3 Copyright (C) 1989-2013 Regents of the University of Illinois
8 Charm++ is a message-passing parallel language and runtime system.
9 It is implemented as a set of libraries for C++, is efficient,
10 and is portable to a wide variety of parallel machines.
11 Source code is provided, and non-commercial use is free.
14 GETTING THE LATEST CHARM SOURCE
15 ===============================
17 You can use anonymous Git access to obtain the latest Charm++ source
20 git clone http://charm.cs.illinois.edu/gerrit/charm
26 First, you need to decide which version of charm++ to use. The "build"
27 script in charm source directory takes several command line options to
28 compile Charm++. The command line syntax is:
30 build <target> <version> [options ...]
31 [--basedir=dir] [--libdir=dir] [--incdir=dir]
34 For detailed help messages, pass -h or --help to the build script, i.e.
39 <target> specifies the parts of Charm++ to compile. The most often used
40 <target> is "charm++", which will compile the key Charm++ executables and
41 runtime libraries. Other common targets are "AMPI" and "FEM".
42 <versions> defines the CPU, OS and Communication layer of the machines. See
43 "How to choose a <version>" below for details.
47 <options> defines more detailed information of the compilations, including
48 compilers, features to support, etc. See "How to choose <option>"
50 [--libdir=dir] specify additional lib paths for building Charm++.
51 [--incdir=dir] specify additional include paths for building Charm++.
52 [--basedir=dir] a shortcut to specify additional include and lib paths for
53 building Charm++, the include path is dir/include and lib path
57 Running build script, a directory of the name of combination of version and
58 options like "<version>-<option1>-<option2>-..." will be created and
59 the build script will compile Charm++ under this directory.
61 For example, on an ordinary Linux PC:
63 ./build charm++ net-linux-x86_64
65 will build charm++ in the directory: net-linux-x86_64/. The communication
66 defaults to UDP packets and the compiler to gcc.
68 For a more complex example, on a Scyld workstation with the Intel C++
69 compiler, where you want the communication to happen over TCP sockets:
71 ./build charm++ net-linux scyld icc tcp
73 will build charm++ in the directory: net-linux-scyld-tcp-icc/.
75 You can specify multiple options, however you can use at most one compiler
76 option. The sequence of the options are not important given in build script,
77 only one directory name will be generated, following the rules:
78 1. compiler option will be at the end;
79 2. other options are sorted alphabetically.
81 **** How to choose a <version> ****
83 Here is the table for choosing correct version. The default setting of compiler
84 in Charm version is gcc/g++. However, one can use <options> to specify other
85 compilers. See the detailed explanation of the <options> below.
86 (Note: this isn't a complete list. Run ./build for a complete listing)
88 Charm version OS Communication Default Compiler
89 ------------- --------- -------------- --------------------
90 net-linux-x86_64 Opteron Linux UDP GNU compiler
91 net-darwin-x86_64 MacOS X UDP GNU C++ compiler
92 net-linux Linux (32 bit) UDP/Myrinet GNU compiler
93 net-sol Solaris UDP GNU compiler
94 net-win32 Win32 UDP MS Visual C++
95 net-cygwin Win32/cygwin UDP GNU compiler
96 mpi-sp IBM A/IX MPI A/IX xlC Compiler
97 net-linux-ia64 IA64 Linux UDP/Myrinet GNU compiler
98 mpi-linux PC Linux MPI GNU compiler
99 mpi-darwin-x86_64 MacOS X MPI GNU C++ compiler
100 mpi-linux-ia64 IA64 Linux MPI GNU compiler
101 mpi-linux-x86_64 Opteron Linux MPI GNU compiler
104 To choose <version>, your choice is determined by three options:
106 1.) The way a parallel program written in Charm++ will communicate:
108 "net-" Charm++ communicates using the regular TCP/IP stack
109 (UDP packets), which works everywhere but is fairly slow. Use this
110 option for networks of workstations, clusters, or single-machine
111 development and testing.
113 "gni-", "bluegene[lp]-", "pami-bluegeneq-" Charm++
114 communicates using direct calls to the machine's communication primitives.
116 "mpi-" Charm++ communicates using MPI calls. Use this for
117 machines with a good MPI implementation (such as the Origin 2000).
119 "multicore-" Charm++ communicates using shared memory within a
122 "sim-" and "uth-" are not actively maintained. These are
123 single-processor versions: "uth-" simulates processors as user-level
124 threads; "sim-" switches between processors and counts communications.
