[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-120809-en":3,"doc-seo-120809-105":29,"detail-sidebar-cat-0-en-105":89},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":20,"language":21,"language_code":22,"site_id":23,"html_lang":22,"table_of_contents":24,"faqs":25,"seo_title":26,"seo_description":14,"update_tm":27,"read_time":28},120809,8796095360427,"Lucas Martin","https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d",8,"Research & Report","Equalization Tuning of the PCIe Physical Layer by Using Machine Learning in Industrial Post-silicon Validation - Abstract","High-speed computer platforms make post-silicon validation an intensive industrial task, especially for HSIO links where Tx and Rx equalizers must be tuned efficiently. PCIe, a critical HSIO interface, is evolving toward Gen6 speeds, but higher rates intensify signal-integrity issues, channel attenuation, distortion, and inter-symbol interference. The work addresses EQ coefficient selection by replacing map-based, expert-driven processes with machine learning: it clusters post-silicon data by channel conditions using unsupervised learning and then trains Gaussian-process regression models to predict eye-diagram margins and optimize tuning settings, validated on functional measurements of an industrial platform.","We3G-2  \nEqualization Tuning of the PCIe Physical Layer by Using Machine Learning in Industrial Post-silicon Validation  \nFrancisco E. Rangel-Patiño\\#*1, Andres Viveros-Wacher\\#2, Chintan Rajyaguru$3, Edgar A. VegaOchoa\\#4, Sofia D. Rodriguez-Saenz\\#5, Johana L. Silva-Cortes\\#6, Hemanth Shival$7, and José E.  \nRayas-Sánchez*8,  \n\\# Intel Corp. Zapopan, Jalisco, 45019 Mexico  \n$ Intel Corp. Folsom, CA, 95630 USA  \n* Department of Electronics, Systems, and Informatics, ITESO – The Jesuit University of Guadalajara, Tlaquepaque, Jalisco, 45604 Mexico  \n1francisco.rangel, 2andres.viveros.wacher, 3chintan.rajyaguru, 4edgar.vega.ochoa, 5sofia.d.rodriguez.saenz,6johana.l.silva.cortes, [7](7hemanth.shival{@intel.com})[hemanth.shival{@intel.com}](7hemanth.shival{@intel.com}), [8](8erayas@iteso.mx)[erayas@iteso.mx](8erayas@iteso.mx)  \nABSTRACT  \nThe increasing complexity of high-speed computer platforms has made post-silicon validation a highly demanding industrial task. A large portion of the circuits to be validated in modern microprocessors corresponds to high-speed input/output (HSIO) links, imposing the need to efficiently tune the transmitter (Tx) and receiver (Rx) equalizers [1] .  \nPeripheral component interconnect express (PCIe) is one of the most complex HSIO interfaces and the primary interface for a host central processing unit (CPU) to connect with input/output (I/O) devices. PCIe has been continuously evolving and the new PCIe Gen6 specification has reached a data rate of 64 giga-transfers per second (GT/s) . However, as transmission speeds increase, undesired signal integrity effects are more severe, causing the signals to become more susceptible to errors [2] . Additionally, PCIe channels are bandwidth-limited by default, causing large signal attenuation at high frequencies. This generates distortion and spreading of the transmitted data over multiple symbols, exacerbating inter-symbol interference (ISI), which can make the signal unreadable at the Rx, producing bit errors. The most practical solution to this problem is signal conditioning to open the eye diagram [3] .  \nPCIe specification defines an adaptive mechanism for equalization (EQ) to determine the optimum values of the Tx and Rx EQ coefficients within a fixed time limit, across the allowed channel types. The most widely used current method consists of using maps of EQ coefficients, which are obtained from massive eye diagram measurements. The EQ maps are used to characterize the PCIe link across different channel losses and devices. Once the full characterization is completed, the best Tx EQ values are selected based on the input of an experienced validation engineer. This is a very time-consuming process and prone to human errors.  \nMachine learning algorithms are useful to build statistical models from examples, which are then used to make predictions when faced with cases not seen before [4] . Unsupervised machine learning algorithms are designed to learn patterns from untagged data. On the other hand, supervised machine learning models are trained to predict outputs from a given set of inputs. The large volume of data generated from typical post-silicon testing suggests the application of machine learning techniques to identify underlying patterns, such as channel effects on the analog behaviour ofthe HSIO link.  \nIn this work, we first use unsupervised machine learning techniques [5] to cluster all available post-silicon data from different channels, dividing them into distinct sets of channel conditions. We then develop statistical supervised machine learning models [6], based on Gaussian process regression (GPR), to predict the eye diagram margins within each data subset. We finally optimize the GPR-based models to obtain the optimal tuning settings for the specific channels. Our proposed method is validated by measurements of the functional eye diagram of an actual industrial computer platform.  \n[1] F. E. Rangel-Patiño, J. L. Chávez-Hurtado, A. Viveros-","cbCaioHxcDQtKd0B","https://ap.wps.com/l/cbCaioHxcDQtKd0B","pdf",134341,1,"English","en",105,"# Overview\n## Post-silicon validation and PCIe challenges\n## EQ tuning approaches and limitations\n# Machine learning method\n## Unsupervised clustering of post-silicon data\n## Supervised prediction with GPR\n## Optimization of tuning settings and validation","[{\"question\":\"Why is post-silicon validation particularly demanding for PCIe HSIO links?\",\"answer\":\"Because high-speed links require efficient Tx/Rx equalizer tuning, and higher transmission rates amplify distortion, attenuation, and inter-symbol interference, which can lead to bit errors at the receiver.\"},{\"question\":\"What problem does the traditional PCIe equalization tuning method face?\",\"answer\":\"It relies on EQ coefficient maps derived from extensive eye measurements and then selects best Tx EQ values based on expert input, making the process time-consuming and vulnerable to human errors.\"},{\"question\":\"How does the proposed machine learning approach determine optimal EQ tuning settings?\",\"answer\":\"It first uses unsupervised learning to cluster post-silicon data into channel-condition subsets, then trains supervised Gaussian process regression models to predict eye-diagram margins within each subset, and finally optimizes the models to produce tuning settings for the specific channels.\"}]","Equalization Tuning of the PCIe Physical Layer by Using Machine Learning in Industrial Post-silicon Validation - Abstract | PDF",1785732130,3,{"code":4,"msg":30,"data":31},"ok",{"site_id":23,"language":22,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":84,"head_meta":86,"extra_data":88,"updated_unix":27},"equalization-tuning-of-the-pcie-physical-layer-by-using-machine-learning-in-industrial-post-silicon-validation-abstract","",{"@graph":35,"@context":83},[36,52,66],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,49],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":28},"https://docshare.wps.com/document/research-report/",{"item":50,"name":13,"@type":42,"position":51},"https://docshare.wps.com/document/equalization-tuning-of-the-pcie-physical-layer-by-using-machine-learning-in-industrial-post-silicon-validation-abstract/120809/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":22,"description":14,"dateModified":60,"datePublished":60,"encodingFormat":59,"isAccessibleForFree":61,"interactionStatistic":62},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":40,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":63,"interactionType":64,"userInteractionCount":4},"InteractionCounter",{"@type":65},"ViewAction",{"@type":67,"mainEntity":68},"FAQPage",[69,75,79],{"name":70,"@type":71,"acceptedAnswer":72},"Why is post-silicon validation particularly demanding for PCIe HSIO links?","Question",{"text":73,"@type":74},"Because high-speed links require efficient Tx/Rx equalizer tuning, and higher transmission rates amplify distortion, attenuation, and inter-symbol interference, which can lead to bit errors at the receiver.","Answer",{"name":76,"@type":71,"acceptedAnswer":77},"What problem does the traditional PCIe equalization tuning method face?",{"text":78,"@type":74},"It relies on EQ coefficient maps derived from extensive eye measurements and then selects best Tx EQ values based on expert input, making the process time-consuming and vulnerable to human errors.",{"name":80,"@type":71,"acceptedAnswer":81},"How does the proposed machine learning approach determine optimal EQ tuning settings?",{"text":82,"@type":74},"It first uses unsupervised learning to cluster post-silicon data into channel-condition subsets, then trains supervised Gaussian process regression models to predict eye-diagram margins within each subset, and finally optimizes the models to produce tuning settings for the specific channels.","https://schema.org",{"og:url":50,"og:type":85,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":87,"canonical":50},"index,follow",{"doc_id":7,"site_id":23},{"code":4,"msg":5,"data":90},[91,95,99,103,108,113,118,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":92,"show_sort_weight":93,"slug":94},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":96,"show_sort_weight":97,"slug":98},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":45,"category_name":100,"show_sort_weight":101,"slug":102},"Exam",70,"exam",{"id":104,"doc_module":4,"doc_module_name":45,"category_name":105,"show_sort_weight":106,"slug":107},5,"Comic",60,"comic",{"id":109,"doc_module":4,"doc_module_name":45,"category_name":110,"show_sort_weight":111,"slug":112},6,"Technology",50,"technology",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":104,"slug":136},19,"General","general"]