TLB-pilot: Mitigating TLB Contention Attack on GPUs with Microarchitecture-Aware Scheduling

Co-running GPU kernels on a single GPU can provide high system throughput and improve hardware utilization, but this raises concerns on application security. We reveal that translation lookaside buffer (TLB) attack, one of the common attacks on CPU, can happen on GPU when multiple GPU kernels co-run. We investigate conditions or principles under which a TLB attack can take effect, including the awareness of GPU TLB microarchitecture, being lightweight, and bypassing existing software and hardware mechanisms. This TLB-based attack can be leveraged to conduct Denial-of-Service (or Degradation-of-Service) attacks. Furthermore, we propose a solution to mitigate TLB attacks. In particular, based on the microarchitecture properties of GPU, we introduce a software-based system, TLB-pilot, that binds thread blocks of different kernels to different groups of streaming multiprocessors by considering hardware isolation of last-level TLBs and the application’s resource requirement. TLB-pilot employs lightweight online profiling to collect kernel information before kernel launches. By coordinating software- and hardware-based scheduling and employing a kernel splitting scheme to reduce load imbalance, TLB-pilot effectively mitigates TLB attacks. The result shows that when under TLB attack, TLB-pilot mitigates the attack and provides on average 56.2% and 60.6% improvement in average normalized turnaround times and overall system throughput, respectively, compared to the traditional Multi-Process Service based co-running solution. When under TLB attack, TLB-pilot also provides up to 47.3% and 64.3% improvement (41% and 42.9% on average) in average normalized turnaround times and overall system throughput, respectively, compared to a state-of-the-art co-running solution for efficiently scheduling of thread blocks.

Perspective—Insights into Solar-Rechargeable Redox Flow Cell Design: A Practical Perspective for Lab-Scale Experiments

Solar-rechargeable redox flow batteries (SRFBs) offer feasible solar energy storage with high flexibility in redox couples and storage capacity. Unlike traditional redox flow batteries, homemade flow cell reactors are commonly used in most solar-rechargeable redox flow batteries integrated with photoelectrochemical devices as it provides high system flexibility. This perspective article discusses current trends of the architectural and material characteristics of state-of-the-art photoelectrochemical flow cells for SRFB applications. Key design aspects and guidelines to build a photoelectrochemical flow cell, considering practical operating conditions, are proposed in this perspective.

Construction of STHA-model of geometric geoid on the Lviv region area

Nowadays there is a need to modernize the high system of Ukraine, which requires its integration in the European Vertical Reference System EVRS. In this regard there is also a need to build a regional model of the geoid on the territory of our country, which would be well consistent with the model of the European geoid EGG2015. To obtain the optimal model, it is necessary to use both gravimetric and geometric data. In this case, the model is called gravimetric-geometric. This approach is used both when building a model of the European geoid and when building geoid models on the territory of different European countries. Aim. The purpose of this work is to build a regional geometric STHA-model of the geoid on the Lviv region area and assess its accuracy. In the future it is planned to build a gravimetric STHA-model of the geoid in the same area and compare the results. Methods. To build a geometric STHA-model of the geoid on the Lviv region area, the heights of the geometric geoid, obtained from GNSS-observations at the points of SGN of I, II and III classes, were used. RMS error of determination of geodetic heights , obtained from GNSS leveling in static mode, did not exceed 15 mm. 205 values of the calculated heights of the geoid were used to build the geoid model. 8 values were not involved in the construction of the model, because they were used for an independent assessment of model accuracy. Results. The regional model of geoid within the “Remove–Compute–Restore” procedure with introduction of regularization parameter is obteined. RMS error of the obtained model, calculated on the basis of the data used in its construction, is 12 mm, and on other independent data is 25 mm. Scientific novelty and practical significance. For the first time STHA-functions were tested to build a regional geoid model. The geometric model of the geoid on the Lviv region are is calculated and the accuracy of the obtained model is estimated on the basis of dependent and independent data. The RMS error of the obtained model was about 2 cm, which corresponds to the accuracy of GNSS-measurements. The obtained model can be used as a transformation field on the Lviv region area.

