from collections import OrderedDict, namedtuple
from operator import attrgetter
from math import ceil
from cached_property import cached_property
import ctypes
from devito.arch import compiler_registry, platform_registry
from devito.data import default_allocator
from devito.exceptions import InvalidOperator
from devito.logger import debug, info, perf, warning, is_log_enabled_for
from devito.ir.equations import LoweredEq, lower_exprs
from devito.ir.clusters import ClusterGroup, clusterize
from devito.ir.iet import (Callable, CInterface, EntryFunction, FindSymbols, MetaCall,
derive_parameters, iet_build)
from devito.ir.support import AccessMode, SymbolRegistry
from devito.ir.stree import stree_build
from devito.operator.profiling import AdvancedProfilerVerbose, create_profile
from devito.operator.registry import operator_selector
from devito.mpi import MPI
from devito.parameters import configuration
from devito.passes import (Graph, lower_index_derivatives, generate_implicit,
generate_macros, minimize_symbols, unevaluate)
from devito.symbolics import estimate_cost
from devito.tools import (DAG, OrderedSet, Signer, ReducerMap, as_tuple, flatten,
filter_sorted, frozendict, is_integer, split, timed_pass,
timed_region, contains_val)
from devito.types import Grid, Evaluable
__all__ = ['Operator']
[docs]
class Operator(Callable):
"""
Generate, JIT-compile and run C code starting from an ordered sequence
of symbolic expressions.
Parameters
----------
expressions : expr-like or list or expr-like
The (list of) expression(s) defining the Operator computation.
**kwargs
* name : str
Name of the Operator, defaults to "Kernel".
* subs : dict
Symbolic substitutions to be applied to ``expressions``.
* opt : str
The performance optimization level. Defaults to ``configuration['opt']``.
* language : str
The target language for shared-memory parallelism. Defaults to
``configuration['language']``.
* platform : str
The architecture the code is generated for. Defaults to
``configuration['platform']``.
* compiler : str
The backend compiler used to jit-compile the generated code.
Defaults to ``configuration['compiler']``.
Examples
--------
The following Operator implements a trivial time-marching method that
adds 1 to every grid point in ``u`` at every timestep.
>>> from devito import Eq, Grid, TimeFunction, Operator
>>> grid = Grid(shape=(4, 4))
>>> u = TimeFunction(name='u', grid=grid)
>>> op = Operator(Eq(u.forward, u + 1))
Multiple expressions can be supplied, and there is no limit to the number of
expressions in an Operator.
>>> v = TimeFunction(name='v', grid=grid)
>>> op = Operator([Eq(u.forward, u + 1),
... Eq(v.forward, v + 1)])
Simple boundary conditions can be imposed easily exploiting the "indexed
notation" for Functions/TimeFunctions.
>>> t = grid.stepping_dim
>>> x, y = grid.dimensions
>>> op = Operator([Eq(u.forward, u + 1),
... Eq(u[t+1, x, 0], 0),
... Eq(u[t+1, x, 2], 0),
... Eq(u[t+1, 0, y], 0),
... Eq(u[t+1, 2, y], 0)])
A semantically equivalent computation can be expressed exploiting SubDomains.
>>> u.data[:] = 0
>>> op = Operator(Eq(u.forward, u + 1, subdomain=grid.interior))
By specifying a SubDomain, the Operator constrains the execution of an expression to
a certain sub-region within the computational domain. Ad-hoc SubDomains can also be
created in application code -- refer to the SubDomain documentation for more info.
Advanced boundary conditions can be expressed leveraging `SubDomain` and
`SubDimension`.
Tensor contractions are supported, but with one caveat: in case of MPI execution, any
global reductions along an MPI-distributed Dimension should be handled explicitly in
user code. The following example shows how to implement the matrix-vector
multiplication ``Av = b`` (inducing a reduction along ``y``).
>>> from devito import Inc, Function
>>> A = Function(name='A', grid=grid)
>>> v = Function(name='v', shape=(3,), dimensions=(y,))
>>> b = Function(name='b', shape=(3,), dimensions=(x,))
>>> op = Operator(Inc(b, A*v))
Dense and sparse computation may be present within the same Operator. In the
following example, interpolation is used to approximate the value of four
sparse points placed at the center of the four quadrants at the grid corners.
