Quantum computing is a novel paradigm that can potentially disrupt the Computer Science field. It promises to overcome the gap between the limitations of current classical processors and the performance demanded by modern applications. Nevertheless, significant challenges persist in constructing reliable quantum computers. Qubits within today’s Quantum Processing Units (QPUs) remain highly unstable and susceptible to noise originating from various sources, ultimately preventing the reliable execution of many applications on quantum processors. To address these issues, the current generation of quantum devices integrates Quantum Error Correction (QEC) mechanisms to recover the computation, thus representing a critical component toward fault tolerance. However, the QEC process faces several constraints, with the latency of error decoding being the most critical limitation. In this scenario, Field Programmable Gate Arrays (FPGAs) have arisen as a promising platform for deploying QEC engines. Their deterministic latency and spatial computing capabilities enable them to meet strict timing requirements, maintaining low power consumption at the same time. Existing approaches leveraging FPGAs for QEC have delivered encouraging results, yet substantial room for improvement remains. In particular, the field lacks consensus on the most significant criteria to deploy and assess QEC engines on FPGA, leaving the research landscape fragmented. Furthermore, beyond the pressing needs of latency optimizations, future solutions must also remain adaptable to emerging QPU technologies to keep pace with this rapidly evolving domain. Lastly, abstraction frameworks and modeling tools are essential to equip quantum engineers with the means to effectively customize QEC engines, ensuring their practical usability. In this context, this dissertation investigates how FPGA technology can advance the field of quantum computing through efficient deployment of QEC engines, enabling the full potential of this emerging paradigm. Specifically, it introduces a series of methodologies to develop high-performance and energy-efficient QEC decoders by leveraging FPGAs. The dissertation begins with a comprehensive review of the State of the Art and, building on this analysis, proposes a set of evaluation pillars to guide the benchmarking and evaluation of novel solutions. It then presents two distinct methodologies for designing FPGA-based QEC decoders. The first, QASBA, implements a gold-standard quantum error decoding algorithm entirely in hardware, ensuring high accuracy. To enhance its practical applicability, an automation framework is also introduced to enable quantum engineers to customize QASBA-based decoders according to the characteristics of the target quantum system. The second methodology forms the basis of QUEKUF, which combines the efficiency of a heuristic decoding algorithm with the computing power of FPGAs to achieve superior performance. Also in this case, to facilitate practical deployment, QUEKUF incorporates a latency-oriented resource optimization model that assists final users in selecting the most effective architectural configuration for their specific quantum environment. This dissertation then contextualizes these contributions within the State of the Art, evaluating their advantages and highlighting opportunities for further refinement. Lastly, it concludes by consolidating the proposed contributions, outlining the limitations of the current work, and suggesting promising future research directions.
Il calcolo quantistico rappresenta un paradigma innovativo con il potenziale di rivoluzionare il campo dell’informatica, offrendo la possibilità di colmare il divario tra i limiti degli attuali processori classici e le prestazioni richieste dalle applicazioni moderne. Tuttavia, la costruzione di calcolatori quantistici affidabili presenta ancora sfide significative. I qubit all’interno delle Unità di Elaborazione Quantistiche (QPU) risultano infatti altamente instabili e vulnerabili al rumore generato da diverse fonti, compromettendo in ultima istanza l’esecuzione affidabile di molte applicazioni su processori quantistici. Per affrontare tali problematiche, l’attuale generazione di dispositivi quantistici integra meccanismi di Correzione degli Errori Quantistici (QEC) per garantire il recupero della computazione, diventando un componente fondamentale per raggiungere la tolleranza agli errori. Tuttavia, l’intero processo di QEC è soggetto a vincoli rilevanti, tra cui la latenza della fase di decodifica degli errori, che rappresenta la limitazione più critica. In tale contesto, le FPGA emergono come una piattaforma promettente per l’implementazione di sistemi di QEC. Grazie alla loro latenza deterministica e alle capacità di calcolo spaziale, esse consentono di rispettare rigorosi vincoli temporali mantenendo al contempo un consumo energetico ridotto. Gli approcci presenti in letteratura che sfruttano le FPGA per la QEC hanno prodotto risultati incoraggianti, ma vi è ancora ampio margine di miglioramento. In particolare, il settore manca di un consenso sui criteri più rilevanti da considerare nell’implementazione di sistemi di QEC su hardware FPGA, lasciando il panorama della ricerca frammentato. In pi` u, oltre all’esigenza di scalabilità e ottimizzazione della latenza, le soluzioni future dovranno necessariamente essere adattabili alle nuove tecnologie di QPU, così da tenere il passo con la rapida evoluzione del settore. Strumenti di astrazione e modelli adeguati risultano inoltre indispensabili per fornire agli ingegneri quantistici i mezzi per personalizzare efficacemente i sistemi di QEC, assicurandone così la reale usabilità. In questo contesto, questa tesi indaga come la tecnologia FPGA possa contribuire all’avanzamento del calcolo quantistico attraverso un’implementazione