Comprehensive Path-sensitive Data-flow Analysis

Data-flow analysis is an integral part of any aggressive optimizing compiler. We propose a framework for improving the precision of data-flow analysis in the presence of complex control-flow. W initially perform data-flow analysis to determine those control-flow merges which cause the loss in data-flow analysis precision. The control-flow graph of the program is then restructured such that performing data-flow analysis on the resulting restructured graph gives more precise results. The proposed framework is both simple, involving the familiar notion of product automata, and also general, since it is applicable to any forward data-flow analysis. Apart from proving that our restructuring process is correct, we also show that restructuring is effective in that it necessarily leads to more optimization opportunities. Furthermore, the framework handles the trade-off between the increase in data-flow precision and the code size increase inherent in the restructuring. We show that determining an optimal restructuring is NP-hard, and propose and evaluate a greedy strategy. The framework has been implemented in the Scale research compiler, and instantiated for the specific problem of Constant Propagation. On the SPECINT 2000 benchmark suite we observe an average speedup of 4% in the running times over Wegman-Zadeck conditional constant propagation algorithm and 2% over a purely path profile guided approach.

Power Flow Analysis in MTDC-AC System—A New Approach

This paper presents a new approach to the power flow analysis in steady state for multiterminal DC-AC systems. A flexible and practical choice of per unit system is used to formulate the DC network and converter equations. A converter is represented by Norton's equivalent of a current source in parallel with the commutation resistance. Unlike in previous literature, the DC network equations are used to derive the controller equations for the DC system using a subset of specifications. The specifications considered are current or power at all terminals except the slack terminal where the DC voltage is specified. The control equations are solved by Newton's method, using the current injections at the converter terminals as state variables. Further, a systematic approach to the handling of constraints is proposed by identifying the priorities in rescheduling of the specified variables. The methodology is illustrated by example of a 5 terminal DC system.

Packet flow analysis in IP networks using data-flow analysis

Static analysis (aka offline analysis) of a model of an IP network is useful for understanding, debugging, and verifying packet flow properties of the network. Data-flow analysis is a method that has typically been applied to static analysis of programs. We propose a new, data-flow based approach for static analysis of packet flows in networks. We also investigate an application of our analysis to the problem of inferring a high-level policy from the network, which has been addressed in the past only for a single router.

Computational Analysis of Flow through a Multiple Nozzle Driven Laser Cavity and Diffuser

Computations have been carried out for simulating supersonic flow through a set of converging-diverging nozzles with their expanding jets forming a laser cavity and flow patterns through diffusers, past the cavity. A thorough numerical investigation with 3-D RANS code is carried out to capture the flow distribution which comprises of shock patterns and multiple supersonic jet interactions. The analysis of pressure recovery characteristics during the flow through the diffusers is an important parameter of the simulation and is critical for the performance of the laser device. The results of the computation have shown a close agreement with the experimentally measured parameters as well as other established results indicating that the flow analysis done is found to be satisfactory.

Comparative studies of transient and steady state analysis for a typical 765kV/400kV EHV transmission system in Indian power system

This paper describes an approach for the analysis and design of 765kV/400kV EHV transmission system which is a typical expansion in Indian power grid system, based on the analysis of steady state and transient over voltages. The approach for transmission system design is iterative in nature. The first step involves exhaustive power flow analysis, based on constraints such as right of way, power to be transmitted, power transfer capabilities of lines, existing interconnecting transformer capabilities etc. Acceptable bus voltage profiles and satisfactory equipment loadings during all foreseeable operating conditions for normal and contingency operation are the guiding criteria. Critical operating strategies are also evolved in this initial design phase. With the steady state over voltages obtained, comprehensive dynamic and transient studies are to be carried out including switching over voltages studies. This paper presents steady state and switching transient studies for alternative two typical configurations of 765kV/400 kV systems and the results are compared. Transient studies are carried out to obtain the peak values of 765 kV transmission systems and are compared with the alternative configurations of existing 400 kV systems.

Investigation of schemes for incorporating generator Q limits in the fast decoupled load flow method

Fast Decoupled Load Flow (FDLF) is a very popular and widely used power flow analysis method because of its simplicity and efficiency. Even though the basic FDLF algorithm is well investigated, the same is not true in the case of additional schemes/modifications required to obtain adjusted load flow solutions using the FDLF method. Handling generator Q limits is one such important feature needed in any practical load flow method. This paper presents a comprehensive investigation of two classes of schemes intended to handle this aspect i.e. the bus type switching scheme and the sensitivity scheme. We propose two new sensitivity based schemes and assess their performance in comparison with the existing schemes. In addition, a new scheme to avoid the possibility of anomalous solutions encountered while using the conventional schemes is also proposed and evaluated. Results from extensive simulation studies are provided to highlight the strengths and weaknesses of these existing and proposed schemes, especially from the point of view of reliability.

A Network Model of Static Foam Drainage

On the basis of a more realistic tetrakaidecahedral structure of foam bubbles, a network model of static foam drainage has been developed. The model considers the foam to be made up of films and Plateau borders. The films drain into the adjacent Plateau borders, which in turn form a network through which the liquid moves from the foam to the liquid pool. From the structure, a unit flow cell was found, which constitutes the foam when stacked together both horizontally and vertically. Symmetry in the unit flow cell indicates that the flow analysis of a part of it can be employed to obtain the drainage for the whole foam. Material balance equations have been written for each segment of this subsection, ensuring connectivity, and solved with the appropriate boundary and initial conditions. The calculated rates of drainage, when compared with the available experimental results, indicate that the model predicts the experimental results well.