>>> import numpy as np
>>> from devito import SparseFunction
>>> grid = Grid(shape=(4, 4), extent=(3.0, 3.0))
>>> f = Function(name='f', grid=grid)
>>> coordinates = np.array([(0.5, 0.5), (0.5, 2.5), (2.5, 0.5), (2.5, 2.5)])
>>> sf = SparseFunction(name='sf', grid=grid, npoint=4, coordinates=coordinates)
>>> op = Operator([Eq(f, f + 1)] + sf.interpolate(f))
The iteration direction is automatically detected by the Devito compiler. Below,
the Operator runs from ``time_M`` (maximum point in the time dimension) down to
``time_m`` (minimum point in the time dimension), as opposed to all of the examples
seen so far, in which the execution along time proceeds from ``time_m`` to ``time_M``
through unit-step increments.
>>> op = Operator(Eq(u.backward, u + 1))
Loop-level optimisations, including SIMD vectorisation and OpenMP parallelism, are
automatically discovered and handled by the Devito compiler. For more information,
refer to the relevant documentation.
"""
_default_headers = [('_POSIX_C_SOURCE', '200809L')]
_default_includes = ['stdlib.h', 'math.h', 'sys/time.h']
_default_globals = []
def __new__(cls, expressions, **kwargs):
if expressions is None:
# Return a dummy Callable. This is exploited by unpickling. Users
# can't do anything useful with it
return super(Operator, cls).__new__(cls, **kwargs)
# Parse input arguments
kwargs = parse_kwargs(**kwargs)
# The Operator type for the given target
cls = operator_selector(**kwargs)
# Normalize input arguments for the selected Operator
kwargs = cls._normalize_kwargs(**kwargs)
cls._check_kwargs(**kwargs)
# Lower to a JIT-compilable object
with timed_region('op-compile') as r:
op = cls._build(expressions, **kwargs)
op._profiler.py_timers.update(r.timings)
# Emit info about how long it took to perform the lowering
op._emit_build_profiling()
return op
@classmethod
def _normalize_kwargs(cls, **kwargs):
return kwargs
@classmethod
def _check_kwargs(cls, **kwargs):
return
@classmethod
def _build(cls, expressions, **kwargs):
# Python- (i.e., compile-) and C-level (i.e., run-time) performance
profiler = create_profile('timers')
# Lower the input expressions into an IET
irs, byproduct = cls._lower(expressions, profiler=profiler, **kwargs)
# Make it an actual Operator
op = Callable.__new__(cls, **irs.iet.args)
Callable.__init__(op, **op.args)
# Header files, etc.
op._headers = OrderedSet(*cls._default_headers)
op._headers.update(byproduct.headers)
op._globals = OrderedSet(*cls._default_globals)
op._globals.update(byproduct.globals)
op._includes = OrderedSet(*cls._default_includes)
op._includes.update(profiler._default_includes)
op._includes.update(byproduct.includes)
# Required for the jit-compilation
op._compiler = kwargs['compiler']
op._language = kwargs['language']
op._lib = None
op._cfunction = None
# Potentially required for lazily allocated Functions
op._mode = kwargs['mode']
op._options = kwargs['options']
op._allocator = kwargs['allocator']
op._platform = kwargs['platform']
# References to local or external routines
op._func_table = OrderedDict()
op._func_table.update(OrderedDict([(i, MetaCall(None, False))
for i in profiler._ext_calls]))
op._func_table.update(OrderedDict([(i.root.name, i) for i in byproduct.funcs]))
# Internal mutable state to store information about previous runs, autotuning
# reports, etc
op._state = cls._initialize_state(**kwargs)
# Produced by the various compilation passes
op._reads = filter_sorted(flatten(e.reads for e in irs.expressions))
op._writes = filter_sorted(flatten(e.writes for e in irs.expressions))
op._dimensions = set().union(*[e.dimensions for e in irs.expressions])
op._dtype, op._dspace = irs.clusters.meta
op._profiler = profiler
return op
def __init__(self, *args, **kwargs):
# Bypass the silent call to __init__ triggered through the backends engine
pass
# Compilation -- Expression level
@classmethod
def _lower(cls, expressions, **kwargs):
"""
Perform the lowering Expressions -> Clusters -> ScheduleTree -> IET.
