A Static Greedy and Dynamic Adaptive Thread Spawning Approach for Loop-Level Parallelism

Thread-level speculation becomes more attractive for the exploitation of thread-level parallelism from irregular sequential applications. But it is common for speculative threads to fail to reach the expected parallel performance. The reason is that the performance of speculative threads is extremely complicated by the fact that it not only suffers from the imprecision of compiler-directed performance estimation due to ambiguous control and data dependences, but also depends on the underlying hardware configuration and program behaviors. Thus, this paper proposes a statically greedy and dynamically adaptive approach for loop-level speculation to dynamically determine the best loop level at runtime. It relies on the compiler to select and optimize all loop candidates greedily, which are then proceeded on the cost-benefit analysis of different loop nesting levels for the determination of the order of loop speculation. Under the runtime loop execution prediction, we dynamically schedule and update the order of loop speculation, and ensure the best loop level to be always parallelized. Two different policies are also examined to maximize overall performance. Compared with traditional static loop selection techniques, our approach （：an achieve comparable or better performance.

Computational fluid dynamics simulation of hydrodynamics in the riser of an external loop airlift reactor

Local hydrodynamics in the riser of an external loop airlift reactor （EL-ALR） are identified and the performances of three drag models are evaluated in computational fluid dynamics simulation. The simulation results show that the Schiller-Naumann drag model underestimated the local gas holdup at lower superficial gas velocity whereas the Tomiyama drag model overestimated that at higher superficial gas velocity. By contrast, the dual-bubble-size （DBS）-local drag model gave more reasonable radial and axial distri-butions of gas holdup in all cases. The reason is that the DBS-local drag model gave correct values of the lumped parameter, i,e., the ratio of the drag coefficient to bubble diameter, for varying operating conditions and radial positions. This ratio is reasonably expected to decrease with increasing superficial gas velocity and be smaller in the center and larger near the wall. Only the DBS-local drag model correctly reproduced these trends. The radial profiles of the axial velocity of the liquid and gas predicted by the DBS-local model also agreed well with experimental data.

Quantitative analysis of the performance of vector tracking algorithms

Vector tracking changes the classical structure of receivers. Combining signal tracking and navigation solution,vector tracking can realize powerful processing capabilities by the fusion technique of receiving channel and feedback correction. In this paper,we try to break through the complicated details of numerical analysis,consider the overall influencing factors of the residual in observed data,and use the intrinsic link between a conventional receiver and a vector receiver. A simple method for performance analysis of the vector tracking algorithm is proposed. Kalman filter has the same steady performance with the classic digital lock loop through the analysis of the relation between gain and band width. The theoretical analysis by the least squares model shows that the reduction of range error is the basis for the superior performance realized by vector tracking. Thus,the bounds of its performance enhancement under weak signal and highly dynamic conditions can be deduced. Simulation results verify the effectiveness of the analysis presented here.

Loop formation and stability of self-avoiding polymer chains

Using 3-dimensional Langevin dynamics simulations, we investigated the dynamics of loop formation of chains with excluded volume interactions, and the stability of the formed loop. The mean looping time ι1/scales with chain length N and corresponding scaling exponent α increases linearly with the capture radius scaled by the Kuhn length a/l due to the effect of finite chain length. We also showed that the probability density function of the looping time is well fitted by a single exponential. Finally, we found that the mean unlooping time ιu hardly depends on chain length N for a given a/l and that the stability of a formed loop is enhanced with increasing a/l.