More Like this

1. Automating CUDA Synchronization via Program Transformation

   https://doi.org/10.1109/ASE.2019.00075  

   Wu, Mingyuan; Zhang, Lingming; Liu, Cong; Tan, Shin Hwei; Zhang, Yuqun (November 2019, IEEE/ACM International Conference on Automated Software Engineering)

   While CUDA has been the most popular parallel computing platform and programming model for general purpose GPU computing, CUDA synchronization undergoes significant challenges for GPU programmers due to its intricate parallel computing mechanism and coding practices. In this paper, we propose AuCS, the first general framework to automate synchronization for CUDA kernel functions. AuCS transforms the original LLVM-level CUDA program control flow graph in a semantic-preserving manner for exploring the possible barrier function locations. Accordingly, AuCS develops mechanisms to correctly place barrier functions for automating synchronization in multiple erroneous (challenging-to-be-detected) synchronization scenarios, including data race, barrier divergence, and redundant barrier functions. To evaluate the effectiveness and efficiency of AuCS, we conduct an extensive set of experiments and the results demonstrate that AuCS can automate 20 out of 24 erroneous synchronization scenarios. 

   more » « less
2. Modeling and analyzing evaluation cost of CUDA kernels

   https://doi.org/10.1145/3434306  

   Muller, Stefan_K; Hoffmann, Jan (January 2021, Proceedings of the ACM on Programming Languages)

   General-purpose programming on GPUs (GPGPU) is becoming increasingly in vogue as applications such as machine learning and scientific computing demand high throughput in vector-parallel applications. NVIDIA's CUDA toolkit seeks to make GPGPU programming accessible by allowing programmers to write GPU functions, called kernels, in a small extension of C/C++. However, due to CUDA's complex execution model, the performance characteristics of CUDA kernels are difficult to predict, especially for novice programmers. This paper introduces a novel quantitative program logic for CUDA kernels, which allows programmers to reason about both functional correctness and resource usage of CUDA kernels, paying particular attention to a set of common but CUDA-specific performance bottlenecks. The logic is proved sound with respect to a novel operational cost semantics for CUDA kernels. The semantics, logic and soundness proofs are formalized in Coq. An inference algorithm based on LP solving automatically synthesizes symbolic resource bounds by generating derivations in the logic. This algorithm is the basis of RaCuda, an end-to-end resource-analysis tool for kernels, which has been implemented using an existing resource-analysis tool for imperative programs. An experimental evaluation on a suite of CUDA benchmarks shows that the analysis is effective in aiding the detection of performance bugs in CUDA kernels. 

   more » « less
3. Learning to Detect and Localize Multilingual Bugs

   https://doi.org/10.1145/3660804  

   Yang, Haoran; Nong, Yu; Zhang, Tao; Luo, Xiapu; Cai, Haipeng (July 2024, Proceedings of the ACM on Software Engineering)

   Increasing studies have shown bugs in multi-language software as a critical loophole in modern software quality assurance, especially those induced by language interactions (i.e., multilingual bugs). Yet existing tool support for bug detection/localization remains largely limited to single-language software, despite the long-standing prevalence of multi-language systems in various real-world software domains. Extant static/dynamic analysis and deep learning (DL) based approaches all face major challenges in addressing multilingual bugs. In this paper, we present xLoc, a DL-based technique/tool for detecting and localizing multilingual bugs. Motivated by results of our bug-characteristics study on top locations of multilingual bugs, xLoc first learns the general knowledge relevant to differentiating various multilingual control-flow structures. This is achieved by pre-training a Transformer model with customized position encoding against novel objectives. Then, xLoc learns task-specific knowledge for the task of multilingual bug detection/localization, through another new position encoding scheme (based on cross-language API vicinity) that allows for the model to attend particularly to control-flow constructs that bear most multilingual bugs during fine-tuning. We have implemented xLoc for Python-C software and curated a dataset of 3,770 buggy and 15,884 non-buggy Python-C samples, which enabled our extensive evaluation of xLoc against two state-of-the-art baselines: fine-tuned CodeT5 and zero-shot ChatGPT. Our results show that xLoc achieved 94.98% F1 and 87.24%@Top-1 accuracy, which are significantly (up to 162.88% and 511.75%) higher than the baselines. Ablation studies further confirmed significant contributions of each of the novel design elements in xLoc. With respective bug-location characteristics and labeled bug datasets for fine-tuning, our design may be applied to other language combinations beyond Python-C. 

   more » « less
4. As Code Testing: Characterizing Test Quality in Open Source Ansible Development

   https://doi.org/10.1109/ICST53961.2022.00031  

   Hassan, Mohammad Mehedi; Rahman, Akond (April 2022, 2022 IEEE Conference on Software Testing, Verification and Validation (ICST))

   Infrastructure as code (IaC) scripts, such as Ansible scripts, are used to provision computing infrastructure at scale. Existence of bugs in IaC test scripts, such as, configuration and security bugs, can be consequential for the provisioned computing infrastructure. A characterization study of bugs in IaC test scripts is the first step to understand the quality concerns that arise during testing of IaC scripts, and also provide recommendations for practitioners on quality assurance. We conduct an empirical study with 4,831 Ansible test scripts mined from 104 open source software (OSS) repositories where we quantify bug frequency, and categorize bugs in test scripts. We further categorize testing patterns, i.e., recurring coding patterns in test scripts, which also correlate with appearance of bugs. From our empirical study, we observe 1.8% of 4,831 Ansible test scripts to include a bug, and 45.2% of the 104 repositories to contain at least one test script that includes bugs. We identify 7 categories of bugs, which includes security bugs and performance bugs that are related with metadata extraction. We also identify 3 testing patterns that correlate with appearance of bugs: 'assertion roulette’, 'local only testing’, and 'remote mystery guest‘. Based on our findings, we advocate for detection and mitigation of the 3 testing patterns as these patterns can have negative implications for troubleshooting failures, reproducible deployments of software, and provisioning of computing infrastructure. 

   more » « less
5. The Role of Antennae in Heat Detection and Feeding Behavior in the Bed Bug (Hemiptera: Cimicidae)

   https://doi.org/10.1093/jee/toaa250  

   Gaire, Sudip; Schal, Coby; Mick, Russell; DeVries, Zachary (October 2020, Journal of Economic Entomology)

   Appel, Arthur (Ed.)

   Abstract The common bed bug (Cimex lectularius L.) is an obligate hematophagous ectoparasite that has significant impacts on human health and well-being. All life stages of bed bugs (except eggs) feed solely on blood, which is required to molt and reproduce. Bed bugs use multiple cues to locate their hosts, including heat, CO2, and body odors. Of these cues, detection of heat appears limited to a short distance of <3 cm. However, it remains unclear if bed bugs can detect radiant heat, what structure(s) are responsible for heat detection, and if heat detection via the antennae is required for feeding. In this study, bed bug response to radiant heat was evaluated using the two-choice T-maze assay with the heat source either in contact with the surface (i.e., conduction) or not in contact (i.e., radiation) in nonantennectomized bed bugs. Further, we systematically ablated the bed bug’s antennal segments (distal tip, first segment, and all four segments) and assessed their responses to heat and feeding in a unique two-choice T-maze assay and individual feeding assays, respectively. Our two-choice assays with contact to or no contact with the surface indicated that bed bugs cannot detect radiant heat. Later, we found that the distal tip of the terminal antennal segment is responsible for orientation toward a heat source. However, >50% of the bed bugs fed even when the entire antenna was removed, suggesting redundancy in sensory cues that drive feeding. These results will be used to better understand the role heat plays in bed bug host attraction and design of traps. 

   more » « less