"""
# Create a symbol registry
kwargs.setdefault('sregistry', SymbolRegistry())
expressions = as_tuple(expressions)
# Input check
if any(not isinstance(i, Evaluable) for i in expressions):
raise InvalidOperator("Only `devito.Evaluable` are allowed.")
# Enable recursive lowering
# This may be used by a compilation pass that constructs a new
# expression for which a partial or complete lowering is desired
kwargs['rcompile'] = cls._rcompile_wrapper(**kwargs)
# [Eq] -> [LoweredEq]
expressions = cls._lower_exprs(expressions, **kwargs)
# [LoweredEq] -> [Clusters]
clusters = cls._lower_clusters(expressions, **kwargs)
# [Clusters] -> ScheduleTree
stree = cls._lower_stree(clusters, **kwargs)
# ScheduleTree -> unbounded IET
uiet = cls._lower_uiet(stree, **kwargs)
# unbounded IET -> IET
iet, byproduct = cls._lower_iet(uiet, **kwargs)
return IRs(expressions, clusters, stree, uiet, iet), byproduct
@classmethod
def _rcompile_wrapper(cls, **kwargs):
def wrapper(expressions, kwargs=kwargs):
return rcompile(expressions, kwargs)
return wrapper
@classmethod
def _initialize_state(cls, **kwargs):
return {}
@classmethod
def _specialize_dsl(cls, expressions, **kwargs):
"""
Backend hook for specialization at the DSL level. The input is made of
expressions and other higher order objects such as Injection or
Interpolation; the expressions are still unevaluated at this stage,
meaning that they are still in tensorial form and derivatives aren't
expanded yet.
"""
return expressions
@classmethod
def _specialize_exprs(cls, expressions, **kwargs):
"""
Backend hook for specialization at the expression level.
"""
return expressions
@classmethod
@timed_pass(name='lowering.Expressions')
def _lower_exprs(cls, expressions, **kwargs):
"""
Expression lowering:
* Apply rewrite rules;
* Evaluate derivatives;
* Flatten vectorial equations;
* Indexify Functions;
* Apply substitution rules;
* Shift indices for domain alignment.
"""
expand = kwargs['options'].get('expand', True)
# Specialization is performed on unevaluated expressions
expressions = cls._specialize_dsl(expressions, **kwargs)
# Lower functional DSL
expressions = flatten([i._evaluate(expand=expand) for i in expressions])
expressions = [j for i in expressions for j in i._flatten]
# A second round of specialization is performed on evaluated expressions
expressions = cls._specialize_exprs(expressions, **kwargs)
# "True" lowering (indexification, shifting, ...)
expressions = lower_exprs(expressions, **kwargs)
processed = [LoweredEq(i) for i in expressions]
return processed
# Compilation -- Cluster level
@classmethod
def _specialize_clusters(cls, clusters, **kwargs):
"""
Backend hook for specialization at the Cluster level.
"""
return clusters
@classmethod
@timed_pass(name='lowering.Clusters')
def _lower_clusters(cls, expressions, profiler=None, **kwargs):
"""
Clusters lowering:
* Group expressions into Clusters;
* Introduce guards for conditional Clusters;
* Analyze Clusters to detect computational properties such
as parallelism.
* Optimize Clusters for performance
"""
sregistry = kwargs['sregistry']
# Build a sequence of Clusters from a sequence of Eqs
clusters = clusterize(expressions, **kwargs)
# Operation count before specialization
init_ops = sum(estimate_cost(c.exprs) for c in clusters if c.is_dense)
clusters = cls._specialize_clusters(clusters, **kwargs)
# Operation count after specialization
final_ops = sum(estimate_cost(c.exprs) for c in clusters if c.is_dense)
try:
profiler.record_ops_variation(init_ops, final_ops)
except AttributeError:
pass
# Generate implicit Clusters from higher level abstractions
clusters = generate_implicit(clusters, sregistry=sregistry)
# Lower all remaining high order symbolic objects
clusters = lower_index_derivatives(clusters, **kwargs)
# Make sure no reconstructions can unpick any of the symbolic
# optimizations performed so far
clusters = unevaluate(clusters)
return ClusterGroup(clusters)
# Compilation -- ScheduleTree level
@classmethod
def _specialize_stree(cls, stree, **kwargs):
"""
Backend hook for specialization at the Schedule tree level.
"""
return stree
@classmethod
@timed_pass(name='lowering.ScheduleTree')
def _lower_stree(cls, clusters, **kwargs):
"""
Schedule tree lowering:
* Turn a sequence of Clusters into a ScheduleTree;
* Derive and attach metadata for distributed-memory parallelism;
* Derive sections for performance profiling
"""
# Build a ScheduleTree from a sequence of Clusters
stree = stree_build(clusters, **kwargs)
stree = cls._specialize_stree(stree)
return stree
# Compilation -- Iteration/Expression tree level
@classmethod
def _specialize_iet(cls, graph, **kwargs):
"""
Backend hook for specialization at the Iteration/Expression tree level.
"""
return graph
@classmethod
@timed_pass(name='lowering.uIET')
def _lower_uiet(cls, stree, profiler=None, **kwargs):
"""
Turn a ScheduleTree into an unbounded Iteration/Expression tree, that is
in essence a "floating" IET where one or more variables may be unbounded
(i.e., no definition placed yet).
"""
# Build an unbounded IET from a ScheduleTree
uiet = iet_build(stree)
# Analyze the IET Sections for C-level profiling
try:
profiler.analyze(uiet)
except AttributeError:
pass
return uiet
@classmethod
@timed_pass(name='lowering.IET')
def _lower_iet(cls, uiet, profiler=None, **kwargs):
"""
Iteration/Expression tree lowering:
* Introduce distributed-memory, shared-memory, and SIMD parallelism;
* Introduce optimizations for data locality;
* Finalize (e.g., symbol definitions, array casts)
"""
name = kwargs.get("name", "Kernel")
sregistry = kwargs['sregistry']
# Wrap the IET with an EntryFunction (a special Callable representing
# the entry point of the generated library)
parameters = derive_parameters(uiet, True)
iet = EntryFunction(name, uiet, 'int', parameters, ())
# Lower IET to a target-specific IET
graph = Graph(iet, sregistry=sregistry)
graph = cls._specialize_iet(graph, **kwargs)
# Instrument the IET for C-level profiling
# Note: this is postponed until after _specialize_iet because during
# specialization further Sections may be introduced
cls._Target.instrument(graph, profiler=profiler, **kwargs)
# Extract the necessary macros from the symbolic objects
generate_macros(graph)
# Target-independent optimizations
minimize_symbols(graph)
return graph.root, graph
# Read-only properties exposed to the outside world
@cached_property
def reads(self):
return tuple(self._reads)
@cached_property
def writes(self):
return tuple(self._writes)
@cached_property
def dimensions(self):
ret = set().union(*[d._defines for d in self._dimensions])
# During compilation other Dimensions may have been produced
dimensions = FindSymbols('dimensions').visit(self)
ret.update(d for d in dimensions if d.is_PerfKnob)
ret = tuple(sorted(ret, key=attrgetter('name')))
return ret
@cached_property
def input(self):
return tuple(i for i in self.parameters if i.is_Input)
@cached_property
def temporaries(self):
return tuple(i for i in self.parameters if i.is_TempFunction)
@cached_property
def objects(self):
return tuple(i for i in self.parameters if i.is_Object)
# Arguments processing
@cached_property
def _access_modes(self):
"""
A table providing the AccessMode of all user-accessible symbols in `self`.
"""
return frozendict({i: AccessMode(i in self.reads, i in self.writes)
for i in self.input})
def _prepare_arguments(self, autotune=None, **kwargs):
"""
Process runtime arguments passed to ``.apply()` and derive
default values for any remaining arguments.
