Tracing Program Research Articles

Comparative evaluation of commercially available AI-based cephalometric tracing programs

ObjectivesCompare the accuracy and diagnostic concordance of three commercially available AI-based lateral cephalometric tracing software.Materials and methodsSixty-three lateral cephalometric radiographs were analyzed using semi-automatic (Dolphin Imaging Systems LLC) and AI-based software programs (WebCeph™, Cephio, and Ceppro DDH Inc.). Intra- and inter-observer reliability were assessed for human expert measurements, and repeated-measures one-way ANOVA was used to compare the AI and human expert measurements. The diagnostic performance was evaluated using sensitivity and specificity tests.ResultsHuman expert reliability was excellent (ICC > 0.9) for most cephalometric parameters. Compared to human experts, significant differences were observed for all three AI-based cephalometric programs (WebCeph™ – 10 of 11, Cephio – 7 of 11, and Ceppro DDH Inc. – 7 of 11 cephalometric measurements). Variations exceeding two units were noted for most parameters, and differences in defining the sagittal and vertical skeletal patterns, dental, and soft tissue characteristics were observed.ConclusionAll three AI-based tracing programs showed inaccuracies compared to human expert measurements and lacked reliability in measuring key cephalometric parameters. Clinicians should exercise caution when relying solely on AI-based analyses for orthodontic treatment planning and assessment.

How much does it cost to retain antiretroviral therapy (ART) clients in their first year? Routine financial costs of retention interventions at Lighthouse Trustin Lilongwe, Malawi.

Introduction Antiretroviral therapy (ART) improves the health of people living with HIV (PLHIV). However, a high loss to follow-up, particularly in the first year after ART initiation, is problematic. The financial expenses related to client retention in low- and middle-income countries (LMICs) in sub-Saharan Africa are not well understood. This study aimed to comprehensively assess and quantify the financial costs associated with routine ART retention services for clients in their first 12 months in care at Lighthouse Trust's (LT) Martin Preuss Centre (MPC), a large, public ART clinic in Lilongwe, Malawi. Methods We performed activity-based microcosting using routine data to assess the expenses related to routine ART retention services at the MPC for 12 months, from January to December 2021. MPC provides an "ART Buddy" from ART initiation to 12 months. The MPC's Back-to-Care (B2C) program traces clients who miss ART visits by at least 14 days. Clients may be traced and return to care multiple times per year. We assessed client retention costs for the first 12 months of treatment with ART and conducted a sensitivity analysis. Results The total annual cost of ART retention interventions at MPC was $237,564. The proactive Buddy phase incurred $108,504; personnel costs contributed $97,764. In the reactive B2C phase, the total cost was $129,060, with personnel expenses remaining substantial at $73,778. The Buddy unit cost was $34 per client. The reactive B2C intervention was $17 per tracing event. On average, the unit cost for ART retention in the first year of ART averaged $22 per client. Conclusion This study sheds light on the financial dimensions of ART retention during the first year on ART among clients in a routine public setting. ART retention is both costly and critical for helping clients adhere to visits and remain in care. Continued investment in the human resources needed for both proactive and reactive retention efforts is critical to engaging and retaining patients on lifetime ART.

Making Sense of Multi-threaded Application Performance at Scale with NonSequitur

Modern multi-threaded systems are highly complex. This makes their behavior difficult to understand. Developers frequently capture behavior in the form of program traces and then manually inspect these traces. Existing tools, however, fail to scale to traces larger than a million events. In this paper we present an approach to compress multi-threaded traces in order to allow developers to visually explore these traces at scale. Our approach is able to compress traces that contain millions of events down to a few hundred events. We use this approach to design and implement a tool called NonSequitur. We present three case studies which demonstrate how we used NonSequitur to analyze real-world performance issues with Meta's storage engine RocksDB and MongoDB's storage engine WiredTiger, two complex database backends. We also evaluate NonSequitur with 42 participants on traces from RocksDB and WiredTiger. We demonstrate that, in some cases, participants on average scored 11 times higher when performing performance analysis tasks on large execution traces. Additionally, for some performance analysis tasks, the participants spent on average three times longer with other tools than with NonSequitur.

