| 1 |
|
\appendix |
| 2 |
|
\chapter{\label{chapt:oopse}Object-Oriented Parallel Simulation Engine} |
| 3 |
|
|
| 4 |
< |
Designing object-oriented software is hard, and designing reusable |
| 5 |
< |
object-oriented scientific software is even harder. Absence of |
| 6 |
< |
applying modern software development practices is the bottleneck of |
| 7 |
< |
Scientific Computing community\cite{Wilson2006}. For instance, in |
| 8 |
< |
the last 20 years , there are quite a few MD packages that were |
| 9 |
< |
developed to solve common MD problems and perform robust simulations |
| 10 |
< |
. However, many of the codes are legacy programs that are either |
| 11 |
< |
poorly organized or extremely complex. Usually, these packages were |
| 12 |
< |
contributed by scientists without official computer science |
| 13 |
< |
training. The development of most MD applications are lack of strong |
| 14 |
< |
coordination to enforce design and programming guidelines. Moreover, |
| 15 |
< |
most MD programs also suffer from missing design and implement |
| 16 |
< |
documents which is crucial to the maintenance and extensibility. |
| 17 |
< |
Along the way of studying structural and dynamic processes in |
| 18 |
< |
condensed phase systems like biological membranes and nanoparticles, |
| 19 |
< |
we developed and maintained an Object-Oriented Parallel Simulation |
| 20 |
< |
Engine ({\sc OOPSE}). This new molecular dynamics package has some |
| 21 |
< |
unique features |
| 4 |
> |
Absence of applying modern software development practices is the |
| 5 |
> |
bottleneck of Scientific Computing community\cite{Wilson2006}. In |
| 6 |
> |
the last 20 years , there are quite a few MD |
| 7 |
> |
packages\cite{Brooks1983, Vincent1995, Kale1999} that were developed |
| 8 |
> |
to solve common MD problems and perform robust simulations . |
| 9 |
> |
Unfortunately, most of them are commercial programs that are either |
| 10 |
> |
poorly written or extremely complicate. Consequently, it prevents |
| 11 |
> |
the researchers to reuse or extend those packages to do cutting-edge |
| 12 |
> |
research effectively. Along the way of studying structural and |
| 13 |
> |
dynamic processes in condensed phase systems like biological |
| 14 |
> |
membranes and nanoparticles, we developed an open source |
| 15 |
> |
Object-Oriented Parallel Simulation Engine ({\sc OOPSE}). This new |
| 16 |
> |
molecular dynamics package has some unique features |
| 17 |
|
\begin{enumerate} |
| 18 |
|
\item {\sc OOPSE} performs Molecular Dynamics (MD) simulations on non-standard |
| 19 |
|
atom types (transition metals, point dipoles, sticky potentials, |
| 59 |
|
program of the package, \texttt{oopse} and it corresponding parallel |
| 60 |
|
version \texttt{oopse\_MPI}, as well as other useful utilities, such |
| 61 |
|
as \texttt{StatProps} (see Sec.~\ref{appendixSection:StaticProps}), |
| 62 |
< |
\texttt{DynamicProps} (see |
| 63 |
< |
Sec.~\ref{appendixSection:appendixSection:DynamicProps}), |
| 64 |
< |
\texttt{Dump2XYZ} (see |
| 70 |
< |
Sec.~\ref{appendixSection:appendixSection:Dump2XYZ}), \texttt{Hydro} |
| 71 |
< |
(see Sec.~\ref{appendixSection:appendixSection:hydrodynamics}) |
| 62 |
> |
\texttt{DynamicProps} (see Sec.~\ref{appendixSection:DynamicProps}), |
| 63 |
> |
\texttt{Dump2XYZ} (see Sec.~\ref{appendixSection:Dump2XYZ}), |
