phi4

See doc/ folder for description.

Build

Supports GCC 4.6, ICC 13.0 and Clang 3.0 or higher.

Requires Intel SVML library coming with ICC for SIMD builds using non-ICC compilers.

To build executable with default configuration use:

make

Building other configurations

Following build options may be combined. Changing build options changes final executable suffix indicating used options.

1. SIMD version via SIMD=<size>, i.e.:

    make SIMD=4
   

2. Enabling OpenMP:

    make OPENMP=1
   

3. Different float type via FLOAT=<type>, i.e.:

    make FLOAT=float
   

4. Different compiler via CC=<compiler>, i.e.:

    make CC=icc
   

5. Enabling cache optimization:

    make CACHE=1
   

6. Enabling index optimization: (implies only 2^n lattice sizes)

    make FAST_INDEXER=1
   

7. Enable SVML fix for ICC<13 when using non-ICC compilers and SIMD

    make SVML_FIX=1
   

8. Use old legacy code (used in publications) max AVX

    make LEGACY=1
   

Notes about source code

Source code in src/ is split into several logical modules and versions described by prefixes.

Modules

• types.h core configuration and type dispatch header

• swep_* files implementing single sweep

• random_* fused Tausworthe/LCG random number generator

• partition_* partitioning (satisfying Markov Chain constraints)

• indexer, *_index memory indexers implementation

• legacy/* legacy versions of modules above (used in publications)

Prefixes

• _simd SIMD version of module

• _test Gtest files (deprecated)

Operations

This is rough estimation of number operations in the algorithm.

NOTE: It skips lattice index calculation operations, memory operations, register transfers and variable initialization. So it estimates minimum number of operations, not average.

1. Corona gather (plus est. for Comp=1 Dim=3)

    Corona = Comp * ( Dim * (4 + 2*(Dim-1)) - Dim + 6 )
    	  ~=	1 * (	3 * (4 + 2*	  2	  ) -	3 + 6 )
    	   = 27
   

2. Fused Tausworthe + LCG random number generator

    Rng	   = Taus * 3 + LCG + 3
    	   = 6 * 3 + 6
    	   = 24
   

3. Initial action before updates (plus est. for Comp=1)

    Init   = Comp * 4 + 2
    	  ~= 6
   

4. Update ops for given Hits (plus est. for Comp=1 Dim=3)

    Update = Hits * ( (Comp+1) * Rng + (Comp-1) * 6 + Log + 10 )
    	  ~= Hits * (		2  * Rng +				  Log + 10 )
    	   = Hits * ( 58 + Log								   )
   

Gives altogether for Comp=1 Dim=3:

Total  = 33 + Hits * ( 58 + Log )

GPU Implementation

GPU implementation is available as separate project phi4_cuda.

License

Copyright (C) 2012-2013 Piotr Bialas, Adam Strzelecki

This file is part of phi4.

This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA