Truong Vinh Truong Duy

Performance evaluation of a Green Scheduling Algorithm for energy savings in Cloud computing

With energy shortages and global climate change leading our concerns these days, the power consumption of datacenters has become a key issue. Obviously, a substantial reduction in energy consumption can be made by powering down servers when they are not in use. This paper aims at designing, implementing and evaluating a Green Scheduling Algorithm integrating a neural network predictor for optimizing server power consumption in Cloud computing. We employ the predictor to predict future load demand based on historical demand. According to the prediction, the algorithm turns off unused servers and restarts them to minimize the number of running servers, thus minimizing the energy use at the points of consumption to benefit all other levels. For evaluation, we perform simulations with two load traces. The results show that the PP20 mode can save up to 46.3% of power consumption with a drop rate of 0.03% on one load trace, and a drop rate of 0.12% with a power reduction rate of 46.7% on the other.

Improving accuracy of host load predictions on computational grids by artificial neural networks

The capability to predict the host load of a system is significant for computational grids to make efficient use of shared resources. This work attempts to improve the accuracy of host load predictions by applying a neural network predictor to reach the goal of best performance and load balance. We describe the feasibility of the proposed predictor in a dynamic environment, and perform experimental evaluation using collected load traces. The results show that the neural network achieves consistent performance improvement with surprisingly low overhead in most cases. Compared with the best previously proposed method, our typical 20:10:1 network reduces the mean of prediction errors by approximately up to 79%. The training and testing time is extremely low, as this network needs only a couple of seconds to be trained with more than 100,000 samples, in order to make tens of thousands of accurate predictions within just a second.

A massively parallel domain decomposition method for large-scale DFT electronic structure calculations

We present a massively parallel domain decomposition method for atoms and grids to enable large-scale density functional theory (DFT) electronic structure calculations. In the atom decomposition, we develop a modified recursive bisection method based on the moment of inertia tensor for reordering the atoms from 3D to 1D along a principal axis so that atoms that are close in real space are also close on the axis to ensure data locality. The atoms are then divided into sub-domains depending on their projections onto the principal axis in a balanced way among the processes. In the grid decomposition, we define four data structures to make data locality consistent with that of the clustered atoms, and propose a 2D decomposition method for solving the Poisson equation using the 3D FFT with communication volume minimized. Benchmark results show that the parallel efficiency at 131,072 cores is 67.7\% compared to the baseline of 16,384 cores on the K computer.

A decomposition method with minimal communication volume for parallelization of multi-dimensional FFTs

We present a decomposition method for the parallelization of multi-dimensional FFTs with two distinguishing features: adaptive decomposition and transpose order awareness for achieving minimal communication volume. Based on a row-wise decomposition that translates the multi-dimensional data into one-dimensional data for equally allocating to the processes, our method can adaptively decompose the data in the lowest possible dimensions to reduce communication volume in the first place, differently from previous works that have pre-defined dimensions of decomposition. Also, this decomposition offers plenty of orders in data transpose, and different transpose orders result in different volumes of communication. By analyzing all the possible cases, we find out the best transpose orders with minimal communication volumes for 3-D, 4-D, and 5-D FFTs.