{"id":"77f8a271-3f7e-48fc-a9ed-656bde06ebc5","arxiv_id":"2508.05608","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Pandora, a circuit-rewrite based compiler, handles billions of gates and outperforms TKET, Qiskit, and MQT.QCEC on large quantum circuits.","lead":"Pandora is an open-source quantum circuit compiler that claims to handle billions of gates and outperform TKET and Qiskit for circuits larger than 10,000 gates. A generalist might read this paper to see whether quantum software can now compile the huge circuits needed for chemistry and for Shor's algorithm.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Semantic soundness of Pandora's rewrite rules is unverified and unverifiable from the supplied text; an unsound rule would invalidate all compilation and equivalence-checking claims.","rationale":"I read the abstract in good faith and found no internal contradiction. The claim that Pandora handles billions of gates and outperforms TKET/Qiskit above 10,000 gates is plausible for a well-engineered rewrite-based tool. However, the central claim's validity is conditional on the rewrite system being semantically sound: a single non-unitary rewrite would break the correctness of every compilation and every equivalence-check result. The supplied full text is corrupted and even contains an extraneous arXiv identifier from a math.GT paper, making it impossible to inspect the rule list, any proofs, or the benchmark setup. The reader's weakest assumption identified exactly this: rewrite-rule semantic preservation. I agree that this is the most load-bearing premise. I considered benchmark matching as an alternative concern, but correctness is logically prior—if the rewrites are unsound, performance comparisons are irrelevant. I therefore propose a single concrete check: obtain the clean paper and verify that a rigorous soundness proof exists. If it does not, the paper remains unverdictable; if it does, the concern is largely resolved. Since we cannot run that check here, the appropriate verdict is unchanged from the reader's UNVERDICTED.","tokens_in":30707,"tokens_out":8097,"duration_ms":84535,"concrete_test":"Download the uncorrupted PDF of arXiv:2508.05608 from arXiv and inspect the section presenting Pandora's rewrite rules (likely Section 3). Check whether the paper proves, via a lemma or machine-checked verification, that every rewrite rule is a unitary circuit identity. If no such proof is present, the central soundness premise is unsupported and the claimed compilations/equivalence results cannot be relied upon.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Pandora's central capability—full compilation of Fermi-Hubbard 100×100 and 1024-bit Shor circuits, plus equivalence checking above 32 qubits—presupposes that every internal rewrite preserves the input unitary. The supplied full text is corrupted (mojibake) and embeds an arXiv line for a different paper (arXiv:2508.05605 [math.GT]), so no rewrite-rule list, correctness proof, or benchmark protocol can be inspected. For circuits above ~30 qubits, exact state-vector verification is computationally intractable, so correctness must rest on local rules being proven unitary identities. The abstract says only 'based on circuit rewrites'; it states no proof, machine-checking, or exhaustive testing of the rules. If even one rewrite rule is unsound, every reported compiled output and equivalence result is semantically invalid, and the scalability/performance numbers become moot. This is more load-bearing than benchmark fairness because it undermines the validity of all downstream claims regardless of how fair the comparisons are.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The abstract announces Pandora, an open-source multithreaded quantum-circuit rewriting tool, claiming it handles billions of gates, compiles Fermi-Hubbard 100×100 and 1024-bit Shor circuits, outperforms TKET and Qiskit on manipulation of circuits with more than 10,000 gates, and outperforms MQT.QCEC on equivalence checking for specific circuits above 32 qubits. The body of the paper as supplied is, however, unreadable mojibake: no algorithm, rewrite-rule list, data structure, correctness proof, benchmark protocol, or numerical result can be inspected. The text even contains an arXiv line belonging to a different paper (arXiv:2508.05605 [math.GT]). As submitted, the manuscript consists of headline assertions without supporting methodology or evidence.","tokens_in":30834,"tokens_out":6627,"duration_ms":72929,"significance":"If substantiated, the claimed capabilities would be practically significant: full compilation of 100×100 Fermi-Hubbard and 1024-bit Shor circuits would exceed typical open-source toolkit benchmarks by orders of magnitude, and streaming resource estimation at billion-gate scale would be useful for fault-tolerant workload sizing. The choice of TKET, Qiskit, and MQT.QCEC as external baselines is appropriate, and the stated open-source, multithreaded, HPC-enabled design is a strength. However, none of this is verifiable from the submitted document. There are no visible derivations, proofs, machine-checked artifacts, or reproducible benchmark data. The contribution is therefore currently an assertion rather than a demonstrated result.","major_comments":[{"comment":"The entire body of the manuscript is illegible mojibake; there is no readable algorithm, complexity analysis, or data-structure section. The abstract's central claims — billion-gate handling, full Fermi-Hubbard 100×100 and 1024-bit Shor compilations, and performance advantages over TKET and Qiskit — depend on exactly this missing content. No tables or figures with run times, gate counts, or memory usage are visible. As supplied, the paper provides no evidence for any of its headline empirical assertions, so the central claims are unverifiable.","section":"Full text (overall)"},{"comment":"The soundness of the rewrite engine is load-bearing: every compiled output and every equivalence-checking result is valid only if each rewrite rule preserves the unitary. The abstract mentions only 'circuit rewrites' and gives no list of rules, no proof of equivalence, and no mention of machine-checking or exhaustive testing. For equivalence checking above 32 qubits, exact state-vector verification is computationally impossible, so a single unsound rule would invalidate all reported compiled outputs and equivalence results. No such correctness evidence is present in the readable portions of the manuscript.","section":"Abstract ('based on circuit rewrites')"},{"comment":"The benchmark comparisons are presented without any protocol. The readable text does not specify matched input circuits, target gate sets, optimization levels, thread counts, machine configuration, timeouts, or how runs were aggregated. Statements such as 'performance advantage for manipulating circuits of more than 10000 gates' and 'outperforms MQT.QCEC on specific circuits' are therefore not quantitative, reproducible claims. Without a documented benchmark protocol, the reported performance advantage cannot be assessed.","section":"Abstract ('Compared to TKET and Qiskit...')"}],"minor_comments":[{"comment":"The text embeds the line 'arXiv:2508.05605v1 [math.GT] 7 Aug 2025', which belongs to a different paper. This should be removed; it is likely the result of corrupted source-file assembly.","section":"Full text (header)"},{"comment":"Once a clean manuscript is supplied, every section heading, equation, and reference will need to be legible; currently no equation number or bibliographic entry can be verified.","section":"Full text (throughout)"},{"comment":"Please define 'full compilation' (target gate set and optimization objective) and 'streaming resource estimation' precisely, so that the scope of the claimed capabilities is unambiguous.","section":"Abstract (terminology)"}],"recommendation":"reject","confidential_remarks":"The supplied full text is not a valid manuscript for review: it is garbled, contains an unrelated arXiv identifier, and offers no legible methodology, proofs, or benchmark data. I cannot review the science in this form. If this is an upload error, a clean, complete version should be requested before any further review; as submitted, the manuscript should not proceed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague—\n\nWhat you should know: this is an abstract-only paper right now. The supplied full text is mojibake from the first line, and it embeds an arXiv header for a different paper (2508.05605, math.GT). So everything below is based on the abstract and the claims it states.\n\nThe claim is worth taking seriously. Pandora is a rewrite-based compiler that says it handles billions of gates, streams circuit partitions for resource estimation, and produces full compilations of a 100×100 Fermi-Hubbard circuit and a 1024-bit Shor circuit. It claims a performance advantage over TKET and Qiskit above 10k gates, and over MQT.QCEC for equivalence checking above 32 qubits. If those numbers hold, that closes a real practical gap in quantum software—resource-estimation teams currently can't touch circuits of that scale. The benchmark targets are named, and the claims are falsifiable, which counts for something.\n\nWhat I can't do is verify any of it. The techniques themselves—circuit rewriting, multi-target compilation, equivalence checking—are established; the novelty is scale and engineering. That makes the empirical protocol the whole ballgame. I need baseline versions, hardware configuration, matched gate sets, and the benchmark script, and none of that is visible. The stress-test concern lands: for equivalence checking above 32 qubits you can't fall back on exact state-vector verification, so correctness rests entirely on the rewrite rules being unitary identities. The abstract doesn't claim a proof or machine-checked rules, so semantic soundness is an open question. I'm not saying the rules are wrong—only that the paper's load-bearing assumption is currently unexamined.\n\nWhat would change my mind, in order: (1) a readable manuscript, (2) an actual public repository shipping the rule set and benchmark scripts, (3) independent reproduction or a formal check of the rewrite rules. The 'open-source' claim in the abstract is a good sign; if the repo is real, that raises confidence meaningfully.