Category: fftw

Posted on

GROMACS 5.0.1, nVidia CUDA Toolkit, And FFTW3 Under Ubuntu 14.04 LTS (64-bit); The Virtues Of VirtualBox

Summarized below are the catches and fixes from a recent effort to build GROMACS 5.0.1 with FFTW3 (single- and double-precision) and GPU support (so, single-precision). Also, a trick I've been doing with great success lately, using a virtual machine to keep my real machine as clean as possible.

0. The Virtues Of VirtualBox

Open source means never having to say you're sorry.

I've made the above proclamation to anyone who'd listen lately who has any interest in using Linux software (because, regardless of what anyone says on the matter, it ain't there yet as an operating system for general scientific users with general computing know-how). You will very likely find yourself stuck at a configure or make step in one or more prerequisite codes to some final build you're trying to do, leaving yourself to google error messages to try to come up with some kind of solution. Invariably, you'll try something that seems to work, only to find it doesn't, potentially leaving a trail of orphaned files, version-breaking changes, and random downgrading only to find something else stupid (or not) fixed your build problems.

I've an install I'm quite happy with that has all of the working code I want on it working – and I've no interest in having to perform re-installs to get back to a working state again.

My solution, which I've used to great success with GAMESS-US, GROMACS, NWChem, and Amber (so far), is to break a virtual instance in VirtualBox first. For those who don't know (and briefly), VirtualBox lets you install a fully-working OS inside of your own OS that simply sits as a file in a Virtual VM folder in your user directory. My procedure has been to install a 60 GB VirtualBox instance of (currently) Ubuntu 14.04 (which I will refer to here as PROTOTYPE), fully update it to the current state of my RealBox (updates, upgrades, program installs, etc.), then copy PROTOTYPE somewhere else on the machine. The only limitation of this approach is that VirtualBox doesn't give you access to the GPU if you're testing CUDA-specific calculations. That said, it does let you install the CUDA Development Toolkit and compile code just fine, so you can at least work your way through a full build to make sure you don't run into problems.

When you're done trashing your VirtualBox after a particularly heinous build, just delete PROTOTYPE from Virtual VM and re-copy your copy back into Virtual VM – voila! You're ready for another build operation (or to make sure your "final" build actually works flawlessly before committing the build to your RealBox.

That's all I have to say on the matter. Consider it as your default procedure (at this point, I won't touch my RealBox with new installs until I know it's safe in VirtualBox).

1. The State Of My Machine Pre-GROMACS And All Other apt-get's Used Below

What follows below is pretty straightforward. Errors you might get that don't appear below might be related to the lack of certain installs on your machine that I installed on VirtualBox. That is, my standard PROTOTYPE comes standard with Intel's Fortran and C Compilers (for code optimization). Those installs required a few installs above the base Ubuntu install. These are (and are pretty standard anyway, so I say install them anyway): 

sudo apt-get install build-essential gcc-multilib rpm openjdk-7-jre-headless

I could have just installed a fresh version of 14.04 onto a machine to try this myself, but I'm not that motivated. Also, note this list does not include the all-important cmake. We'll get to that.

And for the rest of GROMACS (at least for older versions), there were lots of mesa/gnuplot/motif-specific dependencies in older versions of GROMACS to build all of the files included in the GROMACS package. Regardless of GPU builds or not, I tend to default to install all the packages below just to have them (which all, for 14.04, currently apt-get properly).

sudo apt-get install openmpi-bin openmpi-common gfortran csh grace menu x11proto-print-dev motif-clients freeglut3-dev libx11-dev libxmu-dev libxi-dev libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev libgl1-mesa-dri libcurl-ocaml-dev libcurl4-gnutls-dev gnuplot

If you don't install the libblas3gf libblas-doc libblas-dev liblapack3gf liblapack-doc liblapack-dev series, you'll see the following note from your cmake steps in GROMACS.

— A library with BLAS API not found. Please specify library location.
— Using GROMACS built-in BLAS.
— LAPACK requires BLAS
— A library with LAPACK API not found. Please specify library location.
— Using GROMACS built-in LAPACK.

My own preference is to use the (assumedly newer) Ubuntu-specific libraries from apt-get.

sudo apt-get install libblas3gf libblas-doc libblas-dev liblapack3gf liblapack-doc liblapack-dev

GPU-Specific? One More apt-get

My first passes at proper GPU compilation involved several steps for the nVidia Developer Toolkit install. That's now taken care of with apt-get, so perform the final apt-get to complete the component/dependency installations.

sudo apt-get install nvidia-cuda-dev nvidia-cuda-toolkit

With luck, your first attempt at a GPU-based installation will look like the following:

[0%] Building NVCC (Device) object src/gromacs/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_copyrite_gpu.cu.o
…
[100%] Building CXX object src/programs/CMakeFiles/gmx.dir/legacymodules.cpp.o
Linking CXX executable https://www.somewhereville.com/bin/gmx
[100%] Built target gmx

2. Nothing Happens Without cmake

Install cmake! Reproducing the output below to make sure you're using the same versions for everything (in the event something breaks in the future).

sudo apt-get install cmake

Reading package lists… Done
Building dependency tree
Reading state information… Done
The following packages were automatically installed and are no longer required:
linux-headers-3.13.0-32 linux-headers-3.13.0-32-generic
linux-image-3.13.0-32-generic linux-image-extra-3.13.0-32-generic
Use 'apt-get autoremove' to remove them.
The following extra packages will be installed:
cmake-data
Suggested packages:
codeblocks eclipse
The following NEW packages will be installed:
cmake cmake-data
0 upgraded, 2 newly installed, 0 to remove and 0 not upgraded.
Need to get 3,294 kB of archives.
After this operation, 16.6 MB of additional disk space will be used.
Do you want to continue? [Y/n]
Get:1 http://us.archive.ubuntu.com/ubuntu/ trusty/main cmake-data all 2.8.12.2-0ubuntu3 [676 kB]
Get:2 http://us.archive.ubuntu.com/ubuntu/ trusty/main cmake amd64 2.8.12.2-0ubuntu3 [2,618 kB]
Fetched 3,294 kB in 30s (106 kB/s)
Selecting previously unselected package cmake-data.
(Reading database … 258157 files and directories currently installed.)
Preparing to unpack …/cmake-data_2.8.12.2-0ubuntu3_all.deb …
Unpacking cmake-data (2.8.12.2-0ubuntu3) …
Selecting previously unselected package cmake.
Preparing to unpack …/cmake_2.8.12.2-0ubuntu3_amd64.deb …
Unpacking cmake (2.8.12.2-0ubuntu3) …
Processing triggers for man-db (2.6.7.1-1) …
Setting up cmake-data (2.8.12.2-0ubuntu3) …
Setting up cmake (2.8.12.2-0ubuntu3) …

3. First Pass At GROMACS

The make install step will place GROMACS where you want it on your machine, so you're just as good building in $HOME/Downloads as you are anywhere else. I will be performing all operations from $HOME/Downloads unless otherwise stated.

