Translating OpenACC to LLVM IR with SPIR kernels Hao-Wei Peng; Jean Jyh-Jiun Shann
2016 IEEE/ACIS 15th International Conference on Computer and Information Science (ICIS),
06/2016
Conference Proceeding
Open access
In general, highly parallelized programs executed on heterogeneous multiprocessor platforms may get better performance than homogeneous ones. OpenCL is one of the standards for parallel programming ...of heterogeneous multiprocessor platforms and SPIR (Standard Portable Intermediate Representation) is a portable binary format for representing OpenCL kernel code. However, the programming of these programs is usually complex and error-prone for most programmers. Therefore, some standards have been proposed to simplify the programming on heterogeneous multiprocessor platforms, for example, OpenACC (a directive-based parallel programming model). In this paper, we implement our framework on Clang, the C front-end of LLVM, to automatically translate OpenACC to LLVM IR with SPIR kernels. After that, it is optional to optimize the IR code by LLVM optimizer and execute the host LLVM IR by LLVM JIT-compiler. According to the experiment results, our translated programs have significant performance enhancement for some programs while comparing with their corresponding sequential version of programs and have comparable performance while comparing with their manual OpenCL version. Therefore, our design may reduce the difficulty of writing the programs in heterogeneous multiprocessor platform and the translated OpenCL programs are portable and have good performance as that of the manual OpenCL programs written by experienced programmers.
碩士
國立臺北科技大學
機械工程系機電整合碩士班
106
In this study, we aimed to develop a MI-based multi-functional brain-computer interface communication system for patients with ALS. The MI-BCI syetem included three ...functions, Yes/No immediate-reply function, Patients’ needs selection function and communication board-spelling function. We extract fractal dimension as the feature in the MI-BCI system. Fisher criterion-based channel selection strategy is proposed to automatically determine the best patient-dependent channel configuration from 30 EEG recording sites. The average accuracy from 12 patients by using the Top 5-Channel of each patient for resting vs MI classification can achieve 89%. Moreover, in four-class MI classification, using 30-Channel of each patient can achieve 82.67%. Those results can prove a efficiency and feasibility way for further MI-based BCI applications.
This paper proposes a saturation adjustment method based on human vision with YCbCr color model characteristics and luminance changes. In the traditional color adjustment approach, people tried to ...separately adjust the luminance and saturation. However, this approach makes the color over-saturate very easily and makes the image look unnatural. In this work, we try to use the concept of exposure compensation to simulate the brightness changes and to find the relationship among luminance, saturation, and hue. The simulation indicates that saturation changes with the change of luminance and the simulation also shows there are certain relationships between color variation model and YCbCr color model. Together with all these symptoms, we also include the human vision characteristics to propose a new saturation method to enhance the vision effect of an image. The experimental results show that the proposed approach can make the image have better vivid and contrast. Most important of all, unlike the over-saturation caused by the conventional approach, our approach prevents over-saturation and further makes the adjusted image look natural.
Abstract The aqueous reprocessing of spent nuclear fuel generates a considerable amount of plutonium alongside contaminated stainless steel, necessitating meticulous handling for safe decommission ...and long‐term management. The present work investigated the co‐immobilization of CeO 2 (as an inactive surrogate for PuO 2 ) with metallic Fe and Cr (representing a simplified stainless steel) within a zirconolite ceramic wasteform, nominally targeting Ca 1− x Ce x ZrTi 2−2 x Fe x Cr x O 7 . After sintering at 1400°C under an air atmosphere, the zirconolite phase constituted 90.8–95.1 wt.% of the product across the solid solution range of 0.05 ≤ x ≤ 0.25, alongside perovskite and baddeleyite secondary phases in varying proportion; no evidence of the unincorporated CeO 2 or metallic and oxidized Fe or Cr were identified. Above x = 0.30 CeO 2 was detected inferring, the solubility limit was reached. A polytype transformation from zirconolite 2M to 3T was confirmed by X ‐ray diffraction and selected area electron diffraction results, with the relative fraction of the 3T phase gradually increasing to 49.5 wt.% at x = 0.30. Deconvolution of X‐ray photoelectron spectroscopy data revealed the partial reduction of Ce 4+ –Ce 3+ , whereas Fe and Cr species maintained trivalent, in agreement with the targeted substitution scheme. Benefitting from the excellent chemical flexibility of zirconolite structured compounds, the co‐immobilization approach may be an effective disposal pathway for Pu‐containing wastes and contaminated stainless steel residues.
Full text
Available for:
BFBNIB, FZAB, GIS, IJS, KILJ, NLZOH, NUK, OILJ, SBCE, SBMB, UL, UM, UPUK
Objective: To assess the effects of axial vibrations on gene expression and lumbar intervertebral disc degeneration in vivo. Methods: A modified bipedal rat model was established using a brachial ...plexus rhizotomy approach to imitate human upright posture. The experimental animals were randomly divided into three groups: control, vertical vibration, and whole-body vibration. Gene expression in degeneration of the intervertebral discs was assessed by reverse transcription-quantitative polymerase chain reaction. Results: The expression of aggrecan, Col1α1, Col2α1, and decorin were shown to be up-regulated in 14-week-old rats in the vertical vibration and whole-body vibration groups, whereas biglycan and versican expression was down-regulated in 14-week-old rats of the two experimental groups. Furthermore, biglycan and versican expression levels were shown to be lower in the whole-body vibration group than in the vertical vibration group(P<0.05). Conclusions: This in-vivo study demonstrated that vibrations can influence the expression of anabolic genes. Furthermore, whole-body vibrations seem to have a greater effect in this regard than vertical vibrations. A new method is expected to relieve the low back pain of the patients through our research.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
In this paper we investigate the methods for image keypoint detection, description, and matching. We compare four methods, including SIFT, SURF, ORB, and BRISK. The datasets used for evaluation cover ...different changes and transformations in imaging conditions, such as zoom, rotation, blur, viewpoint, light, and JPEG compression. We also compare the computational speed of these four methods.
The lamina cribrosa is affected by intraocular pressure, which is the major risk of glaucoma. However, the capability to evaluate the lamina cribrosa in vivo has been limited until recently due to ...poor image quality and the posterior laminar displacement of glaucomatous eyes. In this study, we propose an automatic method to measure the anterior lamina cribrosa surface depth (ALCSD), including a method for detecting Bruch's membrane opening (BMO) based on k-means and region-based active contour. An anterior lamina cribrosa surface segmentation method based on energy constraint is also proposed. In BMO detection, we initialize the Chan-Vese active contour model by using the segmentation map of the k-means cluster. In the segmentation of anterior lamina cribrosa surface, we utilize the energy function in each A-scan to establish a set of candidates. The points in the set that fail to meet the constraints are removed. Finally, we use the B-spline fitting method to obtain the results. The proposed automatic method can model the posterior laminar displacement by measuring the ALCSD. This method achieves a mean error of 45.34 μm in BMO detection. The mean errors of the anterior lamina cribrosa surface are 94.1% within five pixels and 76.1% within three pixels.