Toroidal distribution of heat load on castellated plasma facing components for lower divertor target in EAST

In tokamak devices, the performance of plasma facing components (PFCs) under high heat loads is a key challenge for achieving long-pulse and high-performance plasma operations. The lower divertor of EAST adopts an ITER-like cassette modules (CMs) structure. However, regular distribution of damage along the toroidal direction have been observed on the divertor target plates, which not only affects the lifespan of PFCs but also limits the long-pulse and high-performance plasma operations. Therefore, investigating the toroidal distribution of the heat load on the divertor target in EAST is crucial for understanding the mechanism of damage on the target plate surface. The toroidal characteristics of heat loads on CMs of the lower divertor in EAST was analyzed using a three-dimensional magnetic field line tracing program (PFCFlux). For the inner vertical target (IVT) and outer vertical target (OVT) within a single CM, the heat load varies monotonically along the toroidal direction. In contrast, the heat load distribution on the outer horizontal target (OHT) is uniform, except at the chamfers, which is directly correlated with the target’s geometric structure and the magnetic field configuration. The surface heat load distribution exhibits toroidal non-uniformity, with the heat deposition pattern being generally consistent and displaying periodic variations. The toroidally uneven heat load distribution is attributed to the change in the magnetic field’s tilt angle. Additionally, the peak heat load is observed at the chamfer positions between cassettes, and an analysis is conducted on the relationship between peak heat load and the size of the chamfer. Smaller tilt angles result in lower surface heat load. Furthermore, a comparative analysis of peak heat loads at the chamfer under different plasma current conditions reveals that higher plasma currents generally result in increased peak heat loads. This investigation into the toroidal heat load distribution on the divertor surface contributes to understanding of the damage mechanism of W PFC as the target plate and provides valuable data for divertor design of future fusion devices.

Sentiment Analysis of Twitter Posts Related to a COVID-19 Test & Trace Program in NYC.

As part of a program evaluation of the New York City Test & Trace program (T2)-one of the largest such programs in the USA-we conducted a study to assess how implementing organizations (NYC Health+ Hospitals, government agencies, CBOs) communicated information about the T2 program on Twitter. Study aims were as follows: (1) quantify user engagement of posts ("tweets") about T2 by NYC organizations on Twitter and (2) examine the emotional tone of social media users' T2-related tweets in our sample of 1987 T2-related tweets. Celebrities and CBOs generated more user engagement (0.26% and 0.07%, respectively) compared to government agencies (e.g., Mayor's Office, 0.0019%), reinforcing the value of collaborating with celebrities and CBOs in social media public health campaigns. Sentiment analysis revealed that positive tweets (46.5%) had higher user engagement than negative tweets (number of likes: R2 = .095, p < .01), underscoring the importance of positively framing messages for effective public health campaigns.

Scale-Up of COVID-19 Testing Services in NYC, 2020-2021: Lessons Learned to Maximize Reach, Equity and Timeliness.

During infectious disease epidemics, accurate diagnostic testing is key to rapidly identify and treat cases, and mitigate transmission. When a novel pathogen is involved, building testing capacity and scaling testing services at the local level can present major challenges to healthcare systems, public health agencies, and laboratories. This mixed methods study examined lessons learned from the scale-up of SARS-CoV-2 testing services in New York City (NYC), as a core part of NYC's Test & Trace program. Using quantitative and geospatial analyses, the authors assessed program success at maximizing reach, equity, and timeliness of SARS-CoV-2 diagnostic testing services across NYC neighborhoods. Qualitative analysis of key informant interviews elucidated key decisions, facilitators, and barriers involved in the scale-up of SARS-CoV-2 testing services. A major early facilitator was the ability to establish working relationships with private sector vendors and contractors to rapidly procure and manufacture necessary supplies locally. NYC residents were, on average, less than 25min away from free SARS-CoV-2 diagnostic testing services by public transport, and services were successfully directed to most neighborhoods with the highest transmission rates, with only one notable exception. A key feature was to direct mobile testing vans and rapid antigen testing services to areas based on real-time neighborhood transmission data. Municipal leaders should prioritize fortifying supply chains, establish cross-sectoral partnerships to support and extend testing services, plan for continuous testing and validation of assays, ensure open communication feedback loops with CBO partners, and maintain infrastructure to support mobile services during infectious disease emergencies.

Barriers and Facilitators to Trust in the COVID-19 New York City Test and Trace Program.

Barriers and Facilitators to Trust in the COVID-19 New York City Test and Trace Program.