127 2.) Your operating system:
130 "win{32,64}" MS Windows with MS Visual C++ compiler (32/64-bit, resp.)
131 "cygwin" MS Windows with Cygnus' Cygwin Unix layer
132 "darwin" Apple Mac OS X
137 3.) Some operating systems have other architecture options, such as:
139 "-x86" For Solaris and Mac OS X, target x86 hardware (instead of
142 "-mips64" MIPS, such as for SiCortex systems
143 "-ia64" Use Itanium(tm) IA-64 instructions (instead of x86).
144 "-x86_64" Use AMD64/EM64T 64-bit x86 instructions (instead of 32 bit).
145 "-cell" Sony/Toshiba/IBM Cell PPE (e.g. Playstation 3,
146 Mercury blades, Roadrunner)
148 Your Charm++ version is made by concatenating the options, e.g.:
150 "net-linux-x86_64" Charm++ for a network of 64-bit Linux workstations,
153 "gni-crayxc" Charm++ for Cray XC30 systems using the system's compiler.
156 **** How to choose <options> ****
158 <version> above defines the most important OS, CPU and Communication of
159 your machine, and most of time, it use the GNU gcc as default compiler.
160 To use different compiler or demand additional special feature support, you
161 need to choose <options> from the following list:
163 * gcc3 - GNU GCC/G++ version 3
164 * acc - HP aC++ compiler
165 * cc - For Sun WorkShop C++ compilers;
166 * cc64 - For 64 bits Sun WorkShop C++ or IBM xlC compilers;
167 * cxx - DIGITAL C++ compiler;
168 * kcc - KAI C++ compiler;
169 * pgcc - Portland Group's C++ compiler;
170 * icc - Intel C/C++ compiler for Linux IA32
171 * ecc - Intel C/C++ compiler for Linux IA64
172 * mpcc - SUN Solaris C++ compiler for MPI.
174 * scyld - support Beowulf Scyld based on bproc;
175 * clustermatic - for Clustermatic Beowulf cluster based on bproc;
176 * gm - support MyriCom's Myrinet GM library;
177 * vmi - support NCSA's VMI library;
179 * tcp - for net- version, default communication is via UDP. Using option
180 tcp will switch to TCP. TCP version of CHarm++ is usually slower
181 than UDP, but it is more reliable.
182 * smp - Enable direct SMP support. An "smp" version communicates using
183 shared memory within a machine; but normal message passing across
184 machines. Because of locking, "smp" may slightly impact non-SMP
185 performance. Try your application to decide if enabling smp mode
186 improves performance.
188 * bigsim - compile Charm++ as running on the BigSim emulator.
189 * help - show supported options for a version. For example, for net-linux,
191 > ./build charm++ net-linux help
193 supported options: gcc3 gm icc kcc pgcc scyld smp bluegene tcp
199 If you have downloaded a binary version of Charm++, you can skip
200 this step-- Charm++ should already be compiled. For win32/win64 systems,
201 see README.win; for Cygwin version, see README.cygwin; for net- version,
204 Once you have decided on a version, unpack Charm++, cd into charm,
207 > ./build <target> <version> <opts>
210 "charm++" The basic Charm++ language.
211 "AMPI" An implementation of MPI on top of Charm++
212 "FEM" A Finite-Element framework on top of Charm++
213 "Tau" TAU's performance profiling/tracing
215 <version> is described above
217 <opts> are build-time options (such as the compiler or "smp"),
218 or command line options passed to the charmc compile script.
219 Common compile time options such as -g, -O, -Ipath, -Lpath, -llib are
222 For example, on a Linux machine, you would run
223 > ./build charm++ net-linux-x86_64 -O
225 This will construct a net-linux-x86_64 directory, link over all
226 the Charm++ source code into net-linux-x86_64/tmp, build the entire
227 Charm++ runtime system in net-linux-x86_64/tmp, and link example programs
228 into net-linux-x86_64/examples.
230 Charm++ can be compiled with several optional features enabled or
231 disabled. These include runtime error checking, tracing, interactive
232 debugging, deterministic record-replay, and more. They can be
233 controlled by passing flags of the form --enable-featurename or
234 --disable-featurename to the build command:
235 > ./build charm++ net-linux-x86_64 --disable-tracing
237 The list of optional features available is shown in the output of
240 Production optimizations: Pass the configure option --with-production
241 to ./build to turn on optimizations in Charm++/Converse. This disables
242 most of the run-time checking performed by Converse and Charm++
243 runtime. This option should be used only after the program has been
244 debugged. Also, this option disables Converse/Charm++ tracing
245 mechanisms such as projections and summary.