| 64 |
> |
\texttt{Hydro} (see Sec.~\ref{appendixSection:hydrodynamics}) |
| 65 |
|
\textit{etc}. |
| 66 |
|
|
| 67 |
|
\begin{figure} |
| 106 |
|
As one of the latest advanced techniques emerged from |
| 107 |
|
object-oriented community, design patterns were applied in some of |
| 108 |
|
the modern scientific software applications, such as JMol, {\sc |
| 109 |
< |
OOPSE}\cite{Meineke05} and PROTOMOL\cite{Matthey05} \textit{etc}. |
| 110 |
< |
The following sections enumerates some of the patterns used in {\sc |
| 111 |
< |
OOPSE}. |
| 109 |
> |
OOPSE}\cite{Meineke2005} and PROTOMOL\cite{Matthey2005} |
| 110 |
> |
\textit{etc}. The following sections enumerates some of the patterns |
| 111 |
> |
used in {\sc OOPSE}. |
| 112 |
|
|
| 113 |
|
\subsection{\label{appendixSection:singleton}Singleton} |
| 114 |
+ |
|
| 115 |
|
The Singleton pattern not only provides a mechanism to restrict |
| 116 |
|
instantiation of a class to one object, but also provides a global |
| 117 |
|
point of access to the object. Currently implemented as a global |
| 121 |
|
pollution.Although the singleton pattern can be implemented in |
| 122 |
|
various ways to account for different aspects of the software |
| 123 |
|
designs, such as lifespan control \textit{etc}, we only use the |
| 124 |
< |
static data approach in {\sc OOPSE}. {\tt IntegratorFactory} class |
| 125 |
< |
is declared as |
| 124 |
> |
static data approach in {\sc OOPSE}. IntegratorFactory class is |
| 125 |
> |
declared as |
| 126 |
> |
|
| 127 |
|
\begin{lstlisting}[float,caption={[A classic Singleton design pattern implementation(I)] The declaration of of simple Singleton pattern.},label={appendixScheme:singletonDeclaration}] |
| 128 |
|
|
| 129 |
|
class IntegratorFactory { |
| 137 |
|
}; |
| 138 |
|
|
| 139 |
|
\end{lstlisting} |
| 140 |
+ |
|
| 141 |
|
The corresponding implementation is |
| 142 |
+ |
|
| 143 |
|
\begin{lstlisting}[float,caption={[A classic implementation of Singleton design pattern (II)] The implementation of simple Singleton pattern.},label={appendixScheme:singletonImplementation}] |
| 144 |
|
|
| 145 |
|
IntegratorFactory::instance_ = NULL; |
| 152 |
|
} |
| 153 |
|
|
| 154 |
|
\end{lstlisting} |
| 155 |
< |
Since constructor is declared as {\tt protected}, a client can not |
| 156 |
< |
instantiate {\tt IntegratorFactory} directly. Moreover, since the |
| 157 |
< |
member function {\tt getInstance} serves as the only entry of access |
| 158 |
< |
to {\tt IntegratorFactory}, this approach fulfills the basic |
| 159 |
< |
requirement, a single instance. Another consequence of this approach |
| 160 |
< |
is the automatic destruction since static data are destroyed upon |
| 161 |
< |
program termination. |
| 155 |
> |
|
| 156 |
> |
Since constructor is declared as protected, a client can not |
| 157 |
> |
instantiate IntegratorFactory directly. Moreover, since the member |
| 158 |
> |
function getInstance serves as the only entry of access to |
| 159 |
> |
IntegratorFactory, this approach fulfills the basic requirement, a |
| 160 |
> |
single instance. Another consequence of this approach is the |
| 161 |
> |
automatic destruction since static data are destroyed upon program |
| 162 |
> |
termination. |
| 163 |
|
|
| 164 |
|
\subsection{\label{appendixSection:factoryMethod}Factory Method} |
| 165 |
|
|
