Search

For system manufacturers, packaging of the ball grid array (BGA) of wireless communication modules can be performed via solder paste printing. As reballing can be conducted via the pretinning process, it can be compatible with its existing surface mount technology (SMT). This study referred to the Taguchi experimental design to plan its experiment. Three quality characteristics related to the consistency of the solder volume are applied for printing, that is, the standard deviation of the deposition volume of solder paste on all bonding pads of a single component, the coplanarity of the solder balls of the component, and the height range of the solder ball of a single component, which were regarded as the response values. Important design and process parameters related to the pretinning process and the relative location of the bonding pads of components on a connecting board are considered as the controlling factors. Then, this study applied and integrated the results of two multivariate analysis methods, i.e., gray correlation of the principal components and a technique for order preference by similarity to ideal solution (TOPSIS), in order to determine the ideal solution and propose the optimal parameter combination: general squeegee: pressure 40 N and speed 30 mm/s; steel: separation distance 3 mm, speed 2 mm/s, and aspect ratio 0.17; and cleaning frequency: cleaning of stencils after the printing of every two boards. This method effectively improves the control of the solder volume in the pretinning process and meets the acceptance criteria of manufacturers.

We retrospectively analyzed 82 adult patients with MALA admitted to the ICU over 20 years. The association between the clinical parameters and mortality post-MALA was estimated using logistic regression analysis.Patients with MALA admitted to the ICU presented with clinical symptoms mainly associated with the head (40.24%), chest (41.46%), and abdomen (35.37%). Additionally, the PLL distribution significantly varied with age, APACHE II = Acute Physiology and Chronic Health Evaluation II (APACHE II) score, various laboratory parameters like nadir arterial bicarbonate level, multiple treatment modalities such as renal replacement therapy, and mortality. The overall mortality rate was 17.07%. After adjustment of age and gender, the significant predictors of mortality were APACHE II score, PLL, vasoactive support, ventilator support, and cardiopulmonary resuscitation.Despite MALA being a rare event, it is necessary to evaluate its clinical characteristics, especially the associated PLL and mortality. In the current study, higher levels of APACHE II score and PLL show a greater likelihood of mortality in MALA patients.

TVM is a AI compiler supports both graph-level and operator-level optimization on machine learning models. It provides an optimizable flow to deploy on diverse target devices. By exploiting TVM schedule, we can optimize the codegen behavior for our RISC-V architecture. Since RISCV is configurable with the selection on different extension, we can enable multiple extensions depends on the application scene. In our work, we present the flow for enabling and optimizing the RISC-V P extension toward QNN models from TVM. With support from LLVM and a customized deep learning runtime (DLR), we verified our work on both FLOAT32 and prequantized models from Tensorflow Lite. Experiments shows that comparing with FLOAT32 models, our work can achieve 2.7-7.0 times of performance improvement with regard to total instruction count at runtime for pre-quantized version with a set of benchmarks including Mobilenet and Inception-v3. As for accuracy issue, the degradation is tiny for quantization version among 500 images. All experiments are running on RISC-V simulator, Spike with P extension support.

Heterogeneous systems that consist of multiple CPUs and GPUs for high-performance computing are becoming increasingly popular, and OpenCL (Open Computing Language) provides a framework for writing programs that can be executed across heterogeneous devices. Compared with OpenCL 1.2, the new features of OpenCL 2.0 provide developers with better expressive power for programming heterogeneous computing environments. Currently, gem5-gpu, which includes gem5 and GPGPU-Sim, can offer an experimental simulation environment for OpenCL. In gem5-gpu, gem5 only supports CUDA, although GPGPU-Sim can support OpenCL by compiling an OpenCL kernel code to PTX code using real GPU drivers. However, this compilation flow in GPGPU-Sim can only support up to OpenCL 1.2. OpenCL 2.0 provides new features such as workgroup built-in functions, extended atomic built-in functions, and device-side enqueue. To support OpenCL 2.0, the compiler must be extended to enable the compilation of OpenCL 2.0 kernel code to PTX code. In this paper, the proposed compiler is modified from the low level virtual machine (LLVM) compiler to extend such features to enhance the emulator to support OpenCL 2.0. The proposed compiler creates local buffers for each workgroup to enable workgroup built-in functions and adds atomic built-in functions with memory order and memory scope for OpenCL 2.0 in NVPTX. Furthermore, the APIs available in CUDA are utilized to implement the OpenCL 2.0 device-side enqueue kernel and compilation schemes in Clang are revised. The AMD APP SDK 3.0 and NTU OpenCL benchmarks are used to verify that the proposed compiler can support the features of OpenCL 2.0.