According to the GROMACS Installation Manual, your quick-and-dirty install need only involve the following:

$ tar xvfz gromacs-src.tar.gz
$ ls
gromacs-src
$ mkdir build
$ cd build
$ cmake ../gromacs-src
$ make

This allows you build "out-of-source" as they put it. Frankly, I just dive right into the GROMACS folder and have at it.

CMake Error: The source directory "/home/user/Downloads/gromacs-5.0.1/build" does not appear to contain CMakeLists.txt.
Specify –help for usage, or press the help button on the CMake GUI.

And did you see the above error? If so, you read the GROMACS quick-and-dirty procedure backwards. I'm not running it this way, so doesn't matter to what follows.

My first attempt at building GROMACS produced the following output from PROTOTYPE (reproducing all the text below).

user@PROTOTYPE:~$ cd Downloads/
user@PROTOTYPE:~/Downloads$ gunzip gromacs-5.0.1.tar.gz 
user@PROTOTYPE:~/Downloads$ tar xvf gromacs-5.0.1.tar

gromacs-5.0.1/README
gromacs-5.0.1/INSTALL
…
gromacs-5.0.1/tests/CppCheck.cmake
gromacs-5.0.1/tests/CMakeLists.txt 

user@PROTOTYPE:~/Downloads$ cd gromacs-5.0.1/
user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=OFF

NOTE: If you just run cmake, you'll get the following…

cmake version 2.8.12.2
Usage

cmake [options] cmake [options] …

… which is to say, cmake requires at least one option be specified. Above, I'm just using -DGMX_GPU=OFF to start the process.

The C compiler identification is GNU 4.8.2
— The CXX compiler identification is GNU 4.8.2
— Check for working C compiler: /usr/bin/cc
— Check for working C compiler: /usr/bin/cc — works
— Detecting C compiler ABI info
— Detecting C compiler ABI info – done
— Check for working CXX compiler: /usr/bin/c++
— Check for working CXX compiler: /usr/bin/c++ — works
— Detecting CXX compiler ABI info
— Detecting CXX compiler ABI info – done
— Checking for GCC x86 inline asm
— Checking for GCC x86 inline asm – supported
— Detecting best SIMD instructions for this CPU
— Detected best SIMD instructions for this CPU – SSE2
— Try OpenMP C flag = [-fopenmp]
— Performing Test OpenMP_FLAG_DETECTED
— Performing Test OpenMP_FLAG_DETECTED – Success
— Try OpenMP CXX flag = [-fopenmp]
— Performing Test OpenMP_FLAG_DETECTED
— Performing Test OpenMP_FLAG_DETECTED – Success
— Found OpenMP: -fopenmp
— Performing Test CFLAGS_WARN
— Performing Test CFLAGS_WARN – Success
— Performing Test CFLAGS_WARN_EXTRA
— Performing Test CFLAGS_WARN_EXTRA – Success
— Performing Test CFLAGS_WARN_REL
— Performing Test CFLAGS_WARN_REL – Success
— Performing Test CFLAGS_WARN_UNINIT
— Performing Test CFLAGS_WARN_UNINIT – Success
— Performing Test CFLAGS_EXCESS_PREC
— Performing Test CFLAGS_EXCESS_PREC – Success
— Performing Test CFLAGS_COPT
— Performing Test CFLAGS_COPT – Success
— Performing Test CFLAGS_NOINLINE
— Performing Test CFLAGS_NOINLINE – Success
— Performing Test CXXFLAGS_WARN
— Performing Test CXXFLAGS_WARN – Success
— Performing Test CXXFLAGS_WARN_EXTRA
— Performing Test CXXFLAGS_WARN_EXTRA – Success
— Performing Test CXXFLAGS_WARN_REL
— Performing Test CXXFLAGS_WARN_REL – Success
— Performing Test CXXFLAGS_EXCESS_PREC
— Performing Test CXXFLAGS_EXCESS_PREC – Success
— Performing Test CXXFLAGS_COPT
— Performing Test CXXFLAGS_COPT – Success
— Performing Test CXXFLAGS_NOINLINE
— Performing Test CXXFLAGS_NOINLINE – Success
— Looking for include file unistd.h
— Looking for include file unistd.h – found
— Looking for include file pwd.h
— Looking for include file pwd.h – found
— Looking for include file dirent.h
— Looking for include file dirent.h – found
— Looking for include file time.h
— Looking for include file time.h – found
— Looking for include file sys/time.h
— Looking for include file sys/time.h – found
— Looking for include file io.h
— Looking for include file io.h – not found
— Looking for include file sched.h
— Looking for include file sched.h – found
— Looking for include file regex.h
— Looking for include file regex.h – found
— Looking for C++ include regex
— Looking for C++ include regex – not found
— Looking for posix_memalign
— Looking for posix_memalign – found
— Looking for memalign
— Looking for memalign – found
— Looking for _aligned_malloc
— Looking for _aligned_malloc – not found
— Looking for gettimeofday
— Looking for gettimeofday – found
— Looking for fsync
— Looking for fsync – found
— Looking for _fileno
— Looking for _fileno – not found
— Looking for fileno
— Looking for fileno – found
— Looking for _commit
— Looking for _commit – not found
— Looking for sigaction
— Looking for sigaction – found
— Looking for sysconf
— Looking for sysconf – found
— Looking for rsqrt
— Looking for rsqrt – not found
— Looking for rsqrtf
— Looking for rsqrtf – not found
— Looking for sqrtf
— Looking for sqrtf – not found
— Looking for sqrt in m
— Looking for sqrt in m – found
— Looking for clock_gettime in rt
— Looking for clock_gettime in rt – found
— Checking for sched.h GNU affinity API
— Performing Test sched_affinity_compile
— Performing Test sched_affinity_compile – Success
— Check if the system is big endian
— Searching 16 bit integer
— Looking for sys/types.h
— Looking for sys/types.h – found
— Looking for stdint.h
— Looking for stdint.h – found
— Looking for stddef.h
— Looking for stddef.h – found
— Check size of unsigned short
— Check size of unsigned short – done
— Using unsigned short
— Check if the system is big endian – little endian
— Found LibXml2: /usr/lib/x86_64-linux-gnu/libxml2.so (found version "2.9.1")
— Looking for xmlTextWriterEndAttribute in /usr/lib/x86_64-linux-gnu/libxml2.so
— Looking for xmlTextWriterEndAttribute in /usr/lib/x86_64-linux-gnu/libxml2.so – found
— Looking for include file libxml/parser.h
— Looking for include file libxml/parser.h – found
— Looking for include file pthread.h
— Looking for include file pthread.h – found
— Looking for pthread_create
— Looking for pthread_create – not found
— Looking for pthread_create in pthreads
— Looking for pthread_create in pthreads – not found
— Looking for pthread_create in pthread
— Looking for pthread_create in pthread – found
— Found Threads: TRUE
— Looking for include file pthread.h
— Looking for include file pthread.h – found
— Atomic operations found
— Performing Test PTHREAD_SETAFFINITY
— Performing Test PTHREAD_SETAFFINITY – Success
— Could NOT find Boost
Boost >= 1.44 not found. Using minimal internal version. This may cause trouble if you plan on compiling/linking other software that uses Boost against Gromacs.
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so – found
— Setting build user/date/host/cpu information
— Setting build user & time – OK
— Checking floating point format
— Checking floating point format – IEEE754 (LE byte, LE word)
— Checking for 64-bit off_t
— Checking for 64-bit off_t – present
— Checking for fseeko/ftello
— Checking for fseeko/ftello – present
— Checking for SIGUSR1
— Checking for SIGUSR1 – found
— Checking for pipe support
— Checking for isfinite
— Performing Test isfinite_compile_ok
— Performing Test isfinite_compile_ok – Success
— Checking for isfinite – yes
— Checking for _isfinite
— Performing Test _isfinite_compile_ok
— Performing Test _isfinite_compile_ok – Failed
— Checking for _isfinite – no
— Checking for _finite
— Performing Test _finite_compile_ok
— Performing Test _finite_compile_ok – Failed
— Checking for _finite – no
— Performing Test CXXFLAG_STD_CXX0X
— Performing Test CXXFLAG_STD_CXX0X – Success
— Performing Test GMX_CXX11_SUPPORTED
— Performing Test GMX_CXX11_SUPPORTED – Success
— Checking for system XDR support
— Checking for system XDR support – present
— Try C compiler SSE2 flag = [-msse2]
— Performing Test C_FLAG_msse2
— Performing Test C_FLAG_msse2 – Success
— Performing Test C_SIMD_COMPILES_FLAG_msse2
— Performing Test C_SIMD_COMPILES_FLAG_msse2 – Success
— Try C++ compiler SSE2 flag = [-msse2]
— Performing Test CXX_FLAG_msse2
— Performing Test CXX_FLAG_msse2 – Success
— Performing Test CXX_SIMD_COMPILES_FLAG_msse2
— Performing Test CXX_SIMD_COMPILES_FLAG_msse2 – Success
— Enabling SSE2 SIMD instructions
— Performing Test _callconv___vectorcall
— Performing Test _callconv___vectorcall – Failed
— Performing Test _callconv___regcall
— Performing Test _callconv___regcall – Failed
— Performing Test _callconv_
— Performing Test _callconv_ – Success
— checking for module 'fftw3f'
— package 'fftw3f' not found
— pkg-config could not detect fftw3f, trying generic detection
Could not find fftw3f library named libfftw3f, please specify its location in CMAKE_PREFIX_PATH or FFTWF_LIBRARY by hand (e.g. -DFFTWF_LIBRARY='/path/to/libfftw3f.so')
CMake Error at cmake/gmxManageFFTLibraries.cmake:76 (MESSAGE):
Cannot find FFTW 3 (with correct precision – libfftw3f for mixed-precision
GROMACS or libfftw3 for double-precision GROMACS). Either choose the right
precision, choose another FFT(W) library (-DGMX_FFT_LIBRARY), enable the
advanced option to let GROMACS build FFTW 3 for you
(-GMX_BUILD_OWN_FFTW=ON), or use the really slow GROMACS built-in fftpack
library (-DGMX_FFT_LIBRARY=fftpack).
Call Stack (most recent call first):
CMakeLists.txt:733 (include)

