Pith. sign in

REVIEW 3 major objections 3 minor

Eleven quantum PDE kernels ranked on the same 1-D heat equation, same grids, same readout contract.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-15 04:07 UTC pith:LGT3IRDH

load-bearing objection Solid subfield benchmark of eleven quantum heat-equation kernels under one contract; the selection-guide claim is only as strong as that fixed readout, which we cannot fully audit from the abstract alone. the 3 major comments →

arxiv 2607.12688 v1 pith:LGT3IRDH submitted 2026-07-14 quant-ph cs.SE

Quantum PDE Solvers in Practice: Application-Driven Benchmarking of the Heat Equation

Mahmoud Elkarargy , Abdelaziz Rahwan , Abdelrahman Elsayed , Forat Hatem This is my paper
classification quant-ph cs.SE
keywords quantum PDE solversheat equationbenchmarkingHHLQSVTVQLSQITESchrödingerisation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Quantum PDE solvers have been hard to compare because papers use different discretizations, output models, reconstruction rules, and hardware assumptions. This paper builds a fixed, reproducible benchmark for the one-dimensional Dirichlet heat equation and runs eleven kernels on identical instances: three initial conditions, grids of 4 to 7 qubits, a CFL-like ratio near 0.4, and final time T=1. Under that contract, statevector runs show QSM and Schade-Hamiltonian matching the classical semi-discrete reference to floating-point precision, Schrödingerisation at about 10^{-4} error, and QITE as the strongest non-transform method for smooth data. With 10^5 shots, HHL collapses to roughly 0.79 relative L2 error while several low-depth methods become readout-limited; a norm-mismatch ablation shows reconstruction normalization alone accounts for 23–29% of the n=7 error for three methods. Compact observables such as total thermal energy need one to three orders of magnitude fewer shots than full-field reconstruction. The public benchmark is meant as a practical selection guide for quantum heat solvers under a shared problem definition.

Core claim

Under identical 1-D Dirichlet heat instances (n=4–7 qubits, r≈0.4, T=1) and a single fixed readout/reconstruction contract, QSM and Schade-Hamiltonian reproduce the semi-discrete classical reference to floating-point precision on statevector, Schrödingerisation reaches ~10^{-4} error, HHL under 10^5 shots degrades to ~0.79 relative ℓ2 error, and compact observables require 1–3 orders of magnitude fewer shots than full-field reconstruction.

What carries the argument

A shared application-driven benchmark that freezes problem instances, finite-difference discretization, CFL-like ratio, final time, and a single readout/reconstruction contract, then separates algorithmic, sampling, and device-noise error via statevector, ideal-shot, and noisy Aer backends.

Load-bearing premise

That one fixed readout contract, one finite-difference discretization, a single CFL-like ratio near 0.4, final time T=1, and the Aer noise model fairly rank all eleven kernels for practical solver selection.

What would settle it

Re-run the same eleven kernels on the same grids and initial conditions but with an adaptive time-stepping scheme or an output model that returns only compact observables instead of full-field reconstruction; if the relative ranking of QSM, Schade-Hamiltonian, Schrödingerisation, QITE and HHL reverses under that change, the selection guide does not generalize.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The manuscript presents a reproducible, application-driven benchmark of eleven quantum PDE kernels (HHL, QSVT, QLS-Fourier, VQLS, QITE, var-QITE, AVQDS, Hamiltonian simulation, Schade-Hamiltonian, Schrödingerisation, and QSM) for the 1-D Dirichlet heat equation under identical problem instances and a single fixed readout/reconstruction contract. It uses three initial conditions, grids n=4–7 qubits (N=16–128), CFL-like ratio r≈0.4, and T=1, and reports statevector, ideal-shot (10^5 shots), and noisy Aer results. On statevector, QSM and Schade-Hamiltonian match the semi-discrete classical reference to floating-point precision, Schrödingerisation reaches ~10^{-4} error, and QITE is strongest among non-transform methods for smooth data; under fixed shots HHL degrades to ~0.79 relative ℓ2 error; a norm-mismatch ablation attributes 23–29% of n=7 error for several methods to reconstruction normalization; compact observables need 1–3 orders of magnitude fewer shots than full-field reconstruction. The abstract frames the ranking as a practical selection guide.