Abstract C053: Pancreatic cancer patient-derived organoids, preclinical tools for therapeutic profiling, patient response prediction, and tumor evolution

Abstract Pancreatic cancer (PC) is characterized by an aggressive biology and an exceptionally high tumor heterogeneity that causes considerable variations in response to chemotherapies, targeted agents, and immunotherapies. Patient-derived organoids (PDOs) accurately reflect parental tumor biological and molecular features and may represent powerful preclinical avatars to predict drug response and support clinical decision-making. We generated a living organoid biobank of &amp;gt;100 PC PDO lines from treatment-naïve and pretreated primary tumor and metastases with a reliable efficacy (60.8%), and previously developed a pharmacotyping-guided prediction model to prognosticate patient therapy response in an initial feasibility trial (Beutel AK et al, 2021). Real-life chemotherapy responses in 46 patients (for a total of 93 therapy lines) matched to pharmacotyping-informed predictions with overall 73.1% accuracy, 85.4% sensitivity (responsiveness), and 63.5% specificity (non-responsiveness). Overall, our model allowed a successful drug-response prediction in naïve patients with an accuracy of 73.0% and 70.0% for first and second-line regimens. Prediction power was nevertheless lower in pretreated and heavily pretreated patients with a precision of 68.2% and 50.0% for subsequent chemotherapy lines, respectively. Interestingly, PDO-based pharmacotyping also precisely predicted patient clinical outcome in the neoadjuvant and adjuvant settings, with respective accuracies of 83.3% and 71.4%. Retrospective analysis of patient clinical data in palliative setting finally showed that the administration of a regimen predicted to be efficient ultimately translated into a significantly longer progression-free survival. Implementing automation and drug screening miniaturization to our workflow also significantly enhanced our process capacity, reduced time before pharmacotyping, and improved compliance to internal quality controls. Tracking clonal evolution in longitudinal biopsies (a set of seven PDO pairs derived from the same patients at two distinct timepoints) revealed lower mutational burden and therapy-driven genetic alterations upon progression. Notably, this approach revealed a CHEK2-mutated patient responded over time to PARP inhibitor maintenance therapy, aligning with our predictions and underscoring the robustness of our method. Single nucleus-resolved RNA and ATAC-seq analysis of a unique longitudinally-collected case uncovered transcriptomic and epigenetic dynamics associated with an oncogenic FGFR2 fusion and with a subsequent resistance-mediating mutation upon targeted treatment with FGFR2 inhibitors. Particularly, FGFR2 constitutive activation correlated with aberrant epigenetic trace and transcription programs of downstream targets as PI3K-AKT and RAS family members and of RNA polymerase II transcription machinery. Here, we report a robust and clinically-relevant preclinical tool for drug-response prediction, one more step towards a PDO therapeutic profiling-guided personalized medicine in clinical routine. Citation Format: Johann Gout, Yazid J Resheq, Jessica Lindenmayer, Julian D Schwab, Elodie Roger, Johann M Kraus, Thomas Ettrich, Alica K Beutel, Lukas Perkhofer, Hans A Kestler, Thomas Seufferlein, Alexander Kleger. Pancreatic cancer patient-derived organoids, preclinical tools for therapeutic profiling, patient response prediction, and tumor evolution [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pancreatic Cancer Research; 2024 Sep 15-18; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(17 Suppl_2):Abstract nr C053.

Impact of a University-Led COVID-19 Case Investigation and Contact Tracing Program.

The coronavirus disease 2019 (COVID-19) pandemic devastated societies and economies worldwide. Given the major disruptions to higher education, reflection on university responses to the COVID-19 pandemic may provide insights for future outbreaks. Here, we describe the epidemiology of COVID-19 on the Emory University campus during the 2020-2021 academic year and provide an evaluation of the performance of a university-led program with the purpose of describing the effectiveness of efforts to augment the public health authority's case investigation and contact tracing efforts during a public health emergency. Evaluation of a case investigation and contact tracing program regarding operations, timeliness, and performance. We analyzed quality metrics to determine the proportion of cases and contacts interviewed and the time to completion of each step from case diagnosis to testing of contacts. During the 2020-2021 academic year, 1267 COVID-19 cases among Emory students, faculty, and staff were confirmed by polymerase chain reaction, with 1132 reported close contacts. Among cases, the median test turnaround time was 1 day (interquartile range: 1, 2). Among both cases and close contacts, 98% were successfully interviewed. The team called a majority of cases on the same day as their test result was reported to the program (87%; n =1052). Almost all (98%; n =1247) cases completed isolation or were advised to isolate during the review period. Close to half (46%; n =513) of contacts interviewed began quarantine before their interview. Among close contacts interviewed, 13% (n=145) subsequently converted to an index case. The impact and performance of Emory's program may provide useful and actionable data for future university-led infectious disease outbreak response programs. The program structure, performance metrics, and information collected via interviews provide practical implications and an organized structure to guide other programs during future outbreaks.

