

pennylane/estimator/templates/select.py

+        if batched_rotations == num_orb - 1:
+            restore_qrom = False
+
+        num_givens_blocks = np.ceil((num_orb - 1) / batched_rotations)


Also need to cast this as int

Co-authored-by: Jay Soni <jbsoni@uwaterloo.ca>

Updated Select THC to take another argument

02acc8e

ddhawan11 requested a review from Jaybsoni November 20, 2025 16:52

ddhawan11 and others added 2 commits November 20, 2025 12:02

Updated changelog

e240c1d

Merge branch 'master' into modify_qubitize_thc

f2e4b7a

ddhawan11 requested a review from soranjh November 20, 2025 17:03

soranjh reviewed Nov 20, 2025

View reviewed changes

ddhawan11 and others added 4 commits November 21, 2025 09:22

Merge branch 'master' into modify_qubitize_thc

4174624

Added tests for compact Hamiltonian classes

851c2ac

removed 1-norm stuff

8c91cbe

Added another example

2311a81

ddhawan11 requested a review from soranjh November 27, 2025 15:39

Jaybsoni reviewed Nov 27, 2025

View reviewed changes

Updated with bugfix PR

970f626

ddhawan11 changed the base branch from master to qubitizeTHC_bugfix December 4, 2025 14:49

ddhawan11 and others added 3 commits December 4, 2025 09:52

updated

59d3f9b

Merge branch 'qubitizeTHC_bugfix' into modify_qubitize_thc

bf771bd

Apply suggestion from @Jaybsoni

a715ead

Co-authored-by: Jay Soni <jbsoni@uwaterloo.ca>

	num_givens_blocks = np.ceil((num_orb - 1) / batched_rotations)
	num_givens_blocks = int(np.ceil((num_orb - 1) / batched_rotations))

		used to trade-off extra wires for reduced circuit depth. Defaults to :code:`None`, which internally
		determines the optimal depth.

		Resources:
		The resources are calculated based on Figure 5 in `arXiv:2011.03494 <https://arxiv.org/abs/2011.03494>`_

Update SelectTHC template to allow batching of Givens rotations #8682

Are you sure you want to change the base?

Update SelectTHC template to allow batching of Givens rotations #8682

Conversation

ddhawan11 commented Nov 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Nov 20, 2025

Uh oh!

codecov bot commented Nov 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

soranjh left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Jaybsoni left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

ddhawan11 commented Nov 20, 2025 •

edited

Loading

codecov bot commented Nov 20, 2025 •

edited

Loading