Multithreaded transactions in scientific computing: New versions of a computer program for kinematical calculations of RHEED intensity oscillations

Writing a concurrent program can be more difficult than writing a sequential program. Programmer needs to think about synchronisation, race conditions and shared variables. Transactions help reduce the inconvenience of using threads. A transaction is an abstraction, which allows programmers to group a sequence of actions on the program into a logical, higher-level computation unit. This paper presents multithreaded versions of the GROWTH program, which allow to calculate the layer coverages during the growth of thin epitaxial films and the corresponding RHEED intensities according to the kinematical approximation. The presented programs also contain graphical user interfaces, which enable displaying program data at run-time.New version program summaryTitles of programs:GROWTHGr, GROWTH06Catalogue identifier:ADVL_v2_0Program summary URL:http://cpc.cs.qub.ac.uk/summaries/ADVL_v2_0Program obtainable from:CPC Program Library, Queen's University of Belfast, N. IrelandCatalogue identifier of previous version:ADVLDoes the new version supersede the original program:NoComputer for which the new version is designed and others on which it has been tested: Pentium-based PCOperating systems or monitors under which the new version has been tested: Windows 9x, XP, NTProgramming language used:Object PascalMemory required to execute with typical data:More than 1 MBNumber of bits in a word:64 bitsNumber of processors used:1No. of lines in distributed program, including test data, etc.:20 931Number of bytes in distributed program, including test data, etc.: 1 311 268Distribution format:tar.gzNature of physical problem: The programs compute the RHEED intensities during the growth of thin epitaxial structures prepared using the molecular beam epitaxy (MBE). The computations are based on the use of kinematical diffraction theory [P.I. Cohen, G.S. Petrich, P.R. Pukite, G.J. Whaley, A.S. Arrott, Surf. Sci. 216 (1989) 222. [1]].Method of solution: Epitaxial growth ofFig. 1. Internal structure of the program.Figure optionsDownload as PowerPoint slide thin films is modelledFig. 2. Static classes model for graphical user interface.Figure optionsDownload as PowerPoint slide by a set ofFig. 3. Activity diagram for the program.Figure optionsDownload as PowerPoint slide non-linear differential equations [P.I. Cohen, G.S. Petrich, P.R. Pukite, G.J. Whaley, A.S. Arrott, Surf. Sci. 216 (1989) 222. [1]].Fig. 4. TTransaction class contents.Figure optionsDownload as PowerPoint slide The Runge–Kutta method with adaptive stepsize control was used for solving initial valueFig. 5. TGrowthTransaction class contents.Figure optionsDownload as PowerPoint slideprs.rt("abs_1st_end");