Parallelization Strategy for Hierarchical Run Length Encoded Data Structures

نویسندگان

  • Lado Filipović
  • Otmar Ertl
  • Siegfried Selberherr
چکیده

An efficient parallelization strategy is presented for a Hierarchical Run Length Encoded (HRLE) data structure, implemented for the Sparse Field Level Set method. In order to achieve high parallel efficiency, computational work must be distributed evenly over all available CPU threads. Since the Level Set surface must be allowed to deform and evolve, thereby increasing the simulation area, there must exist a way to increase the surface domain while keeping an efficient parallelization strategy in place. This is achieved by processing the same number of calculations across each available CPU. The addition of data to HRLE data structures is only permitted in a sequential or lexicographical order, making parallelization more complex. The presented solution uses as many HRLE data structures as there are CPUs available. Approximately 90% of operations can be performed in parallel when using the presented strategy, leading to an efficiency of up to 96% or 78.5% when using two or sixteen CPU cores of an AMD Opteron 8435 processor, clocked at 2.6GHz, respectively. Topographies with one and two moving interfaces were simulated using multi-threading, showing the speedup and efficiency for the presented strategy.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A fast hierarchical traversal strategy for multimodal visualization

In the last years there is a growing demand of multimodal medical rendering systems able to visualize simultaneously data coming from different sources. This paper addresses the Direct Volume Rendering (DVR) of aligned multimodal data in medical applications. Specifically, it proposes a hierarchical representation of the multimodal data set based on the construction of a Fusion Decision Tree (F...

متن کامل

Parallelization Requirements for Heirarchically Structured

In this paper we address a class of problems consisting of highly structured computations on data sets that are described by hierarchical data structures. These are often represented as tree structures to optimize data storage requirements and perform e cient queries for data access. Speci cally, applications that are dynamic and perform many iterations on data are of interest to us, since the ...

متن کامل

A Hybrid Coding Strategy For Optimized Test Data Compression

Store-and-generate techniques encode a given test set and regenerate the original test set during the test with the help of a decoder. Previous research has shown that run-length coding, particularly alternating run-length coding, can provide high compression ratios for the test data. However, experimental data show that longer runlengths are distributed sparsely in the code space and often occ...

متن کامل

Parallelization of Rich Models for Steganalysis of Digital Images using a CUDA-based Approach

There are several different methods to make an efficient strategy for steganalysis of digital images. A very powerful method in this area is rich model consisting of a large number of diverse sub-models in both spatial and transform domain that should be utilized. However, the extraction of a various types of features from an image is so time consuming in some steps, especially for training pha...

متن کامل

Data Dependence Analysis for the Parallelization of Numerical Tree Codes

Data dependence analysis for automatic parallelization of sequential tree codes is discussed. Hierarchical numerical algorithms often use tree data structures for unbalanced, adaptively and dynamically created trees. Moreover, such codes often do not follow a strict divide and conquer concept, but introduce some geometric neighborhood data dependence in addition to parent-children dependencies....

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2010