— Configuring incomplete, errors occurred!
See also "/home/user/Downloads/gromacs-5.0.1/CMakeFiles/CMakeOutput.log".
See also "/home/user/Downloads/gromacs-5.0.1/CMakeFiles/CMakeError.log".

Lots of little things to address here. We'll get to the Boost problem later. Meantime, you can see the critical error is in (1) the lack of FFTW3 and (2) the lack of my specifically asking for -DGMX_BUILD_OWN_FFTW=ON in the cmake process.

NOTE: If you try to fix the FFTW3 problem as described above, you'll get the following error:

-GMX_BUILD_OWN_FFTW=ON

CMake Error: Could not create named generator MX_BUILD_OWN_FFTW=ON

Make sure to put the "D" in:

-DGMX_BUILD_OWN_FFTW=ON

4. If You Don't Use DGMX_BUILD_OWN_FFTW=ON To Build FFTW3…

This is a skip-able section if you're letting cmake do the dirty work (and letting cmake do it is preferred, at least for getting GROMACS built). In trying sudo apt-get install fftw*, you see (currently) the following: fftw2 fftw-dev fftw-docs

No good. So, the procedure is to build FFTW3 from source (which is just as easy as installing from .deb or .rpm files if you installed everything I mentioned above). That said, your attempts to build FFTW3 and build GROMACS may have run into several errors because of how you built FFTW3. Beginning with your extracting and prep for make:

user@PROTOTYPE:~/Downloads$ tar xvf fftw-3.3.4.tar 
user@PROTOTYPE:~/Downloads$ cd fftw-3.3.4/

Any of the combinations below produce the same error:

user@PROTOTYPE:~/Downloads/fftw-3.3.4$ ./configure 
user@PROTOTYPE:~/Downloads/fftw-3.3.4$ ./configure -enable-shared=yes
user@PROTOTYPE:~/Downloads/fftw-3.3.4$ ./configure --enable-threads --enable-float

checking for a BSD-compatible install… /usr/bin/install -c
checking whether build environment is sane… yes
…
config.status: executing depfiles commands
config.status: executing libtool commands 

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=OFF
user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=OFF -DFFTWF_LIBRARY='/usr/local/lib/libfftw3.a'

— The C compiler identification is GNU 4.8.2
— The CXX compiler identification is GNU 4.8.2
— Check for working C compiler: /usr/bin/cc
…
— Performing Test PTHREAD_SETAFFINITY
— Performing Test PTHREAD_SETAFFINITY – Success
— Could NOT find Boost
Boost >= 1.44 not found. Using minimal internal version. This may cause trouble if you plan on compiling/linking other software that uses Boost against Gromacs.
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so – found
…
— checking for module 'fftw3f'
— package 'fftw3f' not found
— pkg-config could not detect fftw3f, trying generic detection
Could not find fftw3f library named libfftw3f, please specify its location in CMAKE_PREFIX_PATH or FFTWF_LIBRARY by hand (e.g. -DFFTWF_LIBRARY='/path/to/libfftw3f.so')
CMake Error at cmake/gmxManageFFTLibraries.cmake:76 (MESSAGE):
Cannot find FFTW 3 (with correct precision – libfftw3f for mixed-precision
GROMACS or libfftw3 for double-precision GROMACS). Either choose the right
precision, choose another FFT(W) library (-DGMX_FFT_LIBRARY), enable the
advanced option to let GROMACS build FFTW 3 for you
(-GMX_BUILD_OWN_FFTW=ON), or use the really slow GROMACS built-in fftpack
library (-DGMX_FFT_LIBRARY=fftpack).
Call Stack (most recent call first):
CMakeLists.txt:733 (include)

— Configuring incomplete, errors occurred!
See also "/home/user/Downloads/gromacs-5.0.1/CMakeFiles/CMakeOutput.log".
See also "/home/user/Downloads/gromacs-5.0.1/CMakeFiles/CMakeError.log".

