import numpy as np
from abc import ABC, abstractmethod
from collections import defaultdict
from contextlib import contextmanager
from lbmpy.stencils import getStencil
from pystencils import Field
from pystencils.parallel.blockiteration import Block, SerialBlock
from pystencils.slicing import normalizeSlice, removeGhostLayers
from pystencils.utils import DotDict
try:
import pycuda.gpuarray as gpuarray
except ImportError:
gpuarray = None
class DataHandling(ABC):
"""
Manages the storage of arrays and maps them to a symbolic field.
Two versions are available: a simple, pure Python implementation for single node
simulations :py:class:SerialDataHandling and a distributed version using waLBerla in :py:class:ParallelDataHandling
Keep in mind that the data can be distributed, so use the 'access' method whenever possible and avoid the
'gather' function that has collects (parts of the) distributed data on a single process.
"""
def __init__(self):
self._preAccessFunctions = defaultdict(list)
self._postAccessFunctions = defaultdict(list)
# ---------------------------- Adding and accessing data -----------------------------------------------------------
@property
@abstractmethod
def dim(self):
"""Dimension of the domain, either 2 or 3"""
@abstractmethod
def addArray(self, name, fSize=1, dtype=np.float64, latexName=None, ghostLayers=None, layout=None, cpu=True, gpu=False):
"""
Adds a (possibly distributed) array to the handling that can be accessed using the given name.
For each array a symbolic field is available via the 'fields' dictionary
:param name: unique name that is used to access the field later
:param fSize: shape of the dim+1 coordinate. DataHandling supports zero or one index dimensions, i.e. scalar
fields and vector fields. This parameter gives the shape of the index dimensions. The default
value of 1 means no index dimension
:param dtype: data type of the array as numpy data type
:param latexName: optional, name of the symbolic field, if not given 'name' is used
:param ghostLayers: number of ghost layers - if not specified a default value specified in the constructor
is used
:param layout: memory layout of array, either structure of arrays 'SoA' or array of structures 'AoS'.
this is only important if fSize > 1
:param cpu: allocate field on the CPU
:param gpu: allocate field on the GPU
"""
@abstractmethod
def hasData(self, name):
"""
Returns true if a field or custom data element with this name was added
"""
@abstractmethod
def addArrayLike(self, name, nameOfTemplateField, latexName=None, cpu=True, gpu=False):
"""
Adds an array with the same parameters (number of ghost layers, fSize, dtype) as existing array
:param name: name of new array
:param nameOfTemplateField: name of array that is used as template
:param latexName: see 'add' method
:param cpu: see 'add' method
:param gpu: see 'add' method
"""
@abstractmethod
def addCustomData(self, name, cpuCreationFunction,
gpuCreationFunction=None, cpuToGpuTransferFunc=None, gpuToCpuTransferFunc=None):
"""
Adds custom (non-array) data to domain
:param name: name to access data
:param cpuCreationFunction: function returning a new instance of the data that should be stored
:param gpuCreationFunction: optional, function returning a new instance, stored on GPU
:param cpuToGpuTransferFunc: function that transfers cpu to gpu version, getting two parameters (gpuInstance, cpuInstance)
:param gpuToCpuTransferFunc: function that transfers gpu to cpu version, getting two parameters (gpuInstance, cpuInstance)
:return:
"""
@property
@abstractmethod
def fields(self):
"""Dictionary mapping data name to symbolic pystencils field - use this to create pystencils kernels"""
@abstractmethod
def ghostLayersOfField(self, name):
"""Returns the number of ghost layers for a specific field/array"""
@abstractmethod
def iterate(self, sliceObj=None, gpu=False, ghostLayers=None):
"""
Iterate over local part of potentially distributed data structure.
"""
@abstractmethod
def gatherArray(self, name, sliceObj=None, allGather=False):
"""
Gathers part of the domain on a local process. Whenever possible use 'access' instead, since this method copies
the distributed data to a single process which is inefficient and may exhaust the available memory
:param name: name of the array to gather
:param sliceObj: slice expression of the rectangular sub-part that should be gathered
:param allGather: if False only the root process receives the result, if True all processes
:return: generator expression yielding the gathered field, the gathered field does not include any ghost layers
"""
@abstractmethod
def runKernel(self, kernelFunc, *args, **kwargs):
"""
Runs a compiled pystencils kernel using the arrays stored in the DataHandling class for all array parameters
Additional passed arguments are directly passed to the kernel function and override possible parameters from
the DataHandling
"""
def registerPreAccessFunction(self, name, fct):
self._preAccessFunctions[name].append(fct)
def registerPostAccessFunction(self, name, fct):
self._postAccessFunctions[name].append(fct)
@contextmanager
def accessWrapper(self, name):
for func in self._preAccessFunctions[name]:
func()
yield
for func in self._postAccessFunctions[name]:
func()
# ------------------------------- CPU/GPU transfer -----------------------------------------------------------------
@abstractmethod
def toCpu(self, name):
"""Copies GPU data of array with specified name to CPU.