Bed Shear from Velocity Profiles: A New Approach

A new binary law of velocity distribution has been developed to describe the velocity profile for the entire flow region. The law is a combination of logarithmic law, valid in the wall (inner) region, and parabolic law, valid in the core (outer) region of the flow. The validity of the law has been established based on earlier data on flat plates, rough and smooth pipes and experimental data obtained from rigid-walled open channels with plane sand beds. A procedure of estimating bed shear stress from the proposed law of velocity distribution using the measured velocity profile has been evolved. Bed shear estimates made according to this procedure are in agreement with the values obtained from uniform flow analysis in the case of open channel flow over a sediment bed. The proposed method of estimating the bed shear stress from the observed velocity profiles is found to be particularly useful in cases where it is difficult to determine precisely the true bed level, such as in the case of flow over sediment beds.

Automatic Compilation of MATLAB Programs for Synergistic Execution on Heterogeneous Processors

MATLAB is an array language, initially popular for rapid prototyping, but is now being increasingly used to develop production code for numerical and scientific applications. Typical MATLAB programs have abundant data parallelism. These programs also have control flow dominated scalar regions that have an impact on the program's execution time. Today's computer systems have tremendous computing power in the form of traditional CPU cores and throughput oriented accelerators such as graphics processing units(GPUs). Thus, an approach that maps the control flow dominated regions to the CPU and the data parallel regions to the GPU can significantly improve program performance. In this paper, we present the design and implementation of MEGHA, a compiler that automatically compiles MATLAB programs to enable synergistic execution on heterogeneous processors. Our solution is fully automated and does not require programmer input for identifying data parallel regions. We propose a set of compiler optimizations tailored for MATLAB. Our compiler identifies data parallel regions of the program and composes them into kernels. The problem of combining statements into kernels is formulated as a constrained graph clustering problem. Heuristics are presented to map identified kernels to either the CPU or GPU so that kernel execution on the CPU and the GPU happens synergistically and the amount of data transfer needed is minimized. In order to ensure required data movement for dependencies across basic blocks, we propose a data flow analysis and edge splitting strategy. Thus our compiler automatically handles composition of kernels, mapping of kernels to CPU and GPU, scheduling and insertion of required data transfer. The proposed compiler was implemented and experimental evaluation using a set of MATLAB benchmarks shows that our approach achieves a geometric mean speedup of 19.8X for data parallel benchmarks over native execution of MATLAB.

Comparison of Optimum Reactive Power Schedule with Different Objectives Using LP Technique

Three algorithms for reactive power optimization are proposed in this paper with three different objective functions. The objectives in the proposed algorithm are to minimize the sum of the squares of the voltage deviations of the load buses, minimization of sum of squares of voltage stability L-indices of load buses (:3L2) algorithm, and also the objective of system real power loss (Ploss) minimization. The approach adopted is an iterative scheme with successive power flow analysis using decoupled technique and solution of the linear programming problem using upper bound optimization technique. Results obtained with all these objectives are compared. The analysis of these objective functions are presented to illustrate their advantages. It is observed comparing different objective functions it is possible to identify critical On Load Tap Changers (OLTCs) that should be made manual to avoid possible voltage instability due to their operation based on voltage improvement criteria under heavy load conditions. These algorithms have been tested under simulated conditions on few test systems. The results obtained on practical systems of 24-node equivalent EHV Indian power network, and for a 205 bus EHV system are presented for illustration purposes.

Automatic compilation of MATLAB programs for synergistic execution on heterogeneous processors

MATLAB is an array language, initially popular for rapid prototyping, but is now being increasingly used to develop production code for numerical and scientific applications. Typical MATLAB programs have abundant data parallelism. These programs also have control flow dominated scalar regions that have an impact on the program's execution time. Today's computer systems have tremendous computing power in the form of traditional CPU cores and throughput oriented accelerators such as graphics processing units(GPUs). Thus, an approach that maps the control flow dominated regions to the CPU and the data parallel regions to the GPU can significantly improve program performance. In this paper, we present the design and implementation of MEGHA, a compiler that automatically compiles MATLAB programs to enable synergistic execution on heterogeneous processors. Our solution is fully automated and does not require programmer input for identifying data parallel regions. We propose a set of compiler optimizations tailored for MATLAB. Our compiler identifies data parallel regions of the program and composes them into kernels. The problem of combining statements into kernels is formulated as a constrained graph clustering problem. Heuristics are presented to map identified kernels to either the CPU or GPU so that kernel execution on the CPU and the GPU happens synergistically and the amount of data transfer needed is minimized. In order to ensure required data movement for dependencies across basic blocks, we propose a data flow analysis and edge splitting strategy. Thus our compiler automatically handles composition of kernels, mapping of kernels to CPU and GPU, scheduling and insertion of required data transfer. The proposed compiler was implemented and experimental evaluation using a set of MATLAB benchmarks shows that our approach achieves a geometric mean speedup of 19.8X for data parallel benchmarks over native execution of MATLAB.

Numerical and experimental investigation and optimization of an open wheeled race car cooling air duct

In this study the cooling performance due to air flow and aerodynamics of the Formula Student open wheeled race car has been investigated and optimized with the help of CFD simulations and experimental validation. The race car in context previously suffered from overheating problems. Flow analysis was carried out based on the detailed race car 3D model (NITK Racing 2012 formula student race car). Wind tunnel experiments were carried out on the same. The results obtained from the computer simulations are compared with experimental results obtained from wind tunnel testing of the full car. Through this study it was possible to locate the problem areas and hence choose the best configuration for the cooling duct. The CFD analysis helped in calculating the mass flow rate, pressure and velocity distribution for different velocities of the car which is then used to determine the heat dissipated by the radiator. Area of flow separation could be visualized and made sure smooth airflow into the radiator core area. This significantly increased the cooling performance of the car with reduction in drag.