Significance. If the fixed-contract ranking is fair and the reported numbers hold under inspection of the full methods and reconstruction definitions, the work would be a useful community resource: a controlled, multi-kernel comparison that separates algorithmic, sampling, and device-noise errors for a standard parabolic PDE, with concrete shot-budget guidance for compact observables. Public, reproducible benchmarks of this kind are scarce in quantum PDE literature and would help practitioners choose among heterogeneous kernels. The explicit norm-mismatch ablation and the separation of backends are strengths of the design as described. Significance is conditional on the fairness of the single readout/reconstruction contract across methods whose natural outputs differ (linear-system solutions, unitary evolutions, post-selected states, variational ansatze).

major comments (3)
  1. The central claim is a practical selection guide obtained by ranking eleven kernels under one fixed readout/reconstruction contract. The abstract does not define that contract (how amplitudes or post-selected states are mapped to the classical field, how normalization is applied, and how methods whose natural output is a unitary evolution or variational ansatz are forced into the same reconstruction). Without that definition, the ranking and the 23–29% norm-mismatch share at n=7 cannot be audited for fairness; methods optimized for different output models may be systematically mis-ranked. This is load-bearing for the selection-guide claim and must be specified and justified in the full text.
  2. The free parameters r≈0.4, T=1, and fixed 10^5 shots per step are presented as the common experimental setting. The abstract does not show that this CFL-like ratio and fixed final time do not systematically disadvantage adaptive, multi-step, or imaginary-time methods relative to single-step or spectral methods. A short sensitivity study (or an explicit argument that the ranking is stable under modest changes of r and T) is needed to support transfer of the selection guide beyond the chosen point.
  3. Reported statevector agreements (floating-point precision for QSM and Schade-Hamiltonian; ~10^{-4} for Schrödingerisation) and the HHL shot degradation (~0.79 relative ℓ2) are specific but, in an abstract-only review, cannot be checked against the semi-discrete reference definition, the finite-difference stencil, or the precise error metric. The full manuscript must make the classical reference, discretization, and reconstruction rule fully reproducible so that these numbers can be verified independently.
minor comments (3)
  1. Abstract-only review: figure and table clarity, notation consistency, and reference completeness cannot be assessed; these should be checked once the full text is available.
  2. The abstract lists eleven kernels but does not briefly indicate which are coherent linear solvers, which are variational, and which are unitary dilations; a one-line taxonomy in the abstract or early introduction would help readers navigate the ranking.
  3. The phrase "CFL-like ratio r≈0.4" should be defined explicitly (e.g., in terms of Δt, Δx, and diffusivity) so that the discretization is unambiguous.

Circularity Check

0 steps flagged

No circularity: abstract-only benchmark ranks external methods against a classical reference under fixed instances; ranking is not forced by definition or self-citation.

full rationale

Only the abstract is available. It describes a comparative benchmark of eleven quantum PDE kernels (HHL, QSVT, QLS-Fourier, VQLS, QITE, var-QITE, AVQDS, Hamiltonian simulation, Schade-Hamiltonian, Schrödingerisation, QSM) on identical 1-D Dirichlet heat instances (n=4–7, r≈0.4, T=1) with a shared readout/reconstruction contract, reporting errors relative to a semi-discrete classical reference under statevector, shot, and noisy backends. The reported floating-point agreement of QSM and Schade-Hamiltonian, ~10^{-4} Schrödingerisation error, HHL shot degradation, and shot-count savings for compact observables are empirical outcomes of that comparison, not quantities fitted to or defined by the target metrics. Design choices (fixed r, T, shot budget, reconstruction rule) set the experimental frame and may affect fairness of ranking—an external validity concern, not circularity. There are no self-definitional identities, fitted parameters renamed as predictions, load-bearing self-citations, uniqueness theorems imported from the authors, ansatz smuggling, or renaming of known results. Score 0 is the honest finding for an abstract-only methods-comparison paper whose central claims are measured against an external classical reference.