Works only if 'cpu=True' and 'gpu=True' has been used in 'add' method"""
pass
@abstractmethod
def toGpu(self, name):
"""Copies GPU data of array with specified name to GPU.
Works only if 'cpu=True' and 'gpu=True' has been used in 'add' method"""
pass
@abstractmethod
def allToCpu(self, name):
"""Copies data from GPU to CPU for all arrays that have a CPU and a GPU representation"""
pass
@abstractmethod
def allToGpu(self, name):
"""Copies data from CPU to GPU for all arrays that have a CPU and a GPU representation"""
pass
# ------------------------------- Communication --------------------------------------------------------------------
def synchronizationFunctionCPU(self, names, stencil=None, **kwargs):
"""
Synchronizes ghost layers for distributed arrays - for serial scenario this has to be called
for correct periodicity handling
:param names: what data to synchronize: name of array or sequence of names
:param stencil: stencil as string defining which neighbors are synchronized e.g. 'D2Q9', 'D3Q19'
if None, a full synchronization (i.e. D2Q9 or D3Q27) is done
:param kwargs: implementation specific, optional optimization parameters for communication
:return: function object to run the communication
"""
def synchronizationFunctionGPU(self, names, stencil=None, **kwargs):
"""
Synchronization of GPU fields, for documentation see CPU version above
"""
class SerialDataHandling(DataHandling):
class _PassThroughContextManager:
def __init__(self, arr):
self.arr = arr
def __enter__(self, *args, **kwargs):
return self.arr
def __init__(self, domainSize, defaultGhostLayers=1, defaultLayout='SoA', periodicity=False):
"""
Creates a data handling for single node simulations
:param domainSize: size of the spatial domain as tuple
:param defaultGhostLayers: nr of ghost layers used if not specified in add() method
:param defaultLayout: layout used if no layout is given to add() method
"""
super(SerialDataHandling, self).__init__()
self._domainSize = tuple(domainSize)
self.defaultGhostLayers = defaultGhostLayers
self.defaultLayout = defaultLayout
self._fields = DotDict()
self._fieldLatexNameToDataName = {}
self.cpuArrays = DotDict()
self.gpuArrays = DotDict()
self.customDataCpu = DotDict()
self.customDataGpu = DotDict()
self._customDataTransferFunctions = {}
if periodicity is None or periodicity is False:
periodicity = [False] * self.dim
if periodicity is True:
periodicity = [True] * self.dim
self._periodicity = periodicity
self._fieldInformation = {}
@property
def dim(self):
return len(self._domainSize)
@property
def fields(self):
return self._fields
def ghostLayersOfField(self, name):
return self._fieldInformation[name]['ghostLayers']
def addArray(self, name, fSize=1, dtype=np.float64, latexName=None, ghostLayers=None, layout=None,
cpu=True, gpu=False):
if ghostLayers is None:
ghostLayers = self.defaultGhostLayers
if layout is None:
layout = self.defaultLayout
if latexName is None:
latexName = name
assert layout in ('SoA', 'AoS')
kwargs = {
'shape': tuple(s + 2 * ghostLayers for s in self._domainSize),
'dtype': dtype,
'order': 'C' if layout == 'AoS' else 'F',
}
self._fieldInformation[name] = {
'ghostLayers': ghostLayers,
'fSize': fSize,
'layout': layout,
'dtype': dtype,
}
if fSize > 1:
kwargs['shape'] = kwargs['shape'] + (fSize,)
indexDimensions = 1
else:
indexDimensions = 0
if cpu:
if name in self.cpuArrays:
raise ValueError("CPU Field with this name already exists")
self.cpuArrays[name] = np.empty(**kwargs)
if gpu:
if name in self.gpuArrays:
raise ValueError("GPU Field with this name already exists")
self.gpuArrays[name] = gpuarray.empty(**kwargs)
assert all(f.name != latexName for f in self.fields.values()), "Symbolic field with this name already exists"
self.fields[name] = Field.createFixedSize(latexName, shape=kwargs['shape'], indexDimensions=indexDimensions,
dtype=kwargs['dtype'], layout=kwargs['order'])
self._fieldLatexNameToDataName[latexName] = name
def addCustomData(self, name, cpuCreationFunction,
gpuCreationFunction=None, cpuToGpuTransferFunc=None, gpuToCpuTransferFunc=None):
assert name not in self.cpuArrays
assert name not in self.customDataCpu
self.customDataCpu[name] = cpuCreationFunction()
if gpuCreationFunction:
self.customDataGpu[name] = gpuCreationFunction()
if cpuToGpuTransferFunc is None or gpuToCpuTransferFunc is None:
raise ValueError("For GPU data, both transfer functions have to be specified")
self._customDataTransferFunctions[name] = (cpuToGpuTransferFunc, gpuToCpuTransferFunc)
