Integral image, also known as summed area table is a two-dimensional table generated from an input image. Each entry in the table stores the sum of all pixels which locate on the top-left corner of the entry in the input image. Integral image is a very popular and important algorithm in computer vision and computer graphics applications. Especially in real-time computer vision, it is usually used to accelerate calculating the sum of a rectangular area. Integral image algorithm is memory-bounded. There are two typical existed image integral algorithms on GPUs. The first is the Scan-Scan algorithm. The second is the Scan-Transpose-Scan algorithm, which adopts three steps to generate the integral image. The first and the third steps are scan. In order to achieve coalesced global memory access in the third step, a transpose step is added. In this paper, we propose a novel blocked integral algorithm, which has three stages. The first stage is intra-block reduction. The second stage is auxiliary matrix scan and the third stage is intra-block scan. Compared with the Scan-Scan algorithm, our proposed scheme reduces the global memory accesses. At the same time, less local synchronizations and less load imbalance are achieved. Compared with the Scan-Transpose-Scan algorithm, our proposed algorithm only needs about half of the global memory accesses. At the same time, coalesced memory access is achieved. We implemented these three algorithms with OpenCL so that they can run on both Nvidia and AMD GPUs. We also designed an auto-tuning framework to search optimal parameters for different size of input matrix on those two platforms. The experiment result shows that our proposed algorithm gets the best performance compared with the two existed typical integral algorithms.

To achieve scalable data intensive analytics, we investigate methods to integrate general purpose analytic computation into a query pipeline using User Defined Functions (UDFs). However, an existing UDF cannot act as a block operator with chunk-wise input along the tuple-wise query processing pipeline, therefore unable to deal with the application semantics definable on the set of incoming tuples representing a single object or falling in a time window, and unable to leverage external computation engines for efficient batch processing.

While parallel DBMSs do support large scale parallel query processing on partitioned data, the reach of more general applications relies on User Defined Functions (UDFs). However, the existent UDF technology is insufficient both conceptually and practically. A UDF is not a relation-in, relation-out operator, which restricts its ability to model complex applications defined on a set of tuples rather than on a single one, and to be composed with other relational operators in a query. Further, to interact with the query execution efficiently, a UDF must be coded with complex interactions with DBMS internal data structures and system calls which is often beyond the expertise of an analytics application developer.

We consider the problem of joint estimation of direction-of-arrival (DoA), propagation delay, and complex channel gain for antenna-array-based DS/CDMA communication systems over frequency selective multipath channels. We propose a MUSIC-type estimation algorithm which utilizes the spatial smoothing preprocessing technique. The proposed algorithm essentially breaks the multipath-induced coherency within the received signals and recovers the full signal subspace spanned by all dominant signal paths of all users. This allows the use of MUSIC-type DoA and delay estimators for the individual paths of the user of interest. Based on the angle and timing information, we then estimate the multipath fading coefficients. Simulation results illustrate the effectiveness of this approach.

We present a strategy for automatic negotiation that takes the same approach as computer programs that play games such as chess; we build the game tree. For every offer we look at every counteroffer, every counteroffer to each of them, and so on. The strategy then selects the counteroffer that has the largest expected payoff. A number of problems arise that are unique to using this strategy for negotiation. These include uncertainty in the opponent’s goals, the fact that a bad move can penalize both players, and moves that are continuous, as opposed to discrete. We show how the standard methods of building the tree and evaluating the results were adapted to this environment.

Retrograde analysis is an efficient exhaustive search method. It is a powerful tool that can be used in solving problems where end states have known values but starting states do not. It has been widely used to solve mathematically-precise games such as chess endgames, and is potentially usable in energy-minimization problems. With increasing computing power, both in speed and storage capacity, retrograde analysis will become more and more useful. This paper looks at successful applications to games, the challenges ahead and the modifications that are required to utilize distributed hardware. The power and the usefulness of retrograde analysis are still limited by the computing resources one has access to. Today, the best sequential retrograde algorithms are capable of solving problems with about 10/sup 9/ states in a few hours on a standard personal computer. Bigger problems need more powerful computers, or take much longer to solve, or are simply out of the reach of today's technologies. Introducing parallelism to retrograde analysis is a natural way to attack the bigger problems. There are today three main architectures available for doing parallel retrograde analysis, namely symmetric multiprocessor (SMP) systems, high-speed network-based distributed systems and Internet-based distributed systems. In this paper, we discuss some of the key issues in doing parallel retrograde analysis on these different architectures. Technical challenges are addressed in detail, as well as some examples and proposals. These examples and proposals are drawn from various board games, but the ideas can be applied to other problem domains.