efficiente dei sistemi di QEC, permettendo di andare poi a sfruttare il pieno potenziale di questo nuovo paradigma. In particolare, vengono introdotte una serie di metodologie finalizzate allo sviluppo di decodificatori di QEC ad alte prestazioni ed elevata efficienza energetica, sfruttando le capacità delle FPGA. La tesi inizia da una revisione approfondita dello Stato dell’Arte e, a partire da tale analisi, propone un insieme di pilastri utili a guidare il benchmarking e la valutazione di nuove soluzioni. Successivamente, vengono presentate due metodologie distinte per la progettazione di decodificatori di QEC basati su FPGA. La prima, QASBA, implementa interamente in hardware un algoritmo di decodifica degli errori quantistici considerato gold standard, garantendo un’elevata accuratezza. Per rafforzarne l’applicabilità in scenari reali, viene inoltre introdotto un framework di automazione che consente agli ingegneri quantistici di personalizzare i decodificatori basati su QASBA in funzione delle caratteristiche del sistema quantistico di riferimento. La seconda metodologia costituisce il fondamento di QUEKUF, che invece combina l’efficienza un algoritmo euristico di decodifica con la potenza computazionale delle FPGAs per migliorare le prestazioni. Anche in questo caso, per favorirne l’impiego pratico, QUEKUF integra un modello di ottimizzazione delle risorse orientato alla riduzione della latenza, concepito per assistere gli utenti finali nell’individuazione della configurazione architetturale più efficace in base all’ambiente quantistico in cui si va ad operare. Infine, la tesi colloca tali contributi all’interno del più ampio Stato dell’Arte, valutandone i vantaggi e individuando margini di ulteriore miglioramento. Infine, la tesi conclude riassumendo i contributi proposti, delineando le limitazioni del lavoro svolto e suggerendo direzioni promettenti per ricerche future.
On the role of reconfigurable systems in quantum computing: the case of quantum error correction
Beatrice, Branchini
2026
Abstract
Quantum computing is a novel paradigm that can potentially disrupt the Computer Science field. It promises to overcome the gap between the limitations of current classical processors and the performance demanded by modern applications. Nevertheless, significant challenges persist in constructing reliable quantum computers. Qubits within today’s Quantum Processing Units (QPUs) remain highly unstable and susceptible to noise originating from various sources, ultimately preventing the reliable execution of many applications on quantum processors. To address these issues, the current generation of quantum devices integrates Quantum Error Correction (QEC) mechanisms to recover the computation, thus representing a critical component toward fault tolerance. However, the QEC process faces several constraints, with the latency of error decoding being the most critical limitation. In this scenario, Field Programmable Gate Arrays (FPGAs) have arisen as a promising platform for deploying QEC engines. Their deterministic latency and spatial computing capabilities enable them to meet strict timing requirements, maintaining low power consumption at the same time. Existing approaches leveraging FPGAs for QEC have delivered encouraging results, yet substantial room for improvement remains. In particular, the field lacks consensus on the most significant criteria to deploy and assess QEC engines on FPGA, leaving the research landscape fragmented. Furthermore, beyond the pressing needs of latency optimizations, future solutions must also remain adaptable to emerging QPU technologies to keep pace with this rapidly evolving domain. Lastly, abstraction frameworks and modeling tools are essential to equip quantum engineers with the means to effectively customize QEC engines, ensuring their practical usability. In this context, this dissertation investigates how FPGA technology can advance the field of quantum computing through efficient deployment of QEC engines, enabling the full potential of this emerging paradigm. Specifically, it introduces a series of methodologies to develop high-performance and energy-efficient QEC decoders by leveraging FPGAs. The dissertation begins with a comprehensive review of the State of the Art and, building on this analysis, proposes a set of evaluation pillars to guide the benchmarking and evaluation of novel solutions. It then presents two distinct methodologies for designing FPGA-based QEC decoders. The first, QASBA, implements a gold-standard quantum error decoding algorithm entirely in hardware, ensuring high accuracy. To enhance its practical applicability, an automation framework is also introduced to enable quantum engineers to customize QASBA-based decoders according to the characteristics of the target quantum system. The second methodology forms the basis of QUEKUF, which combines the efficiency of a heuristic decoding algorithm with the computing power of FPGAs to achieve superior performance. Also in this case, to facilitate practical deployment, QUEKUF incorporates a latency-oriented resource optimization model that assists final users in selecting the most effective architectural configuration for their specific quantum environment. This dissertation then contextualizes these contributions within the State of the Art, evaluating their advantages and highlighting opportunities for further refinement. Lastly, it concludes by consolidating the proposed contributions, outlining the limitations of the current work, and suggesting promising future research directions.| File | Dimensione | Formato | |
|---|---|---|---|
|
PhD_Thesis-2.pdf
accesso solo da BNCF e BNCR
Licenza:
Tutti i diritti riservati
Dimensione
2.45 MB
Formato
Adobe PDF
|
2.45 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/360940
URN:NBN:IT:POLIMI-360940