Creating a Core Innovation Capability in Airbus

Flight in modern age caused our society more global than that of past age. People can go to school aboard, go to work aboard, and travel aboard with airplanes. Airplane has changed our society. Airplane manufacturing industry requires sophisticated technology and innovative management. This most intelligent industry’s innovation strategy can be one of epitomes for business strategy. Airplane manufacturing can be one long-term oriented business. Airplane industry requires many different fields of study not only natural science, but also social science. To manage both engineering team and management team, significant organization strategy could be one crucial factor to maintain its organization effective. Airplane industries must innovate to offer more reliable service with safety, low price and fast travel. New version of motor, car, aircraft and TV takes usually high system level and incremental innovation. Airbus is one of leading aircraft manufacturer and its market share is second largest globally. This study would bring about understanding of innovation management at highest technology and cutting-edge industry.

Primary Research of a New Zero-Liquid-Discharge Technology of Wet Flue Gas Desulfurization Wastewater by Low-Rank Heat from Flue Gas

Wet flue gas desulfurization (WFGD) wastewater treatment is a key problem in coal-fired plants. Traditional chemical precipitation methods cannot reach zero-liquid discharge (ZLD). In this paper, a new technology using the low-rank heat from flue gas to concentrate the wastewater for ZLD is proposed. A scrubber was built to verify the concentrating process, and the characteristics of the concentrated water were analyzed. The concentrated water was neutralized by adding Ca(OH)2 to raise the pH value. The wastewater can be concentrated 10~25 times to reduce the flow rate. The characteristics of the concentrated wastewater were studied by dosing lime. Then, liquid and solids were separated by filter pressing, the liquid was mainly composed of CaCl2, which accounts for 73.6%. The sludge is composed of CaSO4 and Mg(OH)2, depending on the lime consumption of the dosing process. Finally, the filter liquor after the filter press was mixed with ash to reach zero liquid discharge, and the sludge could be burnt after mixing with the coal or disposed by third-part vendor. This technology is demonstrated in one 600 MW unit and shows a high system reliability. The clean water is recycled by the WFGD wastewater during the evaporation. Binding on the environmental policies and large market demand of the WFGD wastewater, this technology shows a great application prospect in the future.

MemX

This work presents MemX: a biologically-inspired attention-aware eyewear system developed with the goal of pursuing the long-awaited vision of a personalized visual Memex. MemX captures human visual attention on the fly, analyzes the salient visual content, and records moments of personal interest in the form of compact video snippets. Accurate attentive scene detection and analysis on resource-constrained platforms is challenging because these tasks are computation and energy intensive. We propose a new temporal visual attention network that unifies human visual attention tracking and salient visual content analysis. Attention tracking focuses computation-intensive video analysis on salient regions, while video analysis makes human attention detection and tracking more accurate. Using the YouTube-VIS dataset and 30 participants, we experimentally show that MemX significantly improves the attention tracking accuracy over the eye-tracking-alone method, while maintaining high system energy efficiency. We have also conducted 11 in-field pilot studies across a range of daily usage scenarios, which demonstrate the feasibility and potential benefits of MemX.

Performance Evaluation for Cooperative NOMA Using Combined Relay Selection Scheme

Effective relay selection in non-orthogonal multiple access (NOMA) is an efficient way to attain a significant performance gain in terms of achieving a high system capacity, ergodic sum rate and a low outage probability. This become more promising when NOMA is treated under cooperation of more than one relay nodes. Relay selection performs well when the source node has an accurate and updated channel state information (CSI). As having perfect CSI is very difficult, therefore the relay selection with continuous time varying CSI is of paramount importance. This paper proposes a combined relay selection (CRS) scheme with time varying CSI which uses failure probability as an important criterion to improve the overall performance of cooperative NOMA system. The proposed approach is compared with most commonly used relay selection schemes: round robin relay selection (RRRS), opportunistic relay selection (ORS) and balanced criteria relay selection (BCRS) scheme. The simulation results show that proposed scheme outperforms the RRRS, BCRS and ORS in terms of ergodic sum rate, system capacity and outage probability.