247 When Charm++ is built successfully, the diretory structure under the
248 target directory will look like:
252 --- bin/ # all executables
254 --- doc/ # documentations
256 --- include/ # header files
260 --- lib_so/ # dynamic libraries
262 --- examples/ # all example programs
264 --- tests/ # all test programs
266 --- tmp/ # Charm++ build directory
271 To make a sample program, cd into examples/charm++/queens/.
272 This program solves the N-queens problem-- find how many ways there
273 are to arrange N queens on an NxN chess board such that none may
276 To build the program, type make. You should get an
277 executable named "pgm".
283 Following the previous example, to run the program on two processors, type
285 > ./charmrun ./pgm 12 6 +p2
287 This should run for a few seconds, and print out:
288 There are 14200 Solutions to 12 queens. Finish time=4.030000
290 Charmrun is used to provide a uniform interface to run charm programs.
291 On some platforms, charmrun is just a shell script which calls the
292 platform-specific start program, such as mpirun on mpi versions.
294 For net- version, charmrun is an executable which invokes ssh to start
295 node programs on remote machines. You should set up a ~/.nodelist that
296 enumerates all the machines you want to run jobs on, otherwise it will
297 create a default ~/.nodelist for you that contains only localhost. Here is a
298 typical .nodelist file:
300 group main ++shell /bin/ssh
303 The default remote shell program is ssh, but you can define a different remote
304 shell to start remote processes using the ++shell option. You should
305 also make sure that ssh or your alternative can connect to these machines without
306 password authentication. Just type following command to verify:
307 > ssh <machinename> date
308 If this gives you current date immediately, your running environment with this
309 node has been setup correctly.
311 Now, for test running purpose, net- version charmrun comes with an easy-to-use
312 "++local" options. No remote shell invocation is needed in this case. It starts
313 node programs right on your local machine. This could be useful if you just
314 want to run program on only one machine, for example, your laptop. This
315 can save you all the hassle of setting up ssh or charmd daemons.
316 To use this option, just type:
318 > ./charmrun ++local ./pgm 12 100 +p2
320 However, for best performance, you should launch one node program per processor.
324 Build Charm++ in Dynamic libraries
325 =============================
327 In order to compile Charm++ into dynamic libraries, one need to specify
328 "-build-shared" option as one of the Charm compiler script "charmc"
329 at link time. For example, to compile Charm++ under net-linux/tmp, run
331 make charm++ OPTS='-O -build-shared'
333 Charm++'s dynamic libraries are compiled into lib_so/ directory.
334 Typically, they are with ".so" suffix.
336 Note, "-build-shared" option is automatically turned on when building
337 Charm++ using "build" script. So you don't need to pass "-build-shared"
340 One can compile a Charm++ applicaiton linking against Charm++ dynamic
341 libraries, run charmc with "-charm-shared" as one of the link options.
344 charmc -o pgm pgm.o -charm-shared
346 You can then run the program as usual.
347 Note, linking against Charm++ dynamic libraries produces much smaller size
348 binaries and takes much less linking time.
353 The Charm++ web page, with documentation, more programs,
354 and the latest version of Charm++, is at
355 http://charm.cs.illinois.edu/
357 The Charm++ mailing list, for questions, comments, suggestions,
358 improvements, or bug reports is
359 charm@cs.illinois.edu
365 Charm++ was created and is maintained by the Parallel Programming Lab,
366 in the Computer Science department at the University of Illinois at
367 Urbana-Champaign. Our managing professor is Dr. L.V. Kale; students
368 have included (in rough time order) Wennie Shu, Kevin Nomura, Wayne
369 Fenton, Balkrishna Ramkumar, Vikram Saletore, Amitabh B. Sinha, Manish
370 Gupta, Attila Gursoy, Balkrishna Ramkumar, Amitabh B. Sinha, Nimish
371 Shah, Sanjeev Krishnan, Jayant DeSouza, Parthasarathy Ramachandran,
372 Jeff Wright, Michael Lang, Jackie Wang, Fang Hu, Michael Denardo,
373 Joshua Yelon, Narain Jagathesan, Zehra Sura, Krishnan Varadarajan,
374 Sameer Paranjpye, Milind Bhandarkar, Robert Brunner and Jayant Desouza.
375 Current developers include Terry Wilmarth, Gengbin Zheng, Orion Lawlor,
376 Karthik Mahesh, and Neelam Saboo.