Add Azure model #568

NoureldinYosri · 2023-12-07T00:03:00Z

Reproduction of the estimates in https://arxiv.org/abs/2211.07629

This PR replaces #327 and uses the infrastrucutre introduced in #357, #378 and #461

fixes #269

mpharrigan

some initial nits

qualtran/surface_code/azure_cost_model.ipynb

mpharrigan · 2023-12-08T18:58:50Z

qualtran/surface_code/azure_cost_model.ipynb

+    "\n",
+    "Where $a$ and $b$ are constants that depend on the gate set and the approximation protocol. Table 1 of [Kliuchnikov et al](https://arxiv.org/abs/2203.10064) gives estimates for those constants for different combinations of gate sets and protocols.\n",
+    "\n",
+    "In [Beverland et al](https://arxiv.org/abs/2211.07629), Although they don't specify the approximation algorithm their constants of $a=0.53, b=5.3$ are closest to the mixed fallback approximation protocol with a Clifford+T gate set constants of $a=0.53, b=4.86$. "


confusing sentence construction

mpharrigan · 2023-12-08T22:43:33Z

qualtran/surface_code/azure_cost_model.ipynb

+   "source": [
+    "### Magic State Factory\n",
+    "\n",
+    "The magic state factory in our case is a T-state factory. The paper describes 15-to-1 factories in Appendix C, but only the overall summary in Table VII in terms of physical qubits count and generation time is used in estimation in Appendix F."


physical qubits count -> physical qubit count

I also don't really understand what's trying to be expressed in this paragraph. Overall: this notebook should stand alone without having intimate knowledge of the appendices of the referenced paper.

"Appendix F" is used in many places in the write-up. I would probably put something near the top

The paper describes the formulas used for estimating cost in the various appendices. The final estimates are brought together (??) in Appendix F. Throughout, we reproduce the costs from the paper; in particular Appendix F.

Then throughout, don't say appendix F unless you're referring to an equation or something.

Only physical qubit count and generation time is used in the estimation in the Beverland paper.

qualtran/surface_code/azure_cost_model.ipynb

mpharrigan · 2023-12-08T23:01:57Z

qualtran/surface_code/azure_cost_model.ipynb

+    "\n",
+    "Other authors (e.g. [Fowler, Gidney](https://arxiv.org/abs/1808.06709)) assume that the entire process takes a specific time (e.g. $\\approx 1\\mu s$).\n",
+    "\n",
+    "In Qualtran, we use \"error detection circuit time\" rather than fall to the low level hardware parameters. "


"error detection circuit time" isn't what the parameter is actually called. These should be consistent (or consistent-ish).

I like that you contrast the two ways of putting in the physical assumptions. It would be nice for this final sentence to say which one Qualtran follows. If I understand correctly, we're using a Gidney Fowler parameter even though it has this decomposition in the azure model

mpharrigan · 2023-12-08T23:08:28Z

qualtran/surface_code/azure_cost_model.ipynb

+   "metadata": {},
+   "source": [
+    "Comparing our esimates of $Q = 230, C_{min} = 1.8 \\times 10^5, M = 6 \\times 10^5$ to the paper estimates of \n",
+    "$Q = 230, C_{min} = 1.5 \\times 10^5, M = 2.4 \\times 10^6$. We find a match for $Q$ and $C_{min}$ however we are off by 4x for the number of T gates $M$.\n",


explain the cause of this discrepancy

that is the next paragraph

qualtran/surface_code/azure_cost_model.ipynb

mpharrigan · 2023-12-08T23:10:37Z

qualtran/surface_code/azure_cost_model.ipynb

+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "'algorithm run time of %g seconds' % (qec.error_detection_circuit_time_us(d) * 1e-6 * c_min)"


use f-strings or .format()

qualtran/surface_code/azure_cost_model.ipynb

mpharrigan

can you remove the doc rst file changes from this PR. I need to think about organization. Otherwise, looks good to merge

