Benchmarking: Gromacs-4.6 - Revision history

Roshan: /* Intel Compilers */

2014-01-15T08:48:37Z

Intel Compilers

Roshan: /* Intel Compilers */

2014-01-15T08:47:29Z

Intel Compilers

Roshan: /* Run */

2014-01-15T08:41:08Z

Run

Roshan: /* Requirements */

2013-03-07T09:56:04Z

Requirements

Roshan: /* Requirements */

2013-02-22T06:53:02Z

Requirements

Roshan at 06:52, 22 February 2013

2013-02-22T06:52:15Z

Roshan at 06:43, 22 February 2013

2013-02-22T06:43:49Z

Roshan at 06:43, 22 February 2013

2013-02-22T06:43:20Z

Roshan: Created page with "Requirements cmake 2.8 Download Make module load cmake/2.8.10.2 module load cuda/5.0 cmake ../gromacs-4.6 -DGMX_GPU=ON -DGMX_BUILD_OWN_FFTW=ON \ -DCUDA_TOOLKIT_ROO..."

2013-02-22T06:27:03Z

Created page with "Requirements cmake 2.8 Download Make module load cmake/2.8.10.2 module load cuda/5.0 cmake ../gromacs-4.6 -DGMX_GPU=ON -DGMX_BUILD_OWN_FFTW=ON \ -DCUDA_TOOLKIT_ROO..."

New page

Requirements

cmake 2.8

Download

Make

module load cmake/2.8.10.2
module load cuda/5.0

cmake ../gromacs-4.6 -DGMX_GPU=ON -DGMX_BUILD_OWN_FFTW=ON \
-DCUDA_TOOLKIT_ROOT_DIR=/shared/apps/cuda/cuda-5.0 \
-DLIB_INSTALL_DIR=/shared/apps/cuda/gromacs/4.6/lib \
-DCMAKE_INSTALL_PREFIX=/home/roshan/apps/Gromacs-4.6

make -j 24

@@ Line 38: / Line 38: @@
 ===Intel Compilers===
 '''Error'''
-<syntaxhighlight>
+<pre>
 /home/boston/gromacs-4.6.3/include/gmx_math_x86_avx_128_fma_double.h(674): error: a value of type "int" cannot be assigned to an entity of type "__m128d"
        PolyCP0  = _mm_macc_pd(PolyCP1, w, PolyCP0);
@@ Line 47: / Line 47: @@
 make[1]: *** [src/gmxlib/CMakeFiles/gmx.dir/all] Error 2
 make: *** [all] Error 2
-</syntaxhighlight>
+</pre>
 '''Fix'''
 <pre>
@@ Line 62: / Line 62: @@
 $ make install
-It’s probably because Intel compilers do not understand the intrinsic types of AMD (AVX_128_FMA as identified by Gromacs) – meaning: the Intel compiler cannot produce optimized code for AMD.
+It’s probably because Intel compilers do not understand the intrinsic types of AMD (AVX_128_FMA as identified by Gromacs)
 </pre>

@@ Line 49: / Line 49: @@
 </syntaxhighlight>
 '''Fix'''
+<pre>
 You can fix the error by updating the CMakeCache.txt after you cmake Gromacs.
@@ Line 63: / Line 63: @@
 It’s probably because Intel compilers do not understand the intrinsic types of AMD (AVX_128_FMA as identified by Gromacs) – meaning: the Intel compiler cannot produce optimized code for AMD.

← Older revision		Revision as of 08:41, 15 January 2014
Line 34:		Line 34:
	module load gpu/gromacs/4.6		module load gpu/gromacs/4.6
	mdrun -s <input file>		mdrun -s <input file>
		+
		+	==Build Errors==
		+	===Intel Compilers===
		+	'''Error'''
		+	<syntaxhighlight>
		+	/home/boston/gromacs-4.6.3/include/gmx_math_x86_avx_128_fma_double.h(674): error: a value of type "int" cannot be assigned to an entity of type "__m128d"
		+	PolyCP0 = _mm_macc_pd(PolyCP1, w, PolyCP0);
		+	^
		+
		+	compilation aborted for /home/boston/gromacs-4.6.3/src/gmxlib/bondfree.c (code 4)
		+	make[2]: *** [src/gmxlib/CMakeFiles/gmx.dir/bondfree.c.o] Error 4
		+	make[1]: *** [src/gmxlib/CMakeFiles/gmx.dir/all] Error 2
		+	make: *** [all] Error 2
		+	</syntaxhighlight>
		+	'''Fix'''
		+
		+	You can fix the error by updating the CMakeCache.txt after you cmake Gromacs.
		+
		+	Update GMX_CPU_ACCELERATION:STRING=AVX_128_FMA to
		+
		+	//Accelerated CPU kernels. Pick one of: None, SSE2, SSE4.1, AVX_128_FMA,
		+	// AVX_256, IBM_QPX, Sparc64_HPC_ACE
		+	GMX_CPU_ACCELERATION:STRING=None
		+
		+	After which you can proceed with
		+	$ make
		+	$ make install
		+
		+	It’s probably because Intel compilers do not understand the intrinsic types of AMD (AVX_128_FMA as identified by Gromacs) – meaning: the Intel compiler cannot produce optimized code for AMD.

@@ Line 1: / Line 1: @@
 ==Requirements==
-cmake 2.8 (for building) <br>
+cmake 2.8 (for build) <br>
 CUDA 5.0

@@ Line 1: / Line 1: @@
 ==Requirements==
-# cmake 2.8
+cmake 2.8 (for building) <br>
 ==Download==

← Older revision		Revision as of 06:52, 22 February 2013
Line 18:		Line 18:

	make -j 24		make -j 24
		+
		+	==Modulefiles==
		+	File location: /shared/modulefiles <br>
		+	File name: gpu/gromacs/4.6
		+
		+	#%Module 1.0
		+	#
		+	prereq cuda/5.0
		+	system "source /shared/apps/cuda/gromacs/4.6/bin/GMXRC.bash"
		+	prepend-path PATH /shared/apps/cuda/gromacs/4.6/bin
		+	prepend-path LD_LIBRARY_PATH /shared/apps/cuda/gromacs/4.6/lib
		+
		+	==Run==
		+	module load gpu/gromacs/4.6
		+	mdrun -s <input file>

@@ Line 7: / Line 7: @@
   wget ftp://ftp.gromacs.org/pub/gromacs/gromacs-4.6.tar.gz
-==Make==
+==Build==
   module load cmake/2.8.10.2