| 167 |
|
with the problem of creating objects without specifying the exact |
| 168 |
|
class of object that will be created. Factory Method is typically |
| 169 |
|
implemented by delegating the creation operation to the subclasses. |
| 170 |
< |
{\tt Integrator} class Parameterized Factory pattern where factory |
| 171 |
< |
method ({\tt createIntegrator} member function) creates products |
| 172 |
< |
based on the identifier (see |
| 173 |
< |
List.~\ref{appendixScheme:factoryDeclaration}). If the identifier |
| 174 |
< |
has been already registered, the factory method will invoke the |
| 175 |
< |
corresponding creator (see List.~\ref{integratorCreator}) which |
| 176 |
< |
utilizes the modern C++ template technique to avoid subclassing. |
| 179 |
< |
\begin{lstlisting}[float,caption={[The implementation of Parameterized Factory pattern (I)]Source code of {\tt IntegratorFactory} class.},label={appendixScheme:factoryDeclaration}] |
| 170 |
> |
Parameterized Factory pattern where factory method ( |
| 171 |
> |
createIntegrator member function) creates products based on the |
| 172 |
> |
identifier (see List.~\ref{appendixScheme:factoryDeclaration}). If |
| 173 |
> |
the identifier has been already registered, the factory method will |
| 174 |
> |
invoke the corresponding creator (see List.~\ref{integratorCreator}) |
| 175 |
> |
which utilizes the modern C++ template technique to avoid excess |
| 176 |
> |
subclassing. |
| 177 |
|
|
| 178 |
+ |
\begin{lstlisting}[float,caption={[The implementation of Parameterized Factory pattern (I)]Source code of IntegratorFactory class.},label={appendixScheme:factoryDeclaration}] |
| 179 |
+ |
|
| 180 |
|
class IntegratorFactory { |
| 181 |
|
public: |
| 182 |
|
typedef std::map<string, IntegratorCreator*> CreatorMapType; |
| 198 |
|
CreatorMapType creatorMap_; |
| 199 |
|
}; |
| 200 |
|
\end{lstlisting} |
| 201 |
+ |
|
| 202 |
|
\begin{lstlisting}[float,caption={[The implementation of Parameterized Factory pattern (III)]Source code of creator classes.},label={appendixScheme:integratorCreator}] |
| 203 |
|
|
| 204 |
|
class IntegratorCreator { |
| 226 |
|
|
| 227 |
|
\subsection{\label{appendixSection:visitorPattern}Visitor} |
| 228 |
|
|
| 229 |
< |
The purpose of the Visitor Pattern is to encapsulate an operation |
| 230 |
< |
that you want to perform on the elements. The operation being |
| 231 |
< |
performed on a structure can be switched without changing the |
| 232 |
< |
interfaces of the elements. In other words, one can add virtual |
| 233 |
< |
functions into a set of classes without modifying their interfaces. |
| 234 |
< |
Fig.~\ref{appendixFig:visitorUML} demonstrates the structure of |
| 235 |
< |
Visitor pattern which is used extensively in {\tt Dump2XYZ}. In |
| 236 |
< |
order to convert an OOPSE dump file, a series of distinct and |
| 237 |
< |
unrelated operations are performed on different StuntDoubles. |
| 238 |
< |
Visitor allows one to keep related operations together by packing |
| 239 |
< |
them into one class. {\tt BaseAtomVisitor} is a typical example of |
| 240 |
< |
visitor in {\tt Dump2XYZ} program{see |
| 241 |
< |
List.~\ref{appendixScheme:visitor}}. In contrast to the operations, |
| 242 |
< |
the object structure or element classes rarely change(See |
| 243 |
< |
Fig.~\ref{oopseFig:heirarchy} and |
| 244 |
< |