* Add SimpleAddConstant bloq * Address many comments on the simple add constant bloq * Make modifications to SimpleAddConstant to address concerns * Use bit tools functions for getting bitstrings * Reformat on_classical_vals and fix nits * Add test failing due to uint to float bug * Stash testing changes before adding MultiControlX * Add MultiControlX() bloq with functional classical simulation * Small linting changes * Add one more correctness test * Add SimpleAddConstant bloq * Address many comments on the simple add constant bloq * Make modifications to SimpleAddConstant to address concerns * Use bit tools functions for getting bitstrings * Reformat on_classical_vals and fix nits * [autodoc] multi_control_multi_target_pauli (#532) * sympy in autogen * bloq example for multi_control_multi_target_pauli * add to docs * [autodoc] addition bloqs (#519) * import sympy everywhere * default directory for notebooks * Add bloq examples for addition module * Factor out generic bloqs for testing (#564) * Use Oracles for sparse select and prepare (#546) * Use select and prepare oracle. * Format. * Fix imports. * Fix argument. * Fix docstring. * Address review comments. --------- Co-authored-by: Matthew Harrigan <[email protected]> * Adjoint protocol (#565) * Adjoint protocol * get block encoding working * Address review 1 * Pylint * Fixes bug with addition endianness. (#576) * Fix bug with addition endianness. * Test no decompose in same unit test as decompose. --------- Co-authored-by: Matthew Harrigan <[email protected]> * Bump jupyter-server from 2.11.0 to 2.11.2 in /dev_tools/requirements/envs (#563) Bump jupyter-server in /dev_tools/requirements/envs Bumps [jupyter-server](https://github.com/jupyter-server/jupyter_server) from 2.11.0 to 2.11.2. - [Release notes](https://github.com/jupyter-server/jupyter_server/releases) - [Changelog](https://github.com/jupyter-server/jupyter_server/blob/main/CHANGELOG.md) - [Commits](jupyter-server/jupyter_server@v2.11.0...v2.11.2) --- updated-dependencies: - dependency-name: jupyter-server dependency-type: direct:development ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> * Sparse decomp select (#547) * Use select and prepare oracle. * Format. * Fix imports. * Fix argument. * Fix docstring. * Sparse SELECT decomposition. * Temporarily delete sparse notebook. * Remove unused import. --------- Co-authored-by: Matthew Harrigan <[email protected]> * Initial overview documentation of protocols (#504) * Use public default fallbacks * Initial overview documentation of protocols * reword --------- Co-authored-by: Fionn Malone <[email protected]> * An assortment of generalizers (#554) * Generalizers * revert this one * pylint * Add test failing due to uint to float bug * Stash testing changes before adding MultiControlX * Add resource estimation chem tutorial (#529) * Add block encoding bloq. * Fix dagger method. * Fix evolve if bloq has no adjoint. * Fix CSwap wiresymbols. * Work on tidy THC prepare. * Use SelectOracle in THC. * Refactor THC prepare. * Use CSwap not CSwapApprox. * Fix THC Prepare. 1. Use prepare oracle and expose necessary ancilla. 2. Use CSwaps not CSwapApprox. 3. Use bitsize as input for contstructor rather than epsilon. * Fix docs. * Use new autodoc. * Move control register to selection register. The control registers for the SelectOracle are used for building a controlled walk operator, not any internal controls used by select. * Add control_val to THC select. * Remove print. * Resource estimation tutorial. * Tidy. * Reuse controlled Z block for DF and THC. * Tidy up notebook. * Fix format error. * Fixup merge * Remove oks. --------- Co-authored-by: Matthew Harrigan <[email protected]> * Sparse decomposition for prepare (#548) * Use select and prepare oracle. * Format. * Fix imports. * Fix argument. * Fix docstring. * Sparse SELECT decomposition. * Temporarily delete sparse notebook. * Decomposition for sparse prepare. * Add back notebook. * Use algorithm from reference rather than triu. * Remove duplicate property. * Beware thresholding in assertion. * Fix fp float multiplication formula. * Address review comment. * Rename select (#552) * Single factorization bloqs. * Use OF test params. * Better unit tests. * Better tests. * Remove unused import. * Fix QROM data size for SF and add note / reference. * Remove duplicate num_aux / unused variables. * Add missing __init__. * Add missing init / license. * Format. * Use select and prepare oracle. * Format. * Fix imports. * Fix argument. * Fix docstring. * Sparse SELECT decomposition. * Temporarily delete sparse notebook. * Decomposition for sparse prepare. * Add back notebook. * Address review comments. * Fix test. * Replace select.py with select_bloq.py. * Errant select.py. * Remove aspirational