Invasive pulmonary Aspergillus infection during or after severe influenza infection in an immunocompromised patient or a previously healthy person has been reported in the literature [1-2]. In addition, coinfections of Staphylococcus aureus or Streptococcus pneumoniae with influenza are common, whereas simultaneous infections of legionellosis with aspergillosis are unusual in patients with influenza [2]. We herein report a diabetic patient who presented with a community-acquired pneumonia (CAP). The initial sputum culture yielded Aspergillus species and S. aureus, which were coinfected with influenza A, Pneumococcus and Legionella, with a deteriorated fatal course. A 63-year-old man with diabetes mellitus felt general malaise and cough with sputum for 3 days. He was admitted to the intensive care unit (ICU) of the hospital on June 20, 2015 due to pneumonia with respiratory failure. There was no fever, diarrhea, headache, or skin rash. He denied travel and contact history. On examination, a blood pressure of 80/60 mmHg was noticed. The breath sounds revealed bilateral fine rales. Laboratory data included WBC 15,700/μL; band, 18%; CRP, 91.9 mg/L; BUN, 50 mg/dL; creatinine, 2.38 mg/dL; blood sugar, 221 mg/dL and HbA1C, 9.1%. The initial chest x-ray (CXR) film showed mild infiltration in perihilar area of left lung on June 20 (Figures 1A). The result of urinary Legionella Ag test was negative. Empirical antibiotic therapy included piperacillin-tazobactam and levofloxacin.

Previous studies have revealed higher mortality rates in patients with severe influenza who are coinfected with invasive pulmonary aspergillosis (IPA) than in those without IPA coinfection, nonetheless, the clinical impact of IPA on economic burden and risk factors for mortality in critically ill influenza patients remains undefined. The study was retrospectively conducted in three institutes. From 2016 through 2018, all adult patients with severe influenza admitted to an intensive care unit (ICU) were identified. All patients were classified as group 1, patients with concomitant severe influenza and IPA, group 2, severe influenza patients without IPA, and group 3, severe influenza patients without testing for IPA. Overall, there were 201 patients enrolled, including group 1 (n = 40), group 2 (n = 50), and group 3 (n = 111). Group 1 patients had a significantly higher mortality rate (20/40, 50%) than that of group 2 (6/50, 12%) and group 3 (18/11, 16.2%), p &lt, 0.001. The risk factors for IPA occurrence were solid cancer and prolonged corticosteroid use in ICU of &gt, 5 days. Group 1 patients had significantly longer hospital stay and higher medical expenditure than the other two groups. The risk factors for mortality in group 1 patients included patients’ Charlson comorbidity index, presenting APACHE II score, and complication of severe acute respiratory distress syndrome. Overall, IPA has a significant adverse impact on the outcome and economic burden of severe influenza patients, who should be promptly managed based on risk host factors for IPA occurrence and mortality risk factors for coinfection with both diseases.

Aspergillus-associated infection might comprise up to 23–29% of severe influenza patients from the community throughout stay in an intensive care unit (ICU). In Taiwan, cases of severe influenza with aspergillosis are increasingly reported. Therefore, we describe the relative risk of mortality among severe influenza patients with aspergillosis and other coinfections compared to severe influenza patients without Aspergillus coinfections. Methods: We retrospectively reviewed 124 adult patients with severe influenza in a tertiary medical center in southern Taiwan from January 2015 through March 2016. The definition of probable aspergillosis required abnormal radiological findings and positive Aspergillus galactomannan (GM) antigen and/or Aspergillus isolation. Results: Probable aspergillosis (detected throughout the whole course) and other coinfections (only community-acquired) were diagnosed in 21 (17%) and 38 (31%) of all patients respectively. Klebsiella pneumoniae (36.8%), Pseudomonas aeruginosa (31.6%) and Staphylococcus aureus (31.6%) were the most frequent isolates of other coinfections. In-ICU mortality of Aspergillus group (66.7%) was significantly higher than other coinfections (23.7%, p = 0.001) or control group without coinfections (15.4%, p

With the growing computing complexity of 5G baseband, parallel computing and OpenCL become a promising solution to help with the computing needs for baseband computing. In this paper, the issues are investigated in supporting the complex class library with OpenCL C++ class for baseband computing. Using templates and overloaded C++ features, the proposed complex class library works with many data types. This work also addresses the issue with fast math, and can work with standard C++ math with complex number as well as fast-math option in the complex class that the check of Not a Number (NaN) and infimum (INF) are removed. Our work is experimented with two environments. One is with Intel OpenCL 2.1 environment with complex float and complex half, and the other one runs a minimum mean-square error (MMSE) equalizer application in Massive Multi-input Multi-output (MIMO). As compared to the C++ sequential code, the complex class in OpenCL can reduce total execution time. Based on the preliminary results, the kernel execution time could be reduced by about 5% to 45% with fast math to omit NaN and INF check. The C++ class header file will also be illustrated to fit C++ complex class with OpenCL C++ linguistics.