Including –enable-shared takes care of this error and gets you to a successful GROMACS build.

user@PROTOTYPE:~/Downloads/fftw-3.3.4$ ./configure --enable-threads --enable-float --enable-shared

— The C compiler identification is GNU 4.8.2
— The CXX compiler identification is GNU 4.8.2
— Check for working C compiler: /usr/bin/cc
…
— Performing Test PTHREAD_SETAFFINITY
— Performing Test PTHREAD_SETAFFINITY – Success
— Could NOT find Boost
Boost >= 1.44 not found. Using minimal internal version. This may cause trouble if you plan on compiling/linking other software that uses Boost against Gromacs.
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so – found
…
— checking for module 'fftw3f'
— found fftw3f, version 3.3.4
— Looking for fftwf_plan_r2r_1d in /usr/local/lib/libfftw3f.so
— Looking for fftwf_plan_r2r_1d in /usr/local/lib/libfftw3f.so – found
— Looking for fftwf_have_simd_avx in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_simd_avx in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_simd_sse2 in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_simd_sse2 in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_simd_avx in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_simd_avx in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_simd_altivec in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_simd_altivec in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_simd_neon in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_simd_neon in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_sse2 in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_sse2 in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_sse in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_sse in /usr/local/lib/libfftw3f.so – not found
— Looking for fftwf_have_altivec in /usr/local/lib/libfftw3f.so
— Looking for fftwf_have_altivec in /usr/local/lib/libfftw3f.so – not found
CMake Warning at cmake/gmxManageFFTLibraries.cmake:89 (message):
The fftw library found is compiled without SIMD support, which makes it
slow. Consider recompiling it or contact your admin
Call Stack (most recent call first):
CMakeLists.txt:733 (include)

— Using external FFT library – FFTW3
— Looking for sgemm_
…
— Configuring done
— Generating done
— Build files have been written to: /home/user/Downloads/gromacs-5.0.1

And out of a first-pass GROMACS build…

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=OFF

Scanning dependencies of target libgromacs
[0%] Building C object src/gromacs/CMakeFiles/libgromacs.dir/__/external/tng_io/src/compression/bwlzh.c.o
[0%] Building C object src/gromacs/CMakeFiles/libgromacs.dir/__/external/tng_io/src/compression/bwt.c.o
…
[100%] Building CXX object src/programs/CMakeFiles/gmx.dir/legacymodules.cpp.o
Linking CXX executable https://www.somewhereville.com/bin/gmx
[100%] Built target gmx

5. But You Let cmake Build FFTW3. So, Continuing The Build Process

With all of the dependencies above installed, the one note I wanted to address was that for Boost:

…
— Performing Test PTHREAD_SETAFFINITY – Success
— Could NOT find Boost
Boost >= 1.44 not found. Using minimal internal version. This may cause trouble if you plan on compiling/linking other software that uses Boost against Gromacs.
— Looking for zlibVersion in /usr/lib/x86_64-linux-gnu/libz.so
…

It certainly isn't a major issue, but I wanted to try to get an warning-free build. Installing Boost 1.56 produced the following negative result:

user@PROTOTYPE:~/Downloads/boost_1_56_0$ ./bootstrap.sh

Building Boost.Build engine with toolset gcc… tools/build/src/engine/bin.linuxx86_64/b2
Detecting Python version… 2.7
Detecting Python root… /usr
Unicode/ICU support for Boost.Regex?… not found.
Generating Boost.Build configuration in project-config.jam…

Bootstrapping is done. To build, run:

./b2

To adjust configuration, edit 'project-config.jam'.
Further information:

– Command line help:
./b2 –help

– Getting started guide:
http://www.boost.org/more/getting_started/unix-variants.html

– Boost.Build documentation:
http://www.boost.org/boost-build2/doc/html/index.html 

user@PROTOTYPE:~/Downloads/boost_1_56_0$ sudo ./b2 install

Performing configuration checks

– 32-bit : no (cached)
– 64-bit : yes (cached)
– arm : no (cached)
…
…failed updating 58 targets…
…skipped 12 targets…
…updated 11322 targets… 

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=ON -DGMX_DOUBLE=OFF
user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ make

[0%] Building NVCC (Device) object src/gromacs/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_copyrite_gpu.cu.o
[0%] Building NVCC (Device) object src/gromacs/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o
…
[7%] Building CXX object src/gromacs/CMakeFiles/libgromacs.dir/commandline/cmdlinehelpwriter.cpp.o
In file included from /home/user/Downloads/gromacs-5.0.1/src/gromacs/options/basicoptions.h:52:0,
from /home/user/Downloads/gromacs-5.0.1/src/gromacs/commandline/cmdlinehelpwriter.cpp:55:
/home/user/Downloads/gromacs-5.0.1/src/gromacs/options/../utility/gmxassert.h:47:57: fatal error: boost/exception/detail/attribute_noreturn.hpp: No such file or directory
#include
^
compilation terminated.
make[2]: *** [src/gromacs/CMakeFiles/libgromacs.dir/commandline/cmdlinehelpwriter.cpp.o] Error 1
make[1]: *** [src/gromacs/CMakeFiles/libgromacs.dir/all] Error 2
make: *** [all] Error 2

Sadly, the solution is to then include -DGMX_EXTERNAL_BOOST=off and stick with the internal boost, which then "makes" just fine. One page references the use of -DGMX_INTERNAL_BOOST=on, but that produced the following:

CMake Warning:
Manually-specified variables were not used by the project:

GMX_INTERNAL_BOOST

— Build files have been written to: /home/user/Downloads/gromacs-5.0.1

There's more on this issue at: gerrit.gromacs.org/#/c/1232/ and t24960.science-biology-gromacs-development.biotalk.us/compiling-boost-problem-and-error-with-icc-t24960.html, but I've opted not to worry about it.

So, with Boost installed, I simply ignore it (and have not installed Boost on my RealBox).