Axiom & Free-Parameter Ledger

4 free parameters · 3 axioms · 0 invented entities

This is an empirical benchmark, not a derivation. Load-bearing inputs are experimental design choices (grid sizes, r, T, shot budget, Aer noise) and the classical semi-discrete reference used as ground truth. No new physical entities are postulated. Free parameters are the fixed experimental knobs that define the ranking, not fitted theory constants.

free parameters (4)
  • CFL-like ratio r = ≈0.4
    Fixed at ≈0.4 for all compared runs; choice affects stability and error ranking of time-stepping methods.
  • final time T = 1
    Fixed at T=1 for all instances; defines the integration horizon over which errors accumulate.
  • shots per step = 1e5
    Fixed ideal-shot budget of 10^5 per step used to declare readout-limited behavior (e.g., HHL ~0.79 error).
  • grid sizes n=4..7 qubits = n=4,5,6,7
    Discrete problem sizes N=16..128 that bound the claimed scaling and noise conclusions.
axioms (3)
  • domain assumption Classical semi-discrete finite-difference heat equation is the correct reference for quantum solver error.
    All reported ℓ2 and floating-point matches are defined against this reference; abstract treats it as ground truth.
  • ad hoc to paper A single fixed readout/reconstruction contract is a fair basis for ranking heterogeneous quantum PDE kernels.
    Central framing of the benchmark; methods designed for different output models may be disadvantaged.
  • domain assumption Aer noisy backend adequately represents device-noise error for the stated separation of error sources.
    Used to isolate device-noise from algorithmic and sampling error; fidelity to real hardware is assumed.

pith-pipeline@v1.1.0-grok45 · 6280 in / 2736 out tokens · 28820 ms · 2026-07-15T04:07:43.459357+00:00 · methodology

0 comments
read the original abstract

Quantum PDE solvers are difficult to evaluate in practice because published studies use different discretizations, output models, reconstruction rules, and hardware assumptions. We present a reproducible, application-driven benchmark for the 1-D Dirichlet heat equation that compares eleven kernels under the same problem instances and readout contract. The benchmark covers coherent linear solvers (HHL, QSVT, and QLS-Fourier), VQLS, imaginary-time methods (QITE, var-QITE, and AVQDS), real-time Hamiltonian simulation and unitary dilations (Hamiltonian simulation, Schade-Hamiltonian, and Schr"odingerisation), and the spectral quantum simulation method (QSM). We use three initial conditions, four grid sizes from $n=4$ to $7$ qubits ($N=16$ to $128$), a CFL-like ratio $r\approx0.4$, and final time $T=1$. Statevector, ideal-shot ($10^5$ shots per step), and noisy Aer backends separate algorithmic, sampling, and device-noise errors. On statevector, QSM and Schade-Hamiltonian reproduce the semi-discrete reference to floating-point precision, Schr"odingerisation reaches approximately $10^{-4}$ error, and QITE is the strongest non-transform method for smooth data. Under the fixed-shot setting, HHL degrades to approximately $0.79$ relative $\ell_2$ error, while several low-depth or postselected methods become readout-limited. A norm-mismatch ablation attributes 23--29% of the $n=7$ smooth-initial-condition error of Hamiltonian simulation, AVQDS, and QLS-Fourier to reconstruction normalization. Compact observables, including total thermal energy and individual Fourier-mode weights, require 1--3 orders of magnitude fewer shots than full-field reconstruction. The resulting public benchmark provides a practical guide for selecting quantum PDE solvers.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.