\n\nWho this is for: people doing resource estimation for quantum chemistry or Shor at scale, and the quantum-software-tools community. It deserves a serious referee, but only once the file is readable. If I get a clean version, I'd run the scorecard again immediately.","headline":"The abstract makes a big, specific claim about billion-gate circuit compilation, but the supplied full text is corrupted—so the only honest verdict is 'can't verify yet.'","tokens_in":31401,"tokens_out":4546,"would_cite":false,"duration_ms":41038,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Pandora is a circuit-rewrite engine that scales quantum-circuit compilation to billions of gates, a regime far beyond current tools, and puts full compilation of practical quantum algorithms within reach.","keywords":["quantum circuit compilation","circuit rewriting","resource estimation","equivalence checking","large-scale quantum circuits","high-performance computing","Shor's algorithm","Fermi-Hubbard model"],"falsifier":"Take a 20-qubit Clifford circuit, run it through Pandora's full rewrite pipeline, and compare the resulting unitary to the original by exact simulation; any difference between the unitaries disproves the semantic-preservation claim.","tokens_in":30497,"feed_emoji":"⚛️","tokens_out":4163,"duration_ms":43700,"temperature":0.7,"pith_summary":"Pandora is presented as a circuit-rewrite engine that scales to billions of gates, a regime far beyond current quantum software. The paper claims it can fully compile a 100x100 Fermi-Hubbard circuit and a 1024-bit Shor's algorithm circuit, stream circuit partitions for resource estimation, and outperform current compilers and equivalence checkers on large circuits. If these claims hold, Pandora offers a practical route to compiling and resource-estimating fault-tolerant-scale quantum programs.","feed_headline":"One rewrite engine scales quantum circuits to billions of gates","feed_subtitle":"It compiles 1024-bit Shor's algorithm and 100x100 Fermi-Hubbard circuits, outpacing current quantum compilers.","key_machinery":"Circuit rewriting: a set of local, semantics-preserving transformations on gate sequences. Pandora uses these rewrites to restructure and optimize circuits, and the machine's design—multithreaded and HPC-aware—lets it apply rewrites across partitioned circuits, so the work is spread over many cores and large circuits are handled in streaming pieces. The rewrite rule set is the mechanism that carries both the compilation and equivalence-checking functionality.","core_discovery":"The paper introduces Pandora, an open-source, multithreaded, high-performance-computing-enabled tool that manipulates quantum circuits by rewriting gate sequences while preserving the overall unitary. The central claim is that the rewrite-based architecture handles circuits with billions of gates—far beyond what current compilers can touch—while still supporting full compilation, equivalence checking, and streaming resource estimation. As demonstrations, the authors report full compilation of a 100x100 Fermi-Hubbard circuit and a 1024-bit Shor's algorithm circuit. They also report that for circuits above 10,000 gates, Pandora manipulates circuits faster than TKET and Qiskit, and for equivale","pith_inferences":["If the rewrite rules are proven correct, Pandora's approach could serve as a reusable verification layer for modular compilation pipelines, not just a standalone tool.","The same streaming mechanism could be adapted to estimate fault-tolerant overheads (logical qubits, T factories) for circuits that never fit in memory.","The large-circuit advantage is likely driven by memory access patterns and partitioning rather than any single rewrite rule; testing that would require profiling beyond the paper's benchmarks.","A rewrite-based core could become the standard data structure for full-stack quantum compilers, potentially replacing DAG-based intermediate representations."],"forward_implications":["A full compilation of a 1024-bit Shor's algorithm circuit becomes practical as a single-tool operation rather than a theoretical benchmark.","Streaming resource estimation lets a pipeline process a circuit partition by partition without ever holding the full circuit in memory.","Equivalence checking can extend beyond the 32-qubit exact state-vector limit for structured circuits.","The 10,000-gate crossover point gives a concrete guide for when a rewrite-based engine beats conventional compilers."],"supporting_citations":[],"fun_headline_variants":["Quantum rewrites push circuit scale to billions of gates","Pandora compiles Shor's 1024-bit and 100x100 Hubbard circuits","Rewrite-based quantum tool scales to billions of gates","Pandora outruns Qiskit/TKET on 10k+ gate quantum circuits","Handles billions of gates: Pandora rewrites quantum circuits"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"Pandora's rewrite rules must exactly preserve the unitary of every circuit they are applied to; if any rewrite changes the operation, the compilation and equivalence results are unsound.","fun_headline_variants_meta":{"raw":{"variants":["Quantum rewrites push circuit scale to billions of gates","Pandora compiles Shor's 1024-bit and 100x100 Hubbard circuits","Rewrite-based quantum tool scales to billions of gates","Pandora outruns Qiskit/TKET on 10k+ gate quantum circuits","Handles billions of gates: Pandora rewrites quantum circuits"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00064,"raw_usage":{"total_tokens":2770,"prompt_tokens":717,"completion_tokens":2053,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":461,"completion_tokens_details":{"reasoning_tokens":1967}},"tokens_in":461,"tokens_out":2053,"duration_ms":14889,"temperature":1.0,"reasoning_tokens":1967,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T23:13:01.008380+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a 20-qubit Clifford circuit, run it through Pandora's full rewrite pipeline, and compare the resulting unitary to the original by exact simulation; any difference between the unitaries disproves the semantic-preservation claim.","supporting_citations":[],"review_version":1}