This paper investigates the phenomenon of mode switching for GaAs-based high index contrast photonic crystal surface emitting lasers (PCSELs) with deeply etched air holes. The room-temperature lasing characteristics and angular resolved photoluminescence spectra of PCSELs show that the lasing mode would change from the fundamental guided mode to a high-order guided mode as the filling factor (FF) increased, resulting from the diffraction of deeply etched semiconductor/air arrays with high index contrast. Higher order guided modes could coherently couple out of photonic crystal and would significantly influence the laser output of PCSELs with a larger air FF, leading to rich far-field characteristics.

In recent decades, the global incidence of dengue has increased. Affected countries have responded with more effective surveillance strategies to detect outbreaks early, monitor the trends, and implement prevention and control measures. We have applied newly developed machine learning approaches to identify laboratory-confirmed dengue cases from 4,894 emergency department patients with dengue-like illness (DLI) who received laboratory tests. Among them, 60.11% (2942 cases) were confirmed to have dengue. Using just four input variables [age, body temperature, white blood cells counts (WBCs) and platelets], not only the state-of-the-art deep neural network (DNN) prediction models but also the conventional decision tree (DT) and logistic regression (LR) models delivered performances with receiver operating characteristic (ROC) curves areas under curves (AUCs) of the ranging from 83.75% to 85.87% [for DT, DNN and LR: 84.60% ± 0.03%, 85.87% ± 0.54%, 83.75% ± 0.17%, respectively]. Subgroup analyses found all the models were very sensitive particularly in the pre-epidemic period. Pre-peak sensitivities (, Author summary Identifying dengue cases early is crucial but challenging for healthcare professionals. This challenge is increased during large epidemics and is a particular problem in non-endemic areas with limited experienced staff. To improve dengue diagnosis, we investigated how to exploit machine learning (ML)-based prediction models and identified four key variables [age, fever, white blood cell counts (WBCs), and platelet counts], which are compatible with clinical and epidemiological knowledge. With these variables, the ML prediction models [decision tree (DT), deep neural network (DNN)] and the logistic regression model developed for identifying laboratory-confirmed dengue cases produced areas under curve (AUCs) of the receiver operating characteristic (ROC) curves ranging from 83.75% to 85.87%. This implies that the prediction models may serve as a pivotal component of an integrated dengue surveillance system and they required only a single complete blood count (CBC) examination. The sensitivities, positive prediction values, and accuracies for major risk factors in the two machine learning models were close to those of the regression models. For future applications, the DNN models with superior performance can be employed at epidemic sites with adequate computer facilities, while the DT and regression models with interpretable prediction logic can be employed at sites with limited or no computer facilities. Artificial intelligence and clinical parameters identified from this study may aid when laboratories are overwhelmed, but should never replace laboratory confirmation.

In recent years, the parallel computing performance on GPUs (graphics processing units) has grown rapidly. As a result, GPUs have been widely applied in computationally intensive applications such as image processing, deep learning, and artificial intelligence. Because these applications can be modeled by multiple GPU kernels, some of which might even be dependent, it is essential to identify an efficient method for scheduling dependent kernels on GPU cores. Simply observing kernel dependencies by executing them in sequence results in performance degradation. Furthermore, dependent kernels generally need to share data. Consequently, without properly scheduling dependent kernels, unnecessary memory accesses and copies will be generated. Neural network model environments include many operators that are suitable for both parallel and dependent kernels. This paper proposes an efficient and clear method for creating a framework to analyze the ONNX (open neural network exchange) model and find the pattern of dependent kernels, which can then be used for scheduling with the GPU architecture. The preliminary experimental results show that this technique improves the overall performance by 8% and reduces the cache miss rate by 14% on average by combining neural network operators and appropriate memory policies.

Item does not contain fulltext 01 augustus 2019

GPUs have now been widely used in various computation-intensive applications, such as image processing, deep learning, artificial intelligence, etc. As these applications could be modeled by multiple GPU kernels, some of which might even be dependent, it is essential to find an efficient method to schedule dependent kernels on GPU cores. Simply observing dependences of kernels by executing them in sequence will result in performance degradation. Furthermore, dependent kernels generally need to share data. Consequently, without properly scheduling dependent kernels, unnecessary memory accesses and copies will be generated. This paper proposes an efficient method for scheduling dependent kernels on GPUs. Preliminary experimental results show that this technique improves performance by 43% on average when combining with appropriate memory write-back policies.