List.~\ref{appendixScheme:element}). |
| 229 |
> |
The visitor pattern is designed to decouple the data structure and |
| 230 |
> |
algorithms used upon them by collecting related operation from |
| 231 |
> |
element classes into other visitor classes, which is equivalent to |
| 232 |
> |
adding virtual functions into a set of classes without modifying |
| 233 |
> |
their interfaces. Fig.~\ref{appendixFig:visitorUML} demonstrates the |
| 234 |
> |
structure of Visitor pattern which is used extensively in {\tt |
| 235 |
> |
Dump2XYZ}. In order to convert an OOPSE dump file, a series of |
| 236 |
> |
distinct operations are performed on different StuntDoubles (See the |
| 237 |
> |
class hierarchy in Fig.~\ref{oopseFig:hierarchy} and the declaration |
| 238 |
> |
in List.~\ref{appendixScheme:element}). Since the hierarchies |
| 239 |
> |
remains stable, it is easy to define a visit operation (see |
| 240 |
> |
List.~\ref{appendixScheme:visitor}) for each class of StuntDouble. |
| 241 |
> |
Note that using Composite pattern\cite{Gamma1994}, CompositVisitor |
| 242 |
> |
manages a priority visitor list and handles the execution of every |
| 243 |
> |
visitor in the priority list on different StuntDoubles. |
| 244 |
|
|
| 246 |
– |
|
| 245 |
|
\begin{figure} |
| 246 |
|
\centering |
| 247 |
|
\includegraphics[width=\linewidth]{visitor.eps} |
| 249 |
|
diagram of Visitor patten.} \label{appendixFig:visitorUML} |
| 250 |
|
\end{figure} |
| 251 |
|
|
| 252 |
< |
\begin{lstlisting}[float,caption={[The implementation of Visitor pattern (I)]Source code of the visitor classes.},label={appendixScheme:visitor}] |
| 253 |
< |
|
| 254 |
< |
class BaseVisitor{ |
| 255 |
< |
public: |
| 256 |
< |
virtual void visit(Atom* atom); |
| 257 |
< |
virtual void visit(DirectionalAtom* datom); |
| 260 |
< |
virtual void visit(RigidBody* rb); |
| 261 |
< |
}; |
| 262 |
< |
|
| 263 |
< |
class BaseAtomVisitor:public BaseVisitor{ public: |
| 264 |
< |
virtual void visit(Atom* atom); |
| 265 |
< |
virtual void visit(DirectionalAtom* datom); |
| 266 |
< |
virtual void visit(RigidBody* rb); |
| 267 |
< |
}; |
| 252 |
> |
\begin{figure} |
| 253 |
> |
\centering |
| 254 |
> |
\includegraphics[width=\linewidth]{hierarchy.eps} |
| 255 |
> |
\caption[Class hierarchy for ojects in {\sc OOPSE}]{ A diagram of |
| 256 |
> |
the class hierarchy. } \label{oopseFig:hierarchy} |
| 257 |
> |
\end{figure} |
| 258 |
|
|
| 269 |
– |
\end{lstlisting} |
| 270 |
– |
|
| 259 |
|
\begin{lstlisting}[float,caption={[The implementation of Visitor pattern (II)]Source code of the element classes.},label={appendixScheme:element}] |
| 260 |
|
|
| 261 |
< |
class StuntDouble { |
| 274 |
< |
public: |
| 261 |
> |
class StuntDouble { public: |
| 262 |
|
virtual void accept(BaseVisitor* v) = 0; |
| 263 |
|
}; |
| 264 |
|
|
| 265 |
< |
class Atom: public StuntDouble { |
| 279 |
< |
public: |
| 265 |
> |
class Atom: public StuntDouble { public: |
| 266 |
|
virtual void accept{BaseVisitor* v*} { |
| 267 |
|
v->visit(this); |
| 268 |
|
} |
| 269 |
|
}; |
| 270 |
|
|
| 271 |
< |
class DirectionalAtom: public Atom { |
| 286 |
< |
public: |
| 271 |
> |
class DirectionalAtom: public Atom { public: |
| 272 |
|
virtual void accept{BaseVisitor* v*} { |
| 273 |
|
v->visit(this); |
| 274 |
|
} |
| 275 |
|
}; |
| 276 |
|
|
| 277 |
< |