Meeting Social Needs in a Crisis Context: Lessons Learned from Integrating the 'Take Care Initiative' into New York City's Testing and Contact Tracing Program.

The COVID-19 pandemic highlighted the importance of addressing social needs in a crisis context. Some US jurisdictions integrated a social service component into case investigation and contact tracing (CI/CT) programs, including the New York City (NYC) Test & Trace (T2) Program; the Take Care initiative referred NYC residents who tested positive or were exposed to COVID-19 to services to support isolation and quarantine and meet basic needs. More research is needed to determine effective implementation strategies for integrating social needs provision into CI/CT programs. To identify barriers and facilitators to the implementation of the Take Care initiative, we conducted key informant interviews with program staff, community-based organization partners, and cases and contacts as part of a larger evaluation of the T2 program. Interviews were recorded, transcribed, and analyzed using rapid qualitative methods. Key facilitators to implementation included utilizing a case management software system, employing strategies to encourage service uptake, leveraging cross-agency collaborations, and partnering with community-based organizations for resource navigation. Barriers identified included external management of the software system, challenges reaching and engaging the public, administrative complications due to shifting collaborations, and management of CBO partners' structure and hiring. Based on our findings, we provide recommendations to support effective planning and implementation of social needs service provision in a crisis context. Future research should focus on testing promising implementation strategies highlighted in this study and applying them to varied contexts and crisis situations.

Emulating target trials of postexposure vaccines using observational data.

Postexposure vaccination has the potential to prevent or modify the course of clinical disease among those exposed to a pathogen. However, due to logistical constraints, postexposure vaccine trials have been difficult to implement in practice. In place of trials, investigators have used observational data to estimate the effectiveness or optimal timing window for postexposure vaccines, but the relationship between these analyses and those that would be conducted in a trial is often unclear. Here, we define several possible target trials for postexposure vaccination and show how, under certain conditions, they can be emulated using observational data. We emphasize the importance of the incubation period and the timing of vaccination in trial design and emulation. As an example, we specify a protocol for postexposure vaccination against mpox and provide a step-by-step description of how to emulate it using data from a healthcare database or contact tracing program. We further illustrate some of the benefits of the target trial approach through simulation.

Lessons Learned from the Launch and Implementation of the COVID-19 Contact Tracing Program in New York City: a Qualitative Study.

On June 1, 2020, NYC Health + Hospitals, in partnership with the NYC Department of Health and Mental Hygiene, other city agencies, and a large network of community partners, launched the New York City Test & Trace (T2) COVID-19 response program to identify and isolate cases, reduce transmission through contact tracing, and provide support to residents during isolation or quarantine periods. In this paper, we describe lessons learned with respect to planning and implementation of case notification and contact tracing. Our findings are based on extensive document review and analysis of 74 key informant interviews with T2 leadership and frontline staff, cases, and contacts conducted between January and September 2022. Interviews elicited respondent background, history of program development, program leadership and structure, goals of the program, program evolution, staffing, data systems, elements of community engagement, trust with community, program reach, timeliness, equity, general barriers and challenges, general facilitators and best practices, and recommendations/improvement for the program. Facilitators and barriers revealed in the interviews primarily revolved around hiring and managing staff, data and technology, and quality of interactions with the public. Based on these facilitators and barriers, we identify suggestions to support effective planning and response for future case notification and contact tracing programs, including recommendations for planning during latent periods, case management and data systems, and processes for outreach to cases and contacts.

Evaluation of the New York City COVID-19 case investigation and contact tracing program: a cascade of care analysis