factory method. * Fix notebook errors. * Fix lint error. --------- Co-authored-by: Matthew Harrigan <[email protected]> * Add Azure model (#568) * [physical costing] Grid search best parameters with multiple CCZ factories (#561) * grid search best params * multifactory added, some documentation of CCZ grid search, THC compilation notebook * fix docstrings and minor edits * Added references * format * Remove outdated test, fix docstrings * format * format * gitignore vscode settings * add test based on THC * format --------- Co-authored-by: Matthew Harrigan <[email protected]> * Rename qro(a)m target registers to avoid name collision with named qubits (#577) * Rename qro(a)m target registers to avoid name collision with named qubits. If the number of data sets is >= 10 collisions could occur (see #556). * Fix alt_mu/nu register size. * Add link to issue. * Fix bug in auto generated `_decompose_with_context_` from `decompose_bloq` in GateWithRegisters (#542) * Fix bug in auto generated _decompose_with_context_ from decompose_bloq in GateWithRegisters * cleanup * Add test for interop qubit manager * Add test for auto generated decompose from context * Idiomatic adjoints for bloqs (#566) * Basic bloqs support adjoint * And.uncompute * Overhaul * validate MultiAnd fix * Fix porting util example * Alloc/Free adjoint * Revert comparison.py overhaul * Lint * Fix notebook's usage of MultiAnd --------- Co-authored-by: Fionn Malone <[email protected]> * More idiomatic adjoint changes (#583) * [autodoc] Chemistry algorithms (#578) * Use autodoc in first quantization. * Remove old doc autogeneration. * Remove duplicate select. * Add autodocs. * Autodoc for trotter. * Fix format and no return. * Add trotter bloqs notebook and fix some typos. * Add THC too. * Chebyshev polynomial block encoding (#510) * Add chebyshev polynomial block encoding. Adds a Reflection bloq which is quite common (often used to implement 2|0><0| - 1). This is a multi-controlled pauli Z up to some phases, wrap this in a bloq to make nicer diagrams. Also update notebook for MCP following autodocgen. * Add test. * pretty_name - > short_name. * Remove sparse select. (#586) * Remove sparse select. * Fix lint error. * Add `QFTPhaseGradient` which synthesizes rotations via addition into gradient register (#505) * Add QFTPhaseGradient which synthesizes rotations via addition into gradient register * Address comments by Matt and Fionn * Add tests for PlusEqualProduct and fix bug in on_classical_vals * Routine maintenance of the docs (#584) * Update generated notebooks v1 * Update generated notebooks v2 * document how to preview the docs * Support inline block math in html output * Rename misspelled notebook * Include notebooks in toc * Add `HammingWeightPhasingViaPhaseGradient` Bloq (#509) * Add HammingWeightPhasingViaPhaseGradient Bloq * Address nits and comments * Fix docstring * Fix docstrings and add more tests * Add tutorial on writing chemistry algorithms in qualtran (#558) * Fix imports. * Sparse SELECT decomposition. * Decomposition for sparse prepare. * Add back notebook. * Replace select.py with select_bloq.py. * WIP. * Finish tutorial. * Allow a block size parameter for black box qroam. * Fix missing import. * Fix lint errors. * Add test for chem tutorials and run notebooks. * Fix notebook. * Fix lint errors. * Fix test. * Remove errant notebook. * Fix argument. * suppress output from `assert_valid_bloq_decomposition` * Address review comments. --------- Co-authored-by: Matthew Harrigan <[email protected]> * Move ControlledBloq to `_infra` (#587) Co-authored-by: Fionn Malone <[email protected]> * Update NDArray proto and serialization to avoid running eval on serialized code (#588) * Update NDArray proto and serialization to avoid running eval on serialized code * Fix format * GQSP: Implement complementary polynomial computation (#581) * implement complementary polynomial computation (theorem 4 of https://arxiv.org/abs/2308.01501) * fix pylint * minor refactor * fix tests for real polynomials * run more tests with symbolic matrices. * add more comments * fix lint error * fix typos + add comment on why larger_roots / smaller_roots are swapped * remove np.around for Q * typo complimentary -> complementary * Add SimpleAddConstant classical sim tests * Fix some unit tests to work with merged MultiControlX() * Silencing pylint warnings * Added tracking issues for TODOs --------- Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: Matthew Harrigan <[email protected]> Co-authored-by: Fionn Malone <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Noureldin <[email protected]> Co-authored-by: Stefano Polla <[email protected]> Co-authored-by: Tanuj Khattar <[email protected]> Co-authored-by: Anurudh Peduri <[email protected]>

