Getting started

Dependencies

GCC or different C-compiler
MPI implementation, i.e. OpenMPI or MPICH for MPI support
In order to make use of CUDA GPU acceleration compile with CUDA 11.x, for multi-GPU using a CUDA-aware MPI implementation

Compilation

Clone the directory

git clone https://github.com/claudiopica/HiRep

Adjust Make Flags

Adjust the file Make/MkFlags for the right compilation

Gauge group SU(NG) or SO(NG)

NG=3

#CHOICES ARE GAUGE_SUN AND GAUGE_SON
GAUGE_GROUP = GAUGE_SUN
#GAUGE_GROUP = GAUGE_SON

Representation

REPR = REPR_FUNDAMENTAL
#REPR = REPR_SYMMETRIC
#REPR = REPR_ANTISYMMETRIC
#REPR = REPR_ADJOINT

Boundary Conditions

Uncomment the line here, when you want to establish certain boundary conditions into the respective direction.

#T => PERIODIC, ANTIPERIODIC, OPEN, THETA
#X => PERIODIC, ANTIPERIODIC, THETA
#Y => PERIODIC, ANTIPERIODIC, THETA
#Z => PERIODIC, ANTIPERIODIC, THETA

Macro parameters

Then a number of macro parameters follow. Here you have to specify if you want to compile for certain boundary conditions by adding the identifier to the MACRO variable.

#MACRO += -DBC_T_THETA
#MACRO += -DBC_T_PERIODIC
MACRO += -DBC_T_ANTIPERIODIC
#MACRO += -DBC_T_OPEN
MACRO += -DBC_X_PERIODIC
MACRO += -DBC_Y_PERIODIC
MACRO += -DBC_Z_PERIODIC

#MACRO += -DBC_XYZ_TWISTED

#MACRO += -DHALFBG_SF
#MACRO += -DBASIC_SF
#MACRO += -DROTATED_SF

Specify, whether you want to compile with MPI either with or without GPU acceleration by using

#MACRO += -DWITH_MPI

For compilation with GPU acceleration for CUDA GPUs, add the identifier -DWITH_GPU to MACRO.

MACRO += -DWITH_GPU

Compilers, wrappers, preprocessors

A number of example combinations are already given in MkFlags.

For compiling with GCC and OpenMPI one would compile with

CC = gcc
MPICC = mpicc
LDFLAGS =
INCLUDE =

Using Intel compilers and Intel’s MPI implementation, one can use for example

CC = icc
MPICC = mpiicc
LDFLAGS = -L /opt/local/lib/mpich-devel-gcc7/ -L /opt/local/lib/
INCLUDE = -I /opt/local/include/mpich-devel-gcc7/

For CUDA acceleration, use nvcc and adjust the flag -arch according to the compute capability of the CUDA capable device.

CC = nvcc
CFLAGS = -O2 -Xcompiler '-std=c99 -fgcse-sm -fgcse-las -fgcse-after-reload'
GPUFLAGS = --x cu -arch=sm_70 -Xptxas -v -Xptxas -dlcm=ca -dc
LDFLAGS = -lcuda

For compilation with CUDA-aware MPI one needs to pass the MPI wrapper of the MPI implementation to the CUDA preprocessor using the flag -ccbin. For OpenMPI this is mpicc.

CC = nvcc
CFLAGS = -ccbin mpicc -Xcompiler '-std=c99 -fgcse-sm -fgcse-las -fgcse-after-reload'
GPUFLAGS = --x cu -arch=sm_70 -Xptxas -v -Xptxas -dlcm=ca -dc
LDFLAGS = -lcuda -lmpi

Run

Adjust input file

Compile the HiRep library for example in LibHR by typing make. An example of a C-file that generates a binary to run the HMC can be found in HMC, you can navigate into this directory and type make to create a binary. It is necessary to specify a number of parameters using an input file, see HMC/input_file for an example. For basic run variables, one can have a look at the section Run control variables.

run name = run1
save freq = 1
meas freq = 1
conf dir = cnfg
gauge start = random
last conf = +1

The “+” in front of last conf specifies the number of trajectories to be generated after the chosen startup configuration. I.e. if the startup configuration is trajectory number 5 and last conf = 6 then one additional trajectory will be generated. If last conf = +6 then six additional trajectories will be generated, up until trajectory number 11.

Execute Binary

Run the HMC using

$ ./hmc -i input_file

where hmc is the binary generated from hmc.c.