class RigidBody: public StuntDouble { |
| 293 |
< |
public: |
| 277 |
> |
class RigidBody: public StuntDouble { public: |
| 278 |
|
virtual void accept{BaseVisitor* v*} { |
| 279 |
|
v->visit(this); |
| 280 |
|
} |
| 282 |
|
|
| 283 |
|
\end{lstlisting} |
| 284 |
|
|
| 285 |
+ |
\begin{lstlisting}[float,caption={[The implementation of Visitor pattern (I)]Source code of the visitor classes.},label={appendixScheme:visitor}] |
| 286 |
+ |
|
| 287 |
+ |
class BaseVisitor{ |
| 288 |
+ |
public: |
| 289 |
+ |
virtual void visit(Atom* atom); |
| 290 |
+ |
virtual void visit(DirectionalAtom* datom); |
| 291 |
+ |
virtual void visit(RigidBody* rb); |
| 292 |
+ |
}; |
| 293 |
+ |
|
| 294 |
+ |
class BaseAtomVisitor:public BaseVisitor{ public: |
| 295 |
+ |
virtual void visit(Atom* atom); |
| 296 |
+ |
virtual void visit(DirectionalAtom* datom); |
| 297 |
+ |
virtual void visit(RigidBody* rb); |
| 298 |
+ |
}; |
| 299 |
+ |
|
| 300 |
+ |
class SSDAtomVisitor:public BaseAtomVisitor{ public: |
| 301 |
+ |
virtual void visit(Atom* atom); |
| 302 |
+ |
virtual void visit(DirectionalAtom* datom); |
| 303 |
+ |
virtual void visit(RigidBody* rb); |
| 304 |
+ |
}; |
| 305 |
+ |
|
| 306 |
+ |
class CompositeVisitor: public BaseVisitor { |
| 307 |
+ |
public: |
| 308 |
+ |
|
| 309 |
+ |
typedef list<pair<BaseVisitor*, int> > VistorListType; |
| 310 |
+ |
typedef VistorListType::iterator VisitorListIterator; |
| 311 |
+ |
virtual void visit(Atom* atom) { |
| 312 |
+ |
VisitorListIterator i; |
| 313 |
+ |
BaseVisitor* curVisitor; |
| 314 |
+ |
for(i = visitorList.begin();i != visitorList.end();++i) { |
| 315 |
+ |
atom->accept(*i); |
| 316 |
+ |
} |
| 317 |
+ |
} |
| 318 |
+ |
|
| 319 |
+ |
virtual void visit(DirectionalAtom* datom) { |
| 320 |
+ |
VisitorListIterator i; |
| 321 |
+ |
BaseVisitor* curVisitor; |
| 322 |
+ |
for(i = visitorList.begin();i != visitorList.end();++i) { |
| 323 |
+ |
atom->accept(*i); |
| 324 |
+ |
} |
| 325 |
+ |
} |
| 326 |
+ |
|
| 327 |
+ |
virtual void visit(RigidBody* rb) { |
| 328 |
+ |
VisitorListIterator i; |
| 329 |
+ |
std::vector<Atom*> myAtoms; |
| 330 |
+ |
std::vector<Atom*>::iterator ai; |
| 331 |
+ |
myAtoms = rb->getAtoms(); |
| 332 |
+ |
for(i = visitorList.begin();i != visitorList.end();++i) {{ |
| 333 |
+ |
rb->accept(*i); |
| 334 |
+ |
for(ai = myAtoms.begin(); ai != myAtoms.end(); ++ai){ |
| 335 |
+ |
(*ai)->accept(*i); |
| 336 |
+ |
} |
| 337 |
+ |
} |
| 338 |
+ |
|
| 339 |
+ |
void addVisitor(BaseVisitor* v, int priority); |
| 340 |
+ |
|
| 341 |
+ |
protected: |
| 342 |
+ |
VistorListType visitorList; |
| 343 |
+ |
}; |
| 344 |
+ |
|
| 345 |
+ |
\end{lstlisting} |
| 346 |
+ |
|
| 347 |
|
\section{\label{appendixSection:concepts}Concepts} |
| 348 |
|
|
| 349 |
|
OOPSE manipulates both traditional atoms as well as some objects |
| 350 |
|
that {\it behave like atoms}. These objects can be rigid |
| 351 |
|
collections of atoms or atoms which have orientational degrees of |
| 352 |
< |
freedom. A diagram of the class heirarchy is illustrated in |
| 353 |
< |
Fig.~\ref{oopseFig:heirarchy}. Every Molecule, Atom and |
| 352 |
> |
freedom. A diagram of the class hierarchy is illustrated in |
| 353 |
> |
Fig.~\ref{oopseFig:hierarchy}. Every Molecule, Atom and |
| 354 |
|
DirectionalAtom in {\sc OOPSE} have their own names which are |
| 355 |
|
specified in the {\tt .md} file. In contrast, RigidBodies are |