def hasData(self, name):
return name in self.fields
def addArrayLike(self, name, nameOfTemplateField, latexName=None, cpu=True, gpu=False):
self.addArray(name, latexName=latexName, cpu=cpu, gpu=gpu, **self._fieldInformation[nameOfTemplateField])
def iterate(self, sliceObj=None, gpu=False, ghostLayers=None):
if ghostLayers is None:
ghostLayers = self.defaultGhostLayers
if sliceObj is None:
sliceObj = (slice(None, None, None),) * self.dim
sliceObj = normalizeSlice(sliceObj, tuple(s + 2 * ghostLayers for s in self._domainSize))
sliceObj = tuple(s if type(s) is slice else slice(s, s+1, None) for s in sliceObj)
arrays = self.gpuArrays if gpu else self.cpuArrays
iterDict = {}
for name, arr in arrays.items():
fieldGls = self._fieldInformation[name]['ghostLayers']
if fieldGls < ghostLayers:
continue
arr = removeGhostLayers(arr, indexDimensions=len(arr.shape) - self.dim, ghostLayers=fieldGls-ghostLayers)
iterDict[name] = arr
offset = tuple(s.start - ghostLayers for s in sliceObj)
yield SerialBlock(iterDict, offset, sliceObj)
def gatherArray(self, name, sliceObj=None, **kwargs):
with self.accessWrapper(name):
gls = self._fieldInformation[name]['ghostLayers']
arr = self.cpuArrays[name]
arr = removeGhostLayers(arr, indexDimensions=self.fields[name].indexDimensions, ghostLayers=gls)
if sliceObj is not None:
arr = arr[sliceObj]
yield arr
def swap(self, name1, name2, gpu=False):
if not gpu:
self.cpuArrays[name1], self.cpuArrays[name2] = self.cpuArrays[name2], self.cpuArrays[name1]
else:
self.gpuArrays[name1], self.gpuArrays[name2] = self.gpuArrays[name2], self.gpuArrays[name1]
def allToCpu(self):
for name in (self.cpuArrays.keys() & self.gpuArrays.keys()) | self._customDataTransferFunctions.keys():
self.toCpu(name)
def allToGpu(self):
for name in (self.cpuArrays.keys() & self.gpuArrays.keys()) | self._customDataTransferFunctions.keys():
self.toGpu(name)
def runKernel(self, kernelFunc, *args, **kwargs):
dataUsedInKernel = [self._fieldLatexNameToDataName[p.fieldName]
for p in kernelFunc.ast.parameters if p.isFieldPtrArgument]
arrays = self.gpuArrays if kernelFunc.ast.backend == 'gpucuda' else self.cpuArrays
arrayParams = {name: arrays[name] for name in dataUsedInKernel}
arrayParams.update(kwargs)
kernelFunc(*args, **arrayParams)
def toCpu(self, name):
if name in self._customDataTransferFunctions:
transferFunc = self._customDataTransferFunctions[name][1]
transferFunc(self.customDataGpu[name], self.customDataCpu[name])
else:
self.gpuArrays[name].get(self.cpuArrays[name])
def toGpu(self, name):
if name in self._customDataTransferFunctions:
transferFunc = self._customDataTransferFunctions[name][0]
transferFunc(self.customDataGpu[name], self.customDataCpu[name])
else:
self.gpuArrays[name].set(self.cpuArrays[name])
def synchronizationFunctionCPU(self, names, stencilName=None, **kwargs):
return self._synchronizationFunctor(names, stencilName, 'cpu')
def synchronizationFunctionGPU(self, names, stencilName=None, **kwargs):
return self._synchronizationFunctor(names, stencilName, 'gpu')
def _synchronizationFunctor(self, names, stencil, target):
if stencil is None:
stencil = 'D3Q27' if self.dim == 3 else 'D2Q9'
assert stencil in ("D2Q9", 'D3Q27'), "Serial scenario support only D2Q9 or D3Q27 for periodicity sync"
assert target in ('cpu', 'gpu')
if not hasattr(names, '__len__') or type(names) is str:
names = [names]
filteredStencil = []
for direction in getStencil(stencil):
useDirection = True
if direction == (0, 0) or direction == (0, 0, 0):
useDirection = False
for component, periodicity in zip(direction, self._periodicity):
if not periodicity and component != 0:
useDirection = False
if useDirection:
filteredStencil.append(direction)
resultFunctors = []
for name in names:
gls = self._fieldInformation[name]['ghostLayers']
if len(filteredStencil) > 0:
if target == 'cpu':
from pystencils.slicing import getPeriodicBoundaryFunctor
resultFunctors.append(getPeriodicBoundaryFunctor(filteredStencil, ghostLayers=gls))
else:
from pystencils.gpucuda.periodicity import getPeriodicBoundaryFunctor
resultFunctors.append(getPeriodicBoundaryFunctor(filteredStencil, self._domainSize,
indexDimensions=self.fields[name].indexDimensions,
indexDimShape=self._fieldInformation[name]['fSize'],
dtype=self.fields[name].dtype.numpyDtype,
ghostLayers=gls))
if target == 'cpu':
def resultFunctor():
for func in resultFunctors:
func(pdfs=self.cpuArrays[name])
else:
def resultFunctor():
for func in resultFunctors:
func(pdfs=self.gpuArrays[name])
return resultFunctor