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=ON -DGMX_EXTERNAL_BOOST=off

6. Finishing Step If All Above Goes Well: CUDA-Based GROMACS Build

If everything else above has gone smoothly (and if you ignored the Boost install. If you didn't, remember to add -DGMX_EXTERNAL_BOOST=off to the cmake below), you should be able to cleanly run a cmake for a GPU version of GROMACS (below, with the final result to be placed into /opt/gromacs_gpu. You then specify the $PATH after and run with it).

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ cmake -DGMX_GPU=ON -DCMAKE_INSTALL_PREFIX=/opt/gromacs_gpu -DGMX_BUILD_OWN_FFTW=ON

— The C compiler identification is GNU 4.8.2
— The CXX compiler identification is GNU 4.8.2
…
— Generating done
— Build files have been written to: /home/damianallis/Downloads/gromacs-5.0.1 

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ make

The make starts with the FFTW3 download and build…

Scanning dependencies of target fftwBuild
[ 0%] Performing pre-download step for 'fftwBuild'
— downloading…
src='http://www.fftw.org/fftw-3.3.3.tar.gz'
dest='/home/damianallis/Downloads/gromacs-5.0.1/src/contrib/fftw/fftw.tar.gz'
— [download 0% complete]
…
[100%] Building CXX object src/programs/CMakeFiles/gmx.dir/legacymodules.cpp.o
Linking CXX executable https://www.somewhereville.com/bin/gmx
[100%] Built target gmx

Finally, your (sudo) make install places everything into /opt/gromacs_gpu.

user@PROTOTYPE:~/Downloads/gromacs-5.0.1$ sudo make install

— The GROMACS-managed build of FFTW 3 will configure with the following optimizations: –enable-sse2
— Configuring done
— Generating done
— Build files have been written to: /home/damianallis/Downloads/gromacs-5.0.1
[1%] Built target fftwBuild
…
[100%] Building CXX object src/programs/CMakeFiles/gmx.dir/legacymodules.cpp.o
Linking CXX executable https://www.somewhereville.com/bin/gmx
[100%] Built target gmx

Posted on

Compiling LAMMPS (3Apr13, But Likely Others) In Ubuntu 10.04 Part 1. Using MPICH2 And FFTW2 (And Ubuntu Notes On Installing Intel Fortran And C++ Composers XE for Linux)

I'll qualify this post by saying that (1) I have given up on Ubuntu 11.x and 12.x because they are consistently unstable on my hardware (so, if you have issues running this installation on those versions, I may not be of much help (although I suspect things should work)), (2) I am starting this install from a fresh 32-bit Desktop Ubuntu 10.04 install (so do not know if there are any issues with other software one might have installed on a Linux box if a problem comes up), and (3) the procedure comes out of the current lack of an Ubuntu binary currently listed as available (as of 6 April 2013) from the LAMMPS website (lammps.sandia.gov/download.html#ubuntu). If (3) changes and is available in an MPI form, what's below will hopefully be unnecessary.

Building Trouble And Solutions

My initial "just unzip, untar, and make linux" attempt on a fresh 10.04 install produced the following error (which I'm reproducing in the expectation that you found this page by typing one of the errors below into a search engine, so you'll find the error and the solutions). NOTE: I build all my programs in /opt for organizational purposes (so replace accordingly):

user@machine:/opt/lammps-3Apr13/src$ sudo make linux 

make[1]: Entering directory `/opt/lammps-3Apr13/src/Obj_linux'
icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../write_restart.cpp > write_restart.d
/bin/sh: icc: not found
icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../write_data.cpp > write_data.d
/bin/sh: icc: not found
icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../verlet.cpp > verlet.d
/bin/sh: icc: not found
icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../velocity.cpp > velocity.d
/bin/sh: icc: not found
...
icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../angle_charmm.cpp > angle_charmm.d
/bin/sh: icc: not found
icc -O  -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -c ../angle_charmm.cpp
make[1]: icc: Command not found
make[1]: *** [angle_charmm.o] Error 127
make[1]: Leaving directory `/opt/lammps-3Apr13/src/Obj_linux'
make: *** [linux] Error 2

Obviously, problem Number 1 is the lack of the Intel C Compiler (icc). My solution to this was to download the non-commercial version of the Intel C++ Composer XE for Linux (then I grabbed the Intel Fortran Composer XE for Linux because, well, why not?) currently available from software.intel.com/en-us/non-commercial-software-development (which means this link may be subject to change, so search for "intel c++ noncommercial" in the event).

Unzipping, untaring, and running ./install.sh will, on a fresh Ubuntu install, give you errors that g++ and a proper Java runtime environment are not available on the computer (as part of the pre-requisite search). This is easily solved before the Intel installs by the following (one of which is needed for LAMMPS anyway). I specifically chose the openJDK, but Java 6 or 7 should also do.

sudo apt-get install build-essential openjdk-6-*

After these installs, both Intel Composers should install just fine. If you're installing LAMMPS into a directory that you, the user, has access to, then adding /opt/intel/bin to your PATH will provide you no compiler errors (related to location). If you attempt to install LAMMPS in a directory you, the user, do not have access to, you have to run the install with sudo, which then doesn't likely have the /opt/intel/bin directory in the root PATH, in which case the solution is simply to add a symbolic link for icc into /user/local/bin.

sudo ln -s /opt/intel/bin/icc /usr/local/bin

With icc accessible, running another sudo make linux produces the following errors:

sudo make clean-all
sudo make linux 

make[1]: Entering directory `/opt/lammps-3Apr13/src/Obj_linux'
icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../write_restart.cpp > write_restart.d
../write_restart.cpp(15): catastrophic error: cannot open source file "mpi.h"
  #include "mpi.h"
                  ^

icc -O -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../write_data.cpp > write_data.d
../write_data.cpp(15): catastrophic error: cannot open source file "mpi.h"
  #include "mpi.h"
                  ^
...

icc -O  -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -c ../angle_charmm.cpp
../pointers.h(25): catastrophic error: cannot open source file "mpi.h"
  #include "mpi.h"
                  ^

compilation aborted for ../angle_charmm.cpp (code 4)
make[1]: *** [angle_charmm.o] Error 4
make[1]: Leaving directory `/opt/lammps-3Apr13/src/Obj_linux'
make: *** [linux] Error 2

The new problem is now an MPI problem, requiring the installation of (as recommended from the LAMMPS linux Makefile) MPICH. Reading the Makefile (you mean you didn't check this first?) also indicates the need for FFTW2. We can install all of the needed files in one shot with apt-get (and I include build-essential here for completeness, as you either did or didn't install it while trying to solve the icc problem above).

sudo apt-get install build-essential mpich-bin libmpich1.0-dev mpi-doc fftw2 fftw-dev libxaw7-dev libmpich2-dev