BackgroundNew York City (NYC) was the first COVID-19 epicenter in the United States and home to one of the country’s largest contact tracing programs, NYC Test & Trace (T2). Understanding points of attrition along the stages of program implementation and follow-up can inform contact tracing efforts for future epidemics or pandemics. The objective of this study was to evaluate the completeness and timeliness of T2 case and contact notification and monitoring using a “cascade of care” approach.MethodsThis cross-sectional study included all SARS-CoV-2 cases and contacts reported to T2 from May 31, 2020 to January 1, 2022. Attrition along the “cascade of care” was defined as: (1) attempted, (2) reached, (3) completed intake (main outcome), (4) eligible for monitoring, and (5) successfully monitored. Timeliness was assessed: (1) by median days from a case’s date of testing until their positive result was reported to T2, (2) from result until the case was notified by T2, and (3) from a case report of a contact until notification of the contact.ResultsA total of 1.45 million cases and 1.38 million contacts were reported to T2 during this period. For cases, attrition occurred evenly across the first three cascade steps (~-12%) and did not change substantially until the Omicron wave in December 2021. During the Omicron wave, the proportion of cases attempted dropped precipitously. For contacts, the largest attrition occurred between attempting and reaching (-27%), and attrition rose with each COVID-19 wave as contact volumes increased. Attempts to reach contacts discontinued entirely during the Omicron wave. Overall, 67% of cases and 49% of contacts completed intake interviews (79% and 57% prior to Omicron). T2 was timely, with a median of 1 day to receive lab results, 2 days to notify cases, and < 1 day to notify contacts.ConclusionsT2 provided a large volume of NYC residents with timely notification and monitoring. Engagement in the program was lower for contacts than cases, with the largest gap coming from inability to reach individuals during call attempts. To strengthen future test-and-trace efforts, strategies are needed to encourage acceptance of local contact tracer outreach attempts.

A large-scale and PCR-referenced vocal audio dataset for COVID-19

The UK COVID-19 Vocal Audio Dataset is designed for the training and evaluation of machine learning models that classify SARS-CoV-2 infection status or associated respiratory symptoms using vocal audio. The UK Health Security Agency recruited voluntary participants through the national Test and Trace programme and the REACT-1 survey in England from March 2021 to March 2022, during dominant transmission of the Alpha and Delta SARS-CoV-2 variants and some Omicron variant sublineages. Audio recordings of volitional coughs, exhalations, and speech were collected in the ‘Speak up and help beat coronavirus’ digital survey alongside demographic, symptom and self-reported respiratory condition data. Digital survey submissions were linked to SARS-CoV-2 test results. The UK COVID-19 Vocal Audio Dataset represents the largest collection of SARS-CoV-2 PCR-referenced audio recordings to date. PCR results were linked to 70,565 of 72,999 participants and 24,105 of 25,706 positive cases. Respiratory symptoms were reported by 45.6% of participants. This dataset has additional potential uses for bioacoustics research, with 11.3% participants self-reporting asthma, and 27.2% with linked influenza PCR test results.

Experience of Public Health Departments in Implementation of COVID-19 Case Investigation and Contact Tracing Programs.

Case investigation and contact tracing (CI/CT) are fundamental public health efforts widely used during the COVID-19 pandemic to mitigate transmission. This study investigated how state, local, and tribal public health departments used CI/CT during the COVID-19 pandemic, including CI/CT methodology, staffing models, training and support, and efforts to identify or prioritize populations disproportionately affected by COVID-19. During March and April 2022, we conducted key informant interviews with up to 3 public health officials from 43 state, local, and tribal public health departments. From audio-recorded and transcribed interviews, we used the framework method to analyze key themes. Major adjustments to CI/CT protocols during the pandemic included (1) prioritizing populations for outreach; (2) implementing automated outreach for nonprioritized groups, particularly during COVID-19 surges; (3) discontinuing contact tracing and focusing exclusively on case investigation; and (4) adding innovations to provide additional support. Key informants also discussed the utility of having backup staffing to support overwhelmed public health departments and spoke to the difficulty in "right-sizing" the public health workforce, with COVID-19 surges leaving public health departments understaffed as case rates rose and overstaffed as case rates fell. When addressing future epidemics or outbreaks, public health officials should consider strategies that improve the effectiveness of CI/CT efforts over time, such as prioritizing populations based on disproportionate risk, implementing automated outreach, developing models that provide flexible additional staffing resources as cases rise and fall among local public health departments, incorporating demographic data in laboratory reporting, providing community connections and support, and having a system of self-notification of contacts.

The First Computer Program

In 1837, the renowned mathematician and inventor Charles Babbage sketched the first computer program in history. It was a program for the Analytical Engine, an unfinished mechanical computer. The program consists of a sequence of arithmetic operations and the necessary variable addresses (memory locations) of the arguments and the result, displayed in tabular form like a program trace. The program calculates the solutions of a system of two linear equations in two unknowns.

