Search for: All records

In recent years, the MLIR framework has had explosive growth due to the need for extensible deep learning compilers for hardware accelerators. Such examples include Triton [39], CIRCT [14], and ONNX-MLIR [22]. MLIR compilers introduce significant complexities in localizing bugs or inefficiencies because of their layered optimization and transformation process with compilation passes. While existing delta debugging techniques can be used to identify a minimum subset of IR code that reproduces a given bug symptom, their naive application to MLIR is time-consuming because real-world MLIR compilers usually involve a large number of compilation passes. Compiler developers must identify a minimized set of relevant compilation passes to reduce the footprint of MLIR compiler code to be inspected for a bug fix. We propose DuoReduce, a dual- dimensional reduction approach for MLIR bug localization. DuoReduce leverages three key ideas in tandem to design an efficient MLIR delta debugger. First, DuoReduce reduces compiler passes that are irrelevant to the bug by identifying ordering dependencies among the different compilation passes. Second, DuoReduce uses MLIR-semantics-aware transformations to expedite IR code reduction. Finally, DuoReduce leverages cross-dependence between the IR code dimension and the compilation pass dimension by accounting for which IR code segments are related to which compilation passes to reduce unused passes. Experiments with three large-scale MLIR compiler projects find that DuoReduce outperforms syntax-aware reducers such as Perses and Vulcan in terms of IR code reduction by 31.6% and 21.5% respectively. If one uses these reducers by enumerating all possible compilation passes (on average 18 passes), it could take up to 145 hours. By identifying ordering dependencies among compilation passes, DuoReduce reduces this time to 9.5 minutes. By identifying which compilation passes are unused for compiling reduced IR code, DuoReduce reduces the number of passes by 14.6%. This translates to not needing to examine 281 lines of MLIR compiler code on average to fix the bugs. DuoReduce has the potential to significantly reduce debugging effort in MLIR compilers, which serves as the foundation for the current landscape of machine learning and hardware accelerators. 

In recent years, the MLIR framework has had explosive growth due to the need for extensible deep learning compilers for hardware accelerators. Such examples include Triton, CIRCT, and ONNX-MLIR. MLIR compilers introduce significant complexities in localizing bugs or inefficiencies because of their layered optimization and transformation process with compilation passes. While existing delta debugging techniques can be used to identify a minimum subset of IR code that reproduces a given bug symptom, their naive application to MLIR is time-consuming because real-world MLIR compilers usually involve a large number of compilation passes. Compiler developers must identify a minimized set of relevant compilation passes to reduce the footprint of MLIR compiler code to be inspected for a bug fix. We propose DuoReduce, a dual-dimensional reduction approach for MLIR bug localization. DuoReduce leverages three key ideas in tandem to design an efficient MLIR delta debugger. First, DuoReduce reduces compiler passes that are irrelevant to the bug by identifying ordering dependencies among the different compilation passes. Second, DuoReduce uses MLIR-semantics-aware transformations to expedite IR code reduction. Finally, DuoReduce leverages cross-dependence between the IR code dimension and the compilation pass dimension by accounting for which IR code segments are related to which compilation passes to reduce unused passes. Experiments with three large-scale MLIR compiler projects find that DuoReduce outperforms syntax-aware reducers such as Perses and Vulcan in terms of IR code reduction by 31.6% and 21.5% respectively. If one uses these reducers by enumerating all possible compilation passes (on average 18 passes), it could take up to 145 hours. By identifying ordering dependencies among compilation passes, DuoReduce reduces this time to 9.5 minutes. By identifying which compilation passes are unused for compiling reduced IR code, DuoReduce reduces the number of passes by 14.6%. This translates to not needing to examine 281 lines of MLIR compiler code on average to fix the bugs. DuoReduce has the potential to significantly reduce debugging effort in MLIR compilers, which serves as the foundation for the current landscape of machine learning and hardware accelerators. 

Coherent phonons in the Terahertz (THz) regime have gained attention as potential candidates for next-generation high-speed, low-energy information carriers in atomically thin phononic or phonon-integrated on-chip devices. Nevertheless, achieving efficient control of the phonon generation dynamics over THz coherent phonons continues to pose a considerable challenge. In this work, we explore THz coherent phonon generation in exfoliated van der Waals (vdW) flakes of WSe2 on Au (WSe2/Au) and Si (WSe2/Si) by using time-resolved pump–probe spectroscopy. The generation of THz coherent phonons was studied as a function of the WSe2 layer thickness and laser wavelength. Notably, a significant enhancement in THz coherent phonon generation was observed in the WSe2/Au structure, but only within a specific range of WSe2 thicknesses and laser wavelengths. The results from numerical simulations, which consider a self-hybridized optical cavity depending on WSe2 thickness and optical reflectance and Raman spectroscopy measurements, all align well with the time-domain observations of THz coherent phonon generation. We propose that the observed enhancement in THz coherent phonon generation is strongly influenced by light–matter interactions in the WSe2 cavity, a mechanism that may be applicable to a broader range of vdW materials. These findings offer promising insights for the development of THz phononic or phonon-integrated devices. 

Mounting concerns regarding per‐/poly‐fluoroalkyl substances (PFAS) on human health are focusing attention on trace‐level PFAS detection in aqueous environments. Here, we report a readily prepared small molecule, 2,6‐bis(3,5‐diethyl‐1H‐pyrrol‐2‐yl)pyridine (receptor 1), that displays high binding affinities (logKa< = 4.9–6.2) and produces a strong “turn‐on” emission response when exposed to representative PFAS in hexanes. The hydrophobic nature of 1 , and its strong affinity for various PFAS, allowed hexanes solutions of 1 to be used as “turn‐on” emission sensors for dilute aqueous solutions of long‐chain (≥C8) PFAS under acidic conditions (pH 2) by liquid‐phase extraction (LPE). In the case of perfluorooctanoic acid (PFOA), the response was rapid (under 10 min) and sensitive. Limits of detection (LOD) as low as 250 ppt were readily achievable by direct naked‐eye observation. LOD as low as 40 and 100 ppt, respectively, could be reached for deionized and tap water solutions of PFOA using a smartphone color‐scanning application. Little change in the sensitivity was seen in the presence of a range of inorganic and organic species that could act as potential interferants. Support for the present findings came from UV–vis absorbance, fluorescence, 1 

IntroductionThe primary objective of this study was to identify variables that significantly influence the implementation of math Response to Intervention (RTI) at the school level, utilizing the ECLS-K: 2011 dataset. MethodsDue to missing values in the original dataset, a Random Forest algorithm was employed for data imputation, generating a total of 10 imputed datasets. Elastic net logistic regression, combined with nested cross-validation, was applied to each imputed dataset, potentially resulting in 10 models with different variables. Variables for the models derived from the imputed datasets were selected using four methods, leading to four candidate models for final selection. These models were assessed based on their performance of prediction accuracy, culminating in the selection of the final model that outperformed the others. Results and discussionMethod₅₀and Method_coefemerged as the most effective, achieving a balanced accuracy of 0.852. The ultimate model selected relevant variables that effectively predicted RTI. The predictive accuracy of the final model was also demonstrated by the receiver operating characteristic (ROC) plot and the corresponding area under the curve (AUC) value, indicating its ability to accurately forecast math RTI implementation in schools for the following year.