NoureldinYosri and others added 28 commits October 5, 2023 16:16

Add a summary of QEC algorithms class

c3ffd19

improve docstring

e1f9b4f

nit

310c0dc

nit

cf1cde3

nit

4cab323

nit

a24c06f

fix docstring indentation

174c239

will be back

885f732

Merge branch 'main' into azure_model

a1e00c8

ready for review

8fa3272

doc

ea9edc7

address comments

5a14ec6

Merge branch 'main' into azure_model

b81e1ea

tmp

30092d7

update terminalogy

3cc5d90

nit

6c79a33

Merge branch 'main' into azure_model

dd38ded

nit

5d5e7e4

Merge branch 'main' into azure_model

d3916ec

remove formulae.py

ac1ee47

nit

234338a

also fix #535

28b2a8b

address comments

b4e9240

Merge branch 'main' into azure_model

1fa2e68

style

b61d688

Merge branch 'main' into azure_model

b02b29f

add azure model

5422c86

Add Azure model

f87a73f

NoureldinYosri changed the title ~~azure model~~ Add Azure model Dec 7, 2023

nit

8d42dc5

NoureldinYosri and others added 3 commits December 6, 2023 16:23

docs

61c56a2

Merge branch 'main' into azure_model

af028f2

docs

23a8fbb

NoureldinYosri requested a review from mpharrigan December 7, 2023 19:11

NoureldinYosri added 5 commits December 7, 2023 11:30

add notebook to index

66dcc24

nit

effcc8c

nit

8e8a7f4

nit

c8fa614

nit

5e843ee

mpharrigan reviewed Dec 8, 2023

View reviewed changes

address comments

9eaea79

NoureldinYosri force-pushed the azure_model branch from f73f7f5 to 9eaea79 Compare December 9, 2023 01:07

NoureldinYosri and others added 2 commits December 8, 2023 17:10

Merge branch 'main' into azure_model

9149c53

nit

dfe5db6

mpharrigan requested changes Dec 15, 2023

View reviewed changes

NoureldinYosri and others added 2 commits December 15, 2023 11:55

ditch rst file changes

27a1313

Merge branch 'main' into azure_model

cb3ab2f

NoureldinYosri requested a review from mpharrigan December 15, 2023 20:05

NoureldinYosri enabled auto-merge (squash) December 15, 2023 20:06

Merge branch 'main' into azure_model

a105620

mpharrigan approved these changes Dec 19, 2023

View reviewed changes

NoureldinYosri merged commit 42e0e85 into main Dec 19, 2023
9 checks passed

mpharrigan mentioned this pull request Jan 4, 2024

Add physical costing notebook(s) to the docs #582

Closed

mpharrigan added the physical_cost label Jan 8, 2024

fpapa250 pushed a commit to fpapa250/Qualtran that referenced this pull request Jan 19, 2024

Add Azure model (quantumlib#568)

4d2363f

mpharrigan deleted the azure_model branch February 27, 2024 21:00

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add Azure model #568

Add Azure model #568

NoureldinYosri commented Dec 7, 2023

mpharrigan left a comment

mpharrigan Dec 8, 2023

NoureldinYosri Dec 9, 2023

mpharrigan Dec 8, 2023

mpharrigan Dec 8, 2023

NoureldinYosri Dec 9, 2023

mpharrigan Dec 8, 2023

NoureldinYosri Dec 9, 2023

mpharrigan Dec 8, 2023

NoureldinYosri Dec 9, 2023

mpharrigan Dec 8, 2023

mpharrigan left a comment

Add Azure model #568

Add Azure model #568

Conversation

NoureldinYosri commented Dec 7, 2023

mpharrigan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mpharrigan left a comment

Choose a reason for hiding this comment