OpenMP Code Offloading: Splitting GPU Kernels, Pipelining Communication and Computation, and Selecting Better Grid Geometries

Inspired by the success of BitTorrent (BT) in the field of file sharing, in this paper, we present a peer-to-peer (P2P) video-on-demand (VoD) system based on asynchronous data transfer. The system is specified with centralized directory and unstructured overlay. For efficient supplier selection and data transfer among peers, we consider the rareness and the priority of video blocks as well as available network resource. Packet-level simulations show the benefits of cooperated asynchronous data transfer in terms of playback continuity, launch delay and server stress.

Regular blood pressure control has been advocated for early detection and treatment of patients with hypertension. We aimed to study whether long-term survival rate is related to socioeconomic circumstances and biological factors among middle-aged men with screened new hypertension. Of the 22,444 men who attended a general health examination in Malmo, Sweden during 1974--1984, in total, 2033 had a systolic or diastolic blood pressure of at least 160 or 100 mmHg and were without treatment for hypertension. Mean age was 45.8 (SD: 5.3) years. All-cause mortality was studied in relation to smoking, diabetes, hyperlipidaemia, being overweight, marital status, problematic alcohol drinking, and occupation over a mean follow-up of 18.6 (3.6) years. Screened hypertension (> or = 160/ > or = 100 mmHg) was significantly more common among men with low occupational level and among single men. Higher risk factor levels were associated with an adverse social background. Of the men with screened hypertension, 479 (24%) died during the follow-up, half of them from cardiovascular diseases. Compared with married men with normal blood pressure, married men with screened hypertension had a relative mortality risk of 1.5 (95% confidence interval = 1.3-1.7), whereas the relative risk for corresponding single men was 3.0 (2.6-3.4) adjusted for confounders. Among men with screened hypertension, marital status, systolic blood pressure at baseline, smoking, diabetes, hyperlipidaemia, and low occupation level (manual) remained significantly associated with total mortality in the multivariate analysis. Marital status contributes substantially to the survival prognosis for men with screened new hypertension. Systolic blood pressure at baseline and other known risk factors for atherosclerosis were also associated with increased mortality in this group.

The Weather Research and Forecast (WRF) Double Moment 5-class (WDM5) mixed ice microphysics scheme predicts the mixing ratio of hydrometeors and their number concentrations for warm rain species including clouds and rain. WDM5 can be computed in parallel in the horizontal domain using multi-core GPUs. In order to obtain a better GPU performance, we manually rewrite the original WDM5 Fortran module into a highly parallel CUDA C program. We explore the usage of coalesced memory access and asynchronous data transfer. Our GPU-based WDM5 module is scalable to run on multiple GPUs. By employing one NVIDIA Tesla K40 GPU, our GPU optimization effort on this scheme achieves a speedup of 252x with respect to its CPU counterpart Fortran code running on one CPU core of Intel Xeon E5-2603, whereas the speedup for one CPU socket (4 cores) with respect to one CPU core is only 4.2x. We can even boost the speedup of this scheme to 468x with respect to one CPU core when two NVIDIA Tesla K40 GPUs are applied.

This article describes a highly parallel event driven interface between a Pixel Array Detector (PAD) and its processing electronics. The method used was originally developed for the Field Programmable Gate Array (FPGA) X-ray Pixel Array detector to allow for real-time processing of X-ray image data on its processing FPGA. This interface potentially allows for entirely asynchronous data transfer off the detector at rates exceeding 110Gbps and operates without the need for a constantly running clock. It thus presents a potential a low-power, low-noise interface option for 10-bit data transfer from photon-counting detectors at frame rates greater than 1MHz with no input dead time. This interface has been successfully tested on 64 independent 16-pixel channels at a transfer frequency of 294MHz achieving maximum data transfer rates of 16.7 Gbps.

In this paper we introduce a new real-time parallel image processing system based on an IEEE1394-based PC-cluster. PC-clusters are becoming standard tools in the field of computer vision, especially for real-time multi-view image processing. To reduce the cost of system construction, we have employed IEEE1394 bus as a base network of the PC-cluster. IEEE1394 bus has the throughput of up to 400Mbps, which is large enough to transfer uncompressed video data, and supports not only asynchronous data transfer but isochronous one, which is inevitable for real-time data transfer. In addition, since we have several products of IEEE1394-based digital cameras, we can easily integrate real-time image processing system including camera systems. This paper presents an overview of our real-time parallel image processing system on an IEEE1394based PC-cluster, referring to basic features of IEEE1394. We also show some experimental results to evaluate the performance of the PC-cluster, comparing it with the performance of our previous Myrinet-based PC-cluster. Then we show a prototypical application of the system, which is real-time volume reconstruction from multi-view images by visual cone intersection method.

HW/SW Co-designed systems rely on dynamic binary translation and optimizations for efficient execution of binary code. Due to memory ordering properties and other architectural constraints, most binary optimizations are applied to regions of code that are atomically executed. To ensure that the underlying hardware has enough speculative resources to execute the whole atomic region, these systems typically form short atomic regions, with only 20 to 30 instructions. However, the shorter is the atomic region the smaller is the scope for optimizations. We present LAR-CC, a novel technique that enables HW/SW co-designed systems to optimize large atomic regions and dynamically fit them into the available speculative hardware resources by means of conditional commits. The LAR-CC technique consists of two major components: 1) conditional branch instructions to conditionally skip commit operations; 2) code transformations that replace commit operations by conditional commits and enable optimizations to be applied on the large atomic regions. Our experiments show that LAR-CC can effectively achieve dynamic atomic region sizes larger than 1000 instructions, providing sufficiently large scope to apply many advanced optimizations on HW/SW co-designed systems.

The output data produced by numerical climate model simulations have increased greatly in complexity and size. The exploding volume of climate data is becoming a challenge for climate scientists. Our previous work, Climate Fast Input/Output (CFIO) library, implemented a two-phase I/O method to overlap I/O with computations, and achieved high throughput. In this paper, we present CFIO2, which can overlap communications with computations using Remote Direct Memory Access(RDMA) technology. We design a simple communication interactions model to implement asynchronous and concurrent data transfer. The experimental results show that CFIO2 can provide higher throughput than CFIO and can shorten the overall simulation time for climate model significantly.

The security of hardware-software systems is at risk from a wide range of attack vectors that appear at various stages during the execution of machine code. The existing approaches for repairing software defects have numerous restrictions with respect to their applicability and functionality and to the range of vulnerabilities that can be identified and eliminated.We propose an approach for removing software errors in program code that is based on just-in-time compilation in a virtual execution environment. The virtual environment uses static, dynamic, and hybrid analyses of the intermediate representation of vulnerable code and re-compiles such code to be safe.The language of code annotations allows us to manage static and dynamic analyses and code transformations. We can change the level of analysis and the amount of time spent on such analyses by dynamically adapting the precision.

User interface prototypes are effective structure for describing as models for web application development thanks to their graphical and dynamic architecture. It provides the benefits such as language independent development, less error-prone and increased quality. The executable programming codes are created by code transformation feature of model-driven development tools as well. Nevertheless, this technique brings the problems along all its benefits. First of all, code transformation from user interfaces by development tools, creates complicated structure to analyze and modify. Another challenge is to deal with the processes which are triggering each other. In addition, data connected process adaptation is a problem as well. All these issues eventually require the integration of discrete modules. This study provides not only a proposal structure to deal with integration problem in web development, but also provides the evaluation criteria for classification of the integration processes.

TCR down-modulation following binding to MHC/peptide complexes is considered to be instrumental for T cell activation because it allows serial triggering of receptors and the desensitization of stimulated cells. We studied CD3/TCR down-modulation and zeta degradation in T cells from two ZAP-70-immunodeficient patients. We show that, at high occupancy of the TCR, down-modulation of the CD3/TCR is comparable whether T cells express or do not express ZAP-70. However, if TCR occupancy was low, we found that CD3/TCR was down-regulated to a lesser extent in ZAP-70-negative than in ZAP-70-positive T cells. We studied CD3/TCR down-modulation in P116 (a ZAP-70-negative Jurkat cell-derived clone) and in P116 transfected with genes encoding the wild-type or a kinase-dead form of ZAP-70. Down-modulation of the TCR at high occupancy did not require ZAP-70, whereas at low TCR occupancy down-modulation was markedly reduced in the absence of ZAP-70 and in cells expressing a dead kinase mutant of ZAP-70. Thus, the presence of ZAP-70 alone is not sufficient for down-modulation; the kinase activity of this molecule is also required. The degradation of zeta induced by TCR triggering is also severely impaired in T cells from ZAP-70-deficient patients, P116 cells, and P116 cells expressing a kinase-dead form of ZAP-70. This defect in TCR-induced zeta degradation is observed at low and high levels of TCR occupancy. Our results identify ZAP-70, a tyrosine kinase known to be crucial for T cell activation, as a key player in TCR down-modulation and zeta degradation.