Camouflage: Utility-Aware Obfuscation for Accurate Simulation of Sensitive Program Traces

Trace-based simulation is a widely used methodology for system design exploration. It relies on realistic traces that represent a range of behaviors necessary to be evaluated, containing a lot of information about the application, its inputs and the underlying system on which it was generated. Consequently, generating traces from real-world executions risks leakage of sensitive information. To prevent this, traces can be obfuscated before release. However, this can undermine their ideal utility, i.e., how realistically a program behavior was captured. To address this, we propose Camouflage, a novel obfuscation framework, designed with awareness of the necessary architectural properties required to preserve trace utility , while ensuring secrecy of the inputs used to generate the trace. Focusing on memory access traces, our extensive evaluation on various benchmarks shows that camouflaged traces preserve the performance measurements of the original execution, with an average τ correlation of 0.66. We model input secrecy as an input indistinguishability problem and show that the average security loss is 7.8%, which is better than traces generated from the state-of-the-art.

Differential Diagnosis, Prevention Measures, and Therapeutic Interventions for Enhanced Monkeypox (Mpox) Care.

Monkeypox (Mpox) is a rare viral disease that presents considerable challenges in healthcare settings, necessitating enhanced nursing care for effective management. This review thoroughly explores key aspects related to improving nursing care for Mpox. It commences by examining the background information on Mpox, encompassing its etiology, epidemiology, and modes of transmission. The differential diagnosis of Mpox is investigated, elucidating its clinical presentation, symptoms, and diagnostic methods to differentiate it from similar conditions. Prevention and control measures at both the public health and healthcare levels are scrutinized, including surveillance and reporting, contact tracing, isolation, and vaccination programs. In healthcare settings, infection prevention and control strategies, such as proper utilization of personal protective equipment, hand hygiene, and environmental management, are discussed. Furthermore, therapeutic interventions for Mpox, including symptomatic management, antiviral therapy, and supportive care, are outlined, with a specific emphasis on pain management, fever control, and psychosocial support. Nursing care strategies encompass patient assessment and monitoring, infection prevention strategies, psychosocial support, and patient education. The challenges encountered in enhancing nursing care for Mpox are acknowledged, along with research gaps and areas for further investigation. Finally, innovations in nursing practice for improved care, such as technology integration and simulation-based training, are explored. Enhancing nursing care in Mpox is crucial for positive patient outcomes, reducing transmission risks, and promoting overall well-being. By addressing the unique challenges, conducting further research, and embracing innovative practices, healthcare professionals, particularly nurses, can provide optimal care and contribute to better management of Mpox cases.

A novel multi-step-ahead approach for cloud server aging prediction based on hybrid deep learning model

Predicting software aging in cloud servers is essential for preventing cloud system breakdown and unexpected failures. However, the aging time series data in cloud servers often exhibits non-linearity and stochastic fluctuations, posing challenges for accurate prediction. To address this issue, this study proposes a novel multi-step ahead approach for cloud server aging prediction. Firstly, the original aging time series data is denoised using discrete wavelet transformation (DWT) and then reconstructed into multi-dimensional data by phase state reconstruction (PSR) to capture potential aging patterns. Subsequently, a deep long short-term memory network (DLSTM) is developed to predict the aging trend. Additionally, a stacked auto-encoder LSTM (SAE-LSTM) is designed to optimize the initial weight matrix of DLSTM to enhance prediction abilities. Finally, multi-step prediction experiments are conducted on our time series datasets from OpenStack cloud servers and a public dataset from Alibaba Cluster Trace Program. The comparative analysis of the proposed method with five baseline methods confirms its superiority in terms of prediction accuracy. For response time and free memory test sets, the range of root mean square error (RMSE) values are produced by the proposed method distributed in [0.1688–0.1931] and [139.05–146.05] respectively, with respect to the multi-step ahead prediction horizon from 1 up to 10. These results demonstrate that the proposed method can tackle the limitations of one-step ahead prediction, effectively extracting relevant aging features, and achieving low prediction accuracy. The study provides valuable insights into revealing future aging trends in cloud servers and offers a reference for implementing preventive maintenance and prognostics in cloud servers.

ShaderPerFormer: Platform-independent Context-aware Shader Performance Predictor

The ability to model and predict the execution time of GPU computations is crucial for real-time graphics application development and optimization. While there are many existing methodologies for graphics programmers to provide such estimates, those methods are often vendor-dependent, require the platforms to be tested, or fail to capture the contextual influences among shader instructions. To address this challenge, we propose ShaderPerFormer, a platform-independent, context-aware deep-learning approach to model GPU performance and provide end-to-end performance predictions on a per-shader basis. To provide more accurate predictions, our method contains a separate stage to gather platform-independent shader program trace information. We also provide a dataset consisting of a total of 54,667 fragment shader performance samples on 5 different platforms. Compared to the PILR and SH baseline methods, our approach reduces the average MAPE across five platforms by 8.26% and 25.25%, respectively.