An automatic design optimization tool and its application to computational fluid dynamics

Abstract

In this paper we describe the Nimrod/O design optimization tool, and its application in computational fluid dynamics. Nimrod/O facilitates the use of an arbitrary computational model to drive an automatic optimization process. This means that the user can parameterise an arbitrary problem, and then ask the tool to compute the parameter values that minimize or maximise a design objective function. The paper describes the Nimrod/O system, and then discusses a case study in the evaluation of an aerofoil problem. The problem involves computing the shape and angle of attack of the aerofoil that maximises the lift to drag ratio. The results show that our general approach is extremely flexible and delivers better results than a program that was developed specifically for the problem. Moreover, it only took us a few hours to set up the tool for the new problem and required no software development.

Tuning parallel applications in parallel

“…The figure shows the performance of square matrix multiplication (of size 800 × 800) as a function of tiling and unrolling factors. The search space is not smooth and contains multiple 1 Tiling loop transformation partitions a loop nest's iteration space into small blocks and then iterates through those blocks in sequence. The goal is to maintain a small data footprint for the sub-loop nests for better cache usage.…”

“…The Nelder-Mead Simplex algorithm [61], is one of the most widely used direct search methods in auto-tuning systems [1,23,35]. We provide a high-level description of the algorithm in chapter 3.…”

“…We separate the search-based navigation of the code-transformation parameter space and the codegeneration process, which allows us to easily switch between different underlying code-generation tools (e.g. if we are tuning CUDA code, we can switch to a code-transformation framework that supports GPUs via CUDA or OpenCL) 1 . Second, our code-generation utility can take advantage of idle (possibly remote) machines for distributed code-generation and compilation.…”

“…Recently, there has been a growing interest in developing empirical auto-tuning software that helps programmers manage this tedious process of tuning and porting their codes. Empirical autotuning software (or auto-tuners) can be broadly grouped into three categories: (1) compiler-based auto-tuners that automatically generate and search a set of alternative implementations of a computation [16,95,33]; (2) application-level auto-tuners that automate empirical search across a set of parameter values proposed by the application programmer [18,62]; and, (3) runtime auto-tuners that automate on-thefly adaptation of application-level and architecture-specific parameters to react to the changing conditions of the system that executes the application [11,87]. What is common across all these different categories of auto-tuners is the need to search a range of possible configurations to identify the best-performing solution.…”

Model Optimization and Parameter Estimation with Nimrod/O

“…Flow around an airfoil was modelled, [8], using the computational fluid dynamics package FLUENT. Input parameters were the thickness, camber and angle of attack of the airfoil; the aim was to maximize the ratio of lift to drag.…”