This chapter discusses several speculations based on the current experiences of minorities in three closely related areas—health, work, and retirement. Minority-nonminority group comparisons are often based on self-reports of health from subjects in national surveys. The issue is whether these reports mean the same, or are measured the same, for minorities as for nonminorities. Work patterns, and factors determining those work patterns, were found to be different for African American and white men and women. Retirement patterns, and the factors that determine them, appear to be different for whites, African Americans, and Mexican Americans. African and Puerto Rican Americans often have been at the bottom when minorities were ranked on measure of health, work, and retirement. Special attention needs to be devoted to these especially disadvantaged groups—and to Native Americans—in future research, planning, programs, and services. Lower-level occupations and wages and more sporadic work patterns over a lifetime are leading causes of minorities’ low levels of retirement income.

Biomonitoring of the adverse effects induced by the chronic exposure to lead and cadmium on kidney function: Usefulness of alpha-glutathione S-transferase

Lu AA21004, a novel multimodal antidepressant, produces regionally selective increases of multiple neurotransmitters—A rat microdialysis and electrophysiology study

There is an increased interest in measuring the interaction of new or established drugs with their targets, in order to gain a better understanding of their mechanisms of action. PET can provide this information if an appropriate radioligand is available. [18F]MPPF (4-(2'-methoxyphenyl)-1-[2'-(N-2"-pyridinyl)-p-[18F]fluorobenzamido]ethylpiperazine) is a selective radioligand for serotonin 5-HT1A receptors. We have established that the binding potential (BP=Bmax/KD) of [18F]MPPF for cerebral 5-HT1A receptors can be assessed in human brain without arterial sampling. The aim of this study was to assess if 5-HT1A receptor occupancy can be measured through calculation of a drug-related decrease in BP with [18F]MPPF and PET. Six volunteers were scanned twice using a Siemens Exact HR+ camera following injection of 70+/-18 MBq [18F]MPPF (baseline and medicated conditions). Before the second scan, volunteers orally received either 3x10 mg pindolol at T=-15.5 h, T=-6.5 h, and T=-1.5 h (n=3) or 10 mg buspirone in a single dose at T=-1.5 h (n=3). Binding potentials were calculated using the simplified reference tissue model with the cerebellum as reference. Administration of 30 mg pindolol led to a significant reduction in [18F]MPPF binding potential of 42+/-17%. In contrast, no significant reduction of [18F]MPPF binding potential was observed following administration of buspirone (5+/-17%). These results show that [18F]MPPF can be used for measurement of drug-related 5-HT1A receptor occupancy and may be of particular interest in determining the 5-HT1A receptor interaction of new or established drugs in phase 1 and early phase 2 drug trials. Apparently, the 5-HT1A partial agonist buspirone is already clinically effective at low levels of 5-HT1A receptor occupancy.

SummaryManycore accelerators have the potential to significantly improve performance of scientific applications when offloading computationally intensive program portions to accelerators. Directive‐based high‐level programming models, such as OpenACC and OpenMP, are used to create applications for accelerators through annotating regions of code meant for offloading. OpenACC is an emerging directive‐based programming model for programming accelerators that typically enable inexperienced programmers to achieve portable and productive performance within applications. In this paper, we present our research in developing challenges and solutions when creating an open‐source OpenACC compiler in an industrial framework (OpenUH as a branch of Open64). We then discuss in detail techniques we developed for loop scheduling reduction operations on general purpose GPUs. The compiler is evaluated with benchmarks from the NAS Parallel Benchmarks suite and self‐written micro‐benchmarks for reduction operations. This implementation has been designed to serve as a compiler infrastructure for researchers to explore advanced compiler techniques, extend OpenACC to other programming models, and build performance tools used in conjunction with OpenACC programs. Copyright © 2015 John Wiley & Sons, Ltd.