This still does not solve the mpi.h issue above, which is a problem with the settings in Makefile.linux (which, apparently, are not Ubuntu-compatible). The solution is, and this is why this is Part 1, to change Makefile.linux to make it compatible with the mpicc compiler in Ubuntu (I will attempt to get it working with icc next). The tweak to my Makefile.linux is provided below (an amalgam of some other Makefile settings. Comment out or delete what's there and add the sections without the "#" to the file):

#######################################################################
## My modified settings are below:

CC = mpicxx
CCFLAGS = -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK
DEPFLAGS =	-M

LINK = $(CC)
LINKFLAGS = -O
SIZE = size

USRLIB = -lfftw

ARCHIVE =	ar
ARFLAGS =	-rc

#######################################################################
## Comment the following out or delete them:

# CC =		icc
# CCFLAGS =	-O
# SHFLAGS =	-fPIC
# DEPFLAGS =	-M

# LINK =		icc
# LINKFLAGS =	-O
# LIB =           -lstdc++
# SIZE =		size

# ARCHIVE =	ar
# ARFLAGS =	-rc
# SHLIBFLAGS =	-shared

I will say that, in some of my searches, I found reference to the use of USRLIB -lmpi. If this is included in the Makefile.linux, you will get the following error upon mpicc (MPICH2) compilation:

/usr/bin/ld: cannot find -lmpl
collect2: ld returned 1 exit status
make[1]: *** [../lmp_linux] Error 1
make[1]: Leaving directory `/opt/lammps-3Apr13/src/Obj_linux'
make: *** [linux] Error 2

So, don't include -lmpi.

With the above Makefile modifications, building of LAMMPS seems to go quite well until the following error is produced:

mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK  -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -c ../dump_dcd.cpp
mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK  -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -c ../dump_image.cpp
../dump_image.cpp:32:21: error: jpeglib.h: No such file or directory
../dump_image.cpp: In member function ‘virtual int LAMMPS_NS::DumpImage::modify_param(int, char**)':
../dump_image.cpp:904: warning: suggest parentheses around assignment used as truth value
../dump_image.cpp:911: warning: suggest parentheses around assignment used as truth value
../dump_image.cpp:964: warning: suggest parentheses around assignment used as truth value
../dump_image.cpp:971: warning: suggest parentheses around assignment used as truth value
../dump_image.cpp: In member function ‘void LAMMPS_NS::DumpImage::create_image()':
../dump_image.cpp:611: warning: ‘diameter' may be used uninitialized in this function
../dump_image.cpp:612: warning: ‘color' may be used uninitialized in this function
../dump_image.cpp:612: warning: ‘color1' may be used uninitialized in this function
../dump_image.cpp:612: warning: ‘color2' may be used uninitialized in this function
make[1]: *** [dump_image.o] Error 1
make[1]: Leaving directory `/opt/lammps-3Apr13/src/Obj_linux'
make: *** [linux] Error 2

The jpeglib.h error can likely be solved by installing only one of the libraries below, but I didn't bother to identify which one, instead indiscriminately installing a whole set based on the recommendation of dev.xonotic.org/projects/xonotic/wiki/Repository_Access (having found this link at forums.xonotic.org/showthread.php?tid=1252 – but I removed some of the install that I'm sure was not needed):

sudo apt-get install libxxf86dga-dev libxcb-xf86dri0-dev libxpm-dev libxxf86vm-dev libsdl1.2-dev libsdl-image1.2-dev libclalsadrv-dev libasound2-dev libxext-dev

Finally, a successful build!

make[1]: Entering directory `/opt/lammps-3Apr13/src/Obj_linux'
mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../write_restart.cpp > write_restart.d
mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../write_data.cpp > write_data.d
mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../verlet.cpp > verlet.d
mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../velocity.cpp > velocity.d
mpicxx -O -DFFT_FFTW -DLAMMPS_GZIP -DMPICH_IGNORE_CXX_SEEK -DLAMMPS_GZIP -DLAMMPS_JPEG  -DMPICH_SKIP_MPICXX  -DFFT_FFTW   -M ../variable.cpp > variable.d
...
write_data.o write_restart.o  -lmpich -lpthread -lfftw -ljpeg   -o ../lmp_linux
size ../lmp_linux
   text	   data	    bss	    dec	    hex	filename
3654398	   6928	    264	3661590	 37df16	../lmp_linux
make[1]: Leaving directory `/opt/lammps-3Apr13/src/Obj_linux'

At this point, you can move the lmp_linux executable anywhere (or, as I find myself doing, just keep it in the lammps-DATE/src folder for when you perform subsequent builds).

Running An MPI Calculations

If you've never run through an MPI calculation, you'll get the following errors when you first try to run lmp_linux:

cannot connect to local mpd (/tmp/mpd2.console_user); possible causes:
  1. no mpd is running on this host
  2. an mpd is running but was started without a "console" (-n option)
In case 1, you can start an mpd on this host with:
    mpd &
and you will be able to run jobs just on this host.
For more details on starting mpds on a set of hosts, see
the MPICH2 Installation Guide.

**********

user@machine:/opt/lammps-3Apr13/src$ mpd &

configuration file /home/damianallis/.mpd.conf not found
A file named .mpd.conf file must be present in the user's home
directory (/etc/mpd.conf if root) with read and write access
only for the user, and must contain at least a line with:
MPD_SECRETWORD=
One way to safely create this file is to do the following:
  cd $HOME
  touch .mpd.conf
  chmod 600 .mpd.conf
and then use an editor to insert a line like
  MPD_SECRETWORD=mr45-j9z
into the file.  (Of course use some other secret word than mr45-j9z.)

[1]+  Exit 255                mpd

Simply follow the instructions to put a .mdp.conf in your ~/ folder, start mpd with "mpd &," and run LAMMPS with the following command below (which uses 2 cores (-np 2)).