We have explored a novel molecular framework with a lysine core to prepare a probe library for protein kinases. Active site-directed probes, identified through screening of the library, displayed desirable labelling properties, including target specificity, good linear response to the corresponding enzymatic activity and cell permeability. Finally, a two-stage protocol utilizing the newly developed probe with an immunoprecipitation step was successfully established to monitor intracellular Src kinase activity in RK3E-v-Src cells.

A class of neutral tris-bidentate Ir(III) metal complexes incorporating a diphosphine as a chelate is prepared and characterized here for the first time. Treatment of [Ir(dppb)(tht)Cl3] (1) with fppzH afforded the dichloride complexes, trans-(Cl,Cl)[Ir(dppb)(fppz)Cl2] (2) and cis-(Cl,Cl)[Ir(dppb)(fppz)Cl2] (3). The reaction of 3 with the dianionic chelate precursor bipzH2 or mepzH2, in DMF gave the complex [Ir(dppb)(fppz)(bipz)] (4) or [Ir(dppb)(fppz)(mepz)] (5), respectively. In contrast, a hydride complex [Ir(dppb)(fppz)(bipzH)H] (6) was isolated instead of 4 in protic solvent, namely: DGME. All complexes 2 - 6 are luminescent in powder forms and thin films where the dichlorides (2, 3) emit with maxima at 590-627 nm (orange) and quantum yields (Q.Y.s) up to 90% whereas the tris-bidentate (4, 5) and hydride (6) complexes emit at 455-458 nm (blue) with Q.Y.s up to 70%. Hybrid TD-DFT calculations showed considerable MLCT contribution to the orange-emitting 2 and 3 but substantial ligand-centered 3ππ* transition character in the blue-emitting 4 - 6. The dppb does not participate to these radiative transitions in 4 - 6, but it provides the rigidity and steric bulk needed to promote the luminescence by suppressing the self-quenching in the solid state. Fabrication of an OLED with dopant 5 gave a deep blue CIE chromaticity of (0.16, 0.15). Superior blue emitters, which are vital in OLED applications, may be found in other neutral Ir(III) complexes containing phosphine chelates.

OpenCL continues to gather momentum on both desktop and mobile devices. The new features of OpenCL 2.0 provides developers better expressive power in programming heterogeneous computing environments. Currently in the experimental simulation environment, gem5-gpu only supports CUDA, but GPGPU-Sim can support OpenCL by compiling OpenCL kernel code to PTX using real GPU driver. However, this driver compilation in GPGPU-Sim only can support up to OpenCL 1.2. To support OpenCL 2.0, it is necessary to extend the compiler to enable the compilation of OpenCL 2.0 kernel code to PTX. In this paper, our experience in enabling the compiler flow is reported. In OpenCL 2.0, it provides new features such as dynamic parallelism, work-group built-in functions, extend atomic builtin functions, and so on. The proposed compiler that is modified from Low Level Virtual Machine (LLVM) extends such features for enhancing the emulator to support OpenCL 2.0. After the compiler is modified, it can support dynamic parallelism, workgroup built-in functions and extend atomic built-in functions. Using existing dynamic parallelism APIs in CUDA to implement OpenCL 2.0 enqueue kernel and revise compilation scheme in clang. Furthermore, the proposed compiler also creates local buffers for each work group to use for work-group builtin functions, and adds atomic built-in functions with memory order and memory scope for OpenCL 2.0 in NVPTX. From benchmarks, the proposed compiler can support the claim target.

Despite the fast boom of thermally activated delayed fluorescence (TADF) emitters bearing borane-based acceptor, so far, no TADF emitter with a direct B-N linkage between N-donor and boryl acceptor has been reported. The latter should simplify the molecular architecture and hence facilitate the synthetic design and versatility. We report here the preparation and characterization of a new series of N-borylated compounds with functional acridine donor unit; namely: ACBM, PACBM, and SACBM. Spectroscopic studies were performed to explore their photophysical properties that exhibited prominent solvatochromism and thermally activated delayed fluorescence. The time-dependent DFT calculation indicated the involvement of substantial intramolecular charge transfer character for which HOMO and LUMO are spatially separated. For compound SACBM, fabrication of green emitting OLED gave CIE chromaticity of (0.22, 0.59) and maximum external quantum efficiency, luminance efficiency and power efficiency of 19.1%, 60.9 cd/A, and 43.6 lm/W, respectively, demonstrating for the first time the highly efficient OLEDs using N-borylated TADF emitters.