Repository navigation
Conversation
Signed-off-by: Scott Thornton <wsttiger@gmail.com>
The six-angle-bit, 1e-3 default executes a degree-125 QSVT sequence on 18 qubits: 1380 s on qpp-cpu with two threads (the CI example timeout is 600 s and the wheel-test lanes run on four-vCPU runners), against 4 s for example 07. Make four angle bits and a 1e-2 inverse target the default (16 qubits, executed degree 99, 244 s under the same conditions; all checks pass with 3.4% full-vector error, dominated by the four-bit matrix quantization) and keep the 18-qubit configuration as the documented --value-bits 6 --inverse-epsilon 1e-3 precision run. Update the qubit, B and degree figures in the getting-started page and the block-encodings guide accordingly. Remove the "physical RHS norm restores linear scaling" check: it compared 2.5 * reconstructed with 2.5 * numpy_solution against 2.5 * total_bound, which is the preceding check multiplied through by a constant and verifies nothing about the recovery factor. State in the guide what does verify it: the factor beta/(c*alpha) is fixed analytically and the comparison with the quantized NumPy solve checks it. Signed-off-by: Scott Thornton <wsttiger@gmail.com>
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds a self-verifying sparse inverse application: one backward-Euler heat-diffusion step in a rod with fixed zero endpoints. Arithmetic banded oracles encode the interior SPD matrix. Independent NumPy constructions check the analytic spectrum/default rational solution, angle-quantized matrix, circuit block, raw complex QSVT response, physical scale restoration, full-vector temperatures, and residual. Coefficient, inverse-polynomial, phase, and circuit errors have separate derived budgets.
The six-bit, 1e-3 default retains B=206 of the CKS construction. A rigorously bounded Chebyshev tail (7.33e-11 uniformly on [-1,1]) reduces its executed degree from 411 to 125; matrix precision and requested inverse accuracy are unchanged. The tail is propagated into the error budget.
--full-cksexecutes all 411 degrees. Full-state real-response extraction is explicitly simulator postprocessing, with no fitted scale or isolated-filter success claim.Validation on CUDA-Q 0.16.0.post1, Python 3.12.3:
python3 -m pytest tests/python/test_examples.py -k '10_sparse_heat_equation or 07_matrix_inversion_qsvt' -q: 2 passed in 451.08 s on qpp-cpu with OMP_NUM_THREADS=8; the existing 600-second example timeout is unchanged.--full-ckspassed in 9.19 s. Full CKS execution is considerably slower on CPU.-W --keep-going, YAPF 0.43.0, and whitespace checks passed.Independent plan 08 PR against main; includes its own helpers and documentation, with no dependency on plan 05 or other proposed examples. QSPPACK is required; Matplotlib is lazily imported only for
--plotand is not added to core dependencies.