NOTE: My LAMMPS test (being brand new to it) was to run the first test available from icme.hpc.msstate.edu/mediawiki/index.php/LAMMPS_Help. Following the instructions therein (DOWNLOAD: Al99.eam.alloy (780 KB) and calc_fcc.in (1 KB)):

sudo mpirun -np 2 ./lmp_linux < calc_fcc.in

Which should produce the following output:

LAMMPS (3 Apr 2013)
Lattice spacing in x,y,z = 4 4 4
Created orthogonal box = (0 0 0) to (4 4 4)
  1 by 1 by 2 MPI processor grid
Lattice spacing in x,y,z = 4 4 4
Created 4 atoms
Replicating atoms ...
  orthogonal box = (0 0 0) to (4 4 4)
  1 by 1 by 2 MPI processor grid
  4 atoms
WARNING: Resetting reneighboring criteria during minimization (../min.cpp:173)
Setting up minimization ...
Memory usage per processor = 2.40372 Mbytes
Step PotEng Lx Ly Lz Press Pxx Pyy Pzz eatoms 
       0   -13.417787            4            4            4     29590.11     29590.11     29590.11     29590.11   -13.417787 
      10   -13.439104         4.04         4.04         4.04    5853.9553    5853.9553    5853.9553    5853.9553   -13.439104 
      14       -13.44         4.05         4.05         4.05     2.726913     2.726913     2.726913     2.726913       -13.44 
Loop time of 0.0287241 on 2 procs for 14 steps with 4 atoms

Minimization stats:
  Stopping criterion = linesearch alpha is zero
  Energy initial, next-to-last, final = 
        -13.4177872966     -13.4399999525     -13.4399999525
  Force two-norm initial, final = 3.54599 0.000335006
  Force max component initial, final = 3.54599 0.000335006
  Final line search alpha, max atom move = 0.0625 2.09379e-05
  Iterations, force evaluations = 14 19

Pair  time (%) = 0.00206506 (7.18931)
Neigh time (%) = 0 (0)
Comm  time (%) = 0.00481999 (16.7803)
Outpt time (%) = 0.000162005 (0.564006)
Other time (%) = 0.021677 (75.4664)

Nlocal:    2 ave 2 max 2 min
Histogram: 2 0 0 0 0 0 0 0 0 0
Nghost:    603 ave 603 max 603 min
Histogram: 2 0 0 0 0 0 0 0 0 0
Neighs:    140 ave 162 max 118 min
Histogram: 1 0 0 0 0 0 0 0 0 1

Total # of neighbors = 280
Ave neighs/atom = 70
Neighbor list builds = 0
Dangerous builds = 0
Total energy (eV) = -13.439999952539944061;
Number of atoms = 4;
Lattice constant (Angstoms) = 4.049999999999998046;
Cohesive energy (eV) = -3.3599999881349860154;
All done!
Posted on

Amber 11 And AmberTools 1.5 In Ubuntu 10.04 LTS (And Related, Including A How-To For EOL 8.10)

Having successfully navigated serial and parallel Amber10 installs under Ubuntu 8.10, I am pleased to report that the process for Amber11 with OpenMPI (from apt-get, one doesn't have to build from scratch) under Ubuntu 10.10 is seemingly much easier (and have it here so I don't forget). There is a bit of persnicketiness to the order of the serial and parallel installs that must be kept track of (and I'm building in serial-to-parallel order), but the process is otherwise straightforward.

For organizational purposes, I'm building amber11 in my $HOME directory. This removes some of the PATH issues with sudo-ing aspects of the install (and can be moved into another directory after the build is complete).

1. apt-get Installs

The search for dependent programs and libraries is a long and involved one given how many programs I have installed. Therefore, instead of trying to find all of the amber-dependent installs for successful building, I'm simply providing the list of everything I have on the test machine. As hard drives are cheap and Ubuntu will warn of conflicts, I recommend simply installing the below and accepting the 100 Mb hit to NOT have to find the smallest apt-get set (yes, some of these are obviously not needed).

sudo apt-get install build-essential cmake doxygen freeglut3-dev g++-multilib gcc-multilib gettext gnuplot ia32-libs lib32asound2 lib32gcc1 lib32gcc1-dbg lib32gfortran3 lib32gomp1 lib32mudflap0 lib32ncurses5 lib32nss-mdns lib32z1 libavdevice52 libc6-dev-i386 libc6-i386 libfreeimage-dev libglew1.5-dev libopenal1 libopenexr-dev libpng12-dev libqt4-dev libssl-dev libstdc++6-4.3-dbg libstdc++6-4.3-dev libstdc++6-4.3-doc libxi-dev libxml-simple-perl libxmu-dev mercurial nfs-common nfs-kernel-server portmap python2.6-dev rpm ssh

The above said, there are some obvious most-important installs that have to be there (according to the "official" Ubuntu amber11 install summary at ambermd.org/ubuntu.html). You could try to work with only these first if you were in a diagnostic mood today:

sudo apt-get install bison csh flex fort77 g++ gcc gfortran libbz2-dev libnetcdf-dev libopenmpi-dev libxext-dev libxt-dev openmpi-bin patch tcsh xorg-dev zlib1g-dev

With that, we move onto the AmberTools 1.5 install.

2. AmberTools 1.5 (Serial)

The AmberTools build process deals with PATH specifications for both it and Amber, then walks you through patching and a successful build.

user@machine:~$ tar xjf AmberTools-1.5.tar.bz2 
user@machine:~$ cd amber11/
user@machine:~/amber11$ echo "export AMBERHOME=$PWD" >> ~/.bashrc
user@machine:~/amber11$ echo "export PATH=$PATH:$AMBERHOME/bin" >> ~/.bashrc
user@machine:~/amber11$ source ~/.bashrc
user@machine:~/amber11$ wget http://ambermd.org/bugfixes/AmberTools/1.5/bugfix.all
user@machine:~/amber11$ patch -p0 < bugfix.all
user@machine:~/amber11$ rm bugfix.all
user@machine:~/amber11$ cd AmberTools/src/
user@machine:~/amber11/AmberTools/src$ ./configure gnu
user@machine:~/amber11/AmberTools/src$ make install
user@machine:~/amber11/AmberTools/src$ cd

3. Amber 11 (Serial Install)

For the Amber build, not building the serial version first will produce the following error (which you may or may not be searching against in google presently):

Warning: Deleted feature: PAUSE statement at (1)
cpp -traditional -P  -DBINTRAJ -DMPI    svbksb.f > _svbksb.f
mpif90 -c -O3 -mtune=generic -ffree-form   -o svbksb.o _svbksb.f
cpp -traditional -P  -DBINTRAJ -DMPI    pythag.f > _pythag.f
mpif90 -c -O3 -mtune=generic -ffree-form   -o pythag.o _pythag.f
Error: a serial version of libFpbsa.a must be built before parallel build.
make[2]: *** [libFpbsa.parallel] Error 2
make[2]: Leaving directory `/home/genomebio/amber11/AmberTools/src/pbsa'
make[1]: *** [libpbsa] Error 2
make[1]: Leaving directory `/home/genomebio/amber11/src/sander'
make: *** [parallel] Error 2

The "gnu" is also important, as there appears to be some kind of formatting (fortran-specific) issue with some files in the non-gnu build attempt that produces the following error if you just blindly run a ./configure:

Error: Unclassifiable statement at (1)
constants.f:39.1:

double precision, parameter :: two       = 2.0d0                        
 1
Error: Non-numeric character in statement label at (1)
constants.f:39.1:

double precision, parameter :: two       = 2.0d0                        
 1
Error: Unclassifiable statement at (1)
constants.f:40.1:

double precision, parameter :: three     = 3.0d0                        
 1
Error: Non-numeric character in statement label at (1)
Fatal Error: Error count reached limit of 25.
make[1]: *** [constants.o] Error 1
make[1]: Leaving directory `/home/user/amber11/src/sander'
make: *** [parallel] Error 2