| 356 |
|
denoted by their membership and index inside a particular molecule: |
| 358 |
|
on the specifics of the simulation). The names of rigid bodies are |
| 359 |
|
generated automatically. For example, the name of the first rigid |
| 360 |
|
body in a DMPC molecule is DMPC\_RB\_0. |
| 361 |
< |
%\begin{figure} |
| 362 |
< |
%\centering |
| 363 |
< |
%\includegraphics[width=\linewidth]{heirarchy.eps} |
| 364 |
< |
%\caption[Class heirarchy for ojects in {\sc OOPSE}]{ A diagram of |
| 365 |
< |
%the class heirarchy. |
| 366 |
< |
%\begin{itemize} |
| 367 |
< |
%\item A {\bf StuntDouble} is {\it any} object that can be manipulated by the |
| 368 |
< |
%integrators and minimizers. |
| 323 |
< |
%\item An {\bf Atom} is a fundamental point-particle that can be moved around during a simulation. |
| 324 |
< |
%\item A {\bf DirectionalAtom} is an atom which has {\it orientational} as well as translational degrees of freedom. |
| 325 |
< |
%\item A {\bf RigidBody} is a collection of {\bf Atom}s or {\bf |
| 326 |
< |
%DirectionalAtom}s which behaves as a single unit. |
| 327 |
< |
%\end{itemize} |
| 328 |
< |
%} \label{oopseFig:heirarchy} |
| 329 |
< |
%\end{figure} |
| 361 |
> |
\begin{itemize} |
| 362 |
> |
\item A {\bf StuntDouble} is {\it any} object that can be manipulated by the |
| 363 |
> |
integrators and minimizers. |
| 364 |
> |
\item An {\bf Atom} is a fundamental point-particle that can be moved around during a simulation. |
| 365 |
> |
\item A {\bf DirectionalAtom} is an atom which has {\it orientational} as well as translational degrees of freedom. |
| 366 |
> |
\item A {\bf RigidBody} is a collection of {\bf Atom}s or {\bf |
| 367 |
> |
DirectionalAtom}s which behaves as a single unit. |
| 368 |
> |
\end{itemize} |
| 369 |
|
|
| 370 |
|
\section{\label{appendixSection:syntax}Syntax of the Select Command} |
| 371 |
|
|
| 372 |
< |
The most general form of the select command is: {\tt select {\it |
| 373 |
< |
expression}}. This expression represents an arbitrary set of |
| 335 |
< |
StuntDoubles (Atoms or RigidBodies) in {\sc OOPSE}. Expressions are |
| 336 |
< |
composed of either name expressions, index expressions, predefined |
| 337 |
< |
sets, user-defined expressions, comparison operators, within |
| 338 |
< |
expressions, or logical combinations of the above expression types. |
| 339 |
< |
Expressions can be combined using parentheses and the Boolean |
| 340 |
< |
operators. |
| 372 |
> |
{\sc OOPSE} provides a powerful selection utility to select |
| 373 |
> |
StuntDoubles. The most general form of the select command is: |
| 374 |
|
|
| 375 |
+ |
{\tt select {\it expression}}. |
| 376 |
+ |
|
| 377 |
+ |
This expression represents an arbitrary set of StuntDoubles (Atoms |
| 378 |
+ |
or RigidBodies) in {\sc OOPSE}. Expressions are composed of either |
| 379 |
+ |
name expressions, index expressions, predefined sets, user-defined |
| 380 |
+ |
expressions, comparison operators, within expressions, or logical |
| 381 |
+ |
combinations of the above expression types. Expressions can be |
| 382 |
+ |
combined using parentheses and the Boolean operators. |
| 383 |
+ |
|
| 384 |
|
\subsection{\label{appendixSection:logical}Logical expressions} |
| 385 |
|
|
| 386 |
|
The logical operators allow complex queries to be constructed out of |