With that, the serial build is below, including bug fixes.

user@machine:~$ tar xfj Amber11.tar.bz2
user@machine:~$ cd $AMBERHOME
user@machine:~/amber11$ wget http://ambermd.org/bugfixes/11.0/bugfix.all
user@machine:~/amber11$ wget http://ambermd.org/bugfixes/11.0/apply_bugfix.x
user@machine:~/amber11$ chmod +x ./apply_bugfix.x
user@machine:~/amber11$ ./apply_bugfix.x bugfix.all
user@machine:~/amber11$ cd AmberTools/src/
user@machine:~/amber11/AmberTools/src$ ./configure gnu
user@machine:~/amber11/AmberTools/src$ cd $AMBERHOME
user@machine:~/amber11$ ./AT15_Amber11.py 
user@machine:~/amber11$ cd src/
user@machine:~/amber11/src$ make serial

4. Amber 11 (Parallel)

Hopefully the serial build ran non-problematically. The parallel install works just as simply provided you run the process in the order below. The key steps are the "make clean," new ./configure, re-run of ./AT_Amber11.py, and the other "make clean."

user@machine:~/amber11/src$ cd $AMBERHOME
user@machine:~/amber11$ cd AmberTools/src/
user@machine:~/amber11/AmberTools/src$ make clean
user@machine:~/amber11/AmberTools/src$ ./configure -mpi gnu
user@machine:~/amber11/AmberTools/src$ cd $AMBERHOME
user@machine:~/amber11$ ./AT15_Amber11.py 
user@machine:~/amber11$ cd src/
user@machine:~/amber11/src$ make clean
user@machine:~/amber11/src$ make parallel

5. Amber 11 (Tests)

Finally, testing the install. Nothing specific to be done as far as the code is concerned, simply running the tests.

user@machine:~/amber11/src$ cd ..
user@machine:~/amber11$ cd test/
user@machine:~/amber11/test$ make -f Makefile
user@machine:~/amber11/test$

From the out-of-the-box installation above, my test results complete as follows:

365 file comparisons passed
15 file comparisons failed
0 tests experienced errors
Test log file saved as logs/test_amber_serial/2011-07-14_11-19-47.log
Test diffs file saved as logs/test_amber_serial/2011-07-14_11-19-47.diff

The failed tests include those already mentioned by the Amber developers to fail. This list is provided at the end of the AT15_Amber11.py results:

NOTE: Because PBSA has changed since Amber 11 was released, some
tests are known to fail and others are known to quit in error. These
can be safely ignored.

Tests that error: Tests in $AMBERHOME/test/sander_pbsa_frc
   Run.argasp.min    Run.dadt.min      Run.dgdc.min
   Run.lysasp.min    Run.polyALA.min   Run.polyAT.min
   Run.argasp.min    Run.dadt.min      Run.dgdc.min
   Run.lysasp.min    Run.polyALA.min   Run.polyAT.min
   Run.argasp.min    Run.dadt.min      Run.dgdc.min
   Run.lysasp.min    Run.polyALA.min   Run.polyAT.min

Tests that produce possible FAILUREs:
   cd sander_pbsa_ipb2   && ./Run.110D.min
   cd sander_pbsa_lpb    && ./Run.lsolver.min (only some of them fail here)
   cd sander_pbsa_tsr    && ./Run.tsrb.min
   cd sander_pbsa_decres && ./Run.pbsa_decres
   mm_pbsa.pl tests 02, 03, and 05

6. Quick Summary

For ease of copy-and-paste-ing, the command list is below:

apt-get

sudo apt-get install build-essential cmake doxygen freeglut3-dev g++-multilib gcc-multilib gettext gnuplot ia32-libs lib32asound2 lib32gcc1 lib32gcc1-dbg lib32gfortran3 lib32gomp1 lib32mudflap0 lib32ncurses5 lib32nss-mdns lib32z1 libavdevice52 libc6-dev-i386 libc6-i386 libfreeimage-dev libglew1.5-dev libopenal1 libopenexr-dev libpng12-dev libqt4-dev libssl-dev libstdc++6-4.3-dbg libstdc++6-4.3-dev libstdc++6-4.3-doc libxi-dev libxml-simple-perl libxmu-dev mercurial nfs-common nfs-kernel-server portmap python2.6-dev rpm ssh

sudo apt-get install bison csh flex fort77 g++ gcc gfortran libbz2-dev libnetcdf-dev libopenmpi-dev libxext-dev libxt-dev openmpi-bin patch tcsh xorg-dev zlib1g-dev

AmberTools

tar xjf AmberTools-1.5.tar.bz2 
cd amber11/
echo "export AMBERHOME=$PWD" >> ~/.bashrc
echo "export PATH=$PATH:$AMBERHOME/bin" >> ~/.bashrc
source ~/.bashrc
wget http://ambermd.org/bugfixes/AmberTools/1.5/bugfix.all
patch -p0 < bugfix.all
rm bugfix.all
cd AmberTools/src/
./configure gnu
make install
cd

Amber 11 (Serial)

tar xfj Amber11.tar.bz2
cd $AMBERHOME
wget http://ambermd.org/bugfixes/11.0/bugfix.all
wget http://ambermd.org/bugfixes/11.0/apply_bugfix.x
chmod +x ./apply_bugfix.x
./apply_bugfix.x bugfix.all
cd AmberTools/src/
./configure gnu
cd $AMBERHOME
./AT15_Amber11.py 
cd src/
make serial

Amber 11 (Parallel)

cd $AMBERHOME
cd AmberTools/src/
make clean
./configure -mpi gnu
cd $AMBERHOME
./AT15_Amber11.py 
cd src/
make clean
make parallel

Amber Tests

cd ..
cd test/
make -f Makefile

7. And Furthermore…

I tried the above on an old linux box running Intrepid Ibex (8.10), which counts as an End-Of-Life (Obsolete) version. Running all of the apt-get installs will work despite 8.10 not existing in the standard package locations, but you have to make the following addition to /etc/apt/sources.list.

sudo pico /etc/apt/sources.list

And copy-and-paste the following (this all taken from help.ubuntu.com/community/EOLUpgrades/Intrepid):

## EOL upgrade sources.list
# Required
deb http://old-releases.ubuntu.com/ubuntu/ intrepid main restricted universe multiverse
deb http://old-releases.ubuntu.com/ubuntu/ intrepid-updates main restricted universe multiverse
deb http://old-releases.ubuntu.com/ubuntu/ intrepid-security main restricted universe multiverse