"""
# Sanity check -- all user-provided keywords must be known to the Operator
if not configuration['ignore-unknowns']:
for k, v in kwargs.items():
if k not in self._known_arguments:
raise ValueError("Unrecognized argument %s=%s" % (k, v))
# Pre-process Dimension overrides. This may help ruling out ambiguities
# when processing the `defaults` arguments. A topological sorting is used
# as DerivedDimensions may depend on their parents
nodes = self.dimensions
edges = [(i, i.parent) for i in self.dimensions
if i.is_Derived and i.parent in set(nodes)]
toposort = DAG(nodes, edges).topological_sort()
futures = {}
for d in reversed(toposort):
if set(d._arg_names).intersection(kwargs):
futures.update(d._arg_values(self._dspace[d], args={}, **kwargs))
overrides, defaults = split(self.input, lambda p: p.name in kwargs)
# Process data-carrier overrides
args = kwargs['args'] = ReducerMap()
for p in overrides:
args.update(p._arg_values(**kwargs))
try:
args.reduce_inplace()
except ValueError:
raise ValueError("Override `%s` is incompatible with overrides `%s`" %
(p, [i for i in overrides if i.name in args]))
# Process data-carrier defaults
for p in defaults:
if p.name in args:
# E.g., SubFunctions
continue
for k, v in p._arg_values(**kwargs).items():
if k not in args:
args[k] = v
elif k in futures:
# An explicit override is later going to set `args[k]`
pass
elif k in kwargs:
# User is in control
# E.g., given a ConditionalDimension `t_sub` with factor `fact` and
# a TimeFunction `usave(t_sub, x, y)`, an override for `fact` is
# supplied w/o overriding `usave`; that's legal
pass
elif is_integer(args[k]) and not contains_val(args[k], v):
raise ValueError("Default `%s` is incompatible with other args as "
"`%s=%s`, while `%s=%s` is expected. Perhaps you "
"forgot to override `%s`?" %
(p, k, v, k, args[k], p))
args = kwargs['args'] = args.reduce_all()
# DiscreteFunctions may be created from CartesianDiscretizations, which in
# turn could be Grids or SubDomains. Both may provide arguments
discretizations = {getattr(kwargs[p.name], 'grid', None) for p in overrides}
discretizations.update({getattr(p, 'grid', None) for p in defaults})
discretizations.discard(None)
# Remove subgrids if multiple grids
if len(discretizations) > 1:
discretizations = {g for g in discretizations
if not any(d.is_Derived for d in g.dimensions)}
for i in discretizations:
args.update(i._arg_values(**kwargs))
# There can only be one Grid from which DiscreteFunctions were created
grids = {i for i in discretizations if isinstance(i, Grid)}
if len(grids) > 1:
# We loosely tolerate multiple Grids for backwards compatibility
# with spacial subsampling, which should be revisited however. And
# With MPI it would definitely break!
if configuration['mpi']:
raise ValueError("Multiple Grids found")
try:
grid = grids.pop()
except KeyError:
grid = None
# An ArgumentsMap carries additional metadata that may be used by
# the subsequent phases of the arguments processing
args = kwargs['args'] = ArgumentsMap(args, grid, self)
# Process Dimensions
for d in reversed(toposort):
args.update(d._arg_values(self._dspace[d], grid, **kwargs))
# Process Objects
for o in self.objects:
args.update(o._arg_values(grid=grid, **kwargs))
# In some "lower-level" Operators implementing a random piece of C, such as
# one or more calls to third-party library functions, there could still be
# at this point unprocessed arguments (e.g., scalars)
kwargs.pop('args')
args.update({k: v for k, v in kwargs.items() if k not in args})
# Sanity check
for p in self.parameters:
p._arg_check(args, self._dspace[p], am=self._access_modes.get(p))
for d in self.dimensions:
if d.is_Derived:
d._arg_check(args, self._dspace[p])
# Turn arguments into a format suitable for the generated code
# E.g., instead of NumPy arrays for Functions, the generated code expects
# pointers to ctypes.Struct
for p in self.parameters:
try:
args.update(kwargs.get(p.name, p)._arg_finalize(args, alias=p))
except AttributeError:
# User-provided floats/ndarray obviously do not have `_arg_finalize`
args.update(p._arg_finalize(args, alias=p))
# Execute autotuning and adjust arguments accordingly
args.update(self._autotune(args, autotune or configuration['autotuning']))
return args
def _postprocess_arguments(self, args, **kwargs):
"""Process runtime arguments upon returning from ``.apply()``."""
for p in self.parameters:
try:
subfuncs = (args[getattr(p, s).name] for s in p._sub_functions)
p._arg_apply(args[p.name], *subfuncs, alias=kwargs.get(p.name))
except AttributeError:
p._arg_apply(args[p.name], alias=kwargs.get(p.name))
@cached_property
def _known_arguments(self):
"""The arguments that can be passed to ``apply`` when running the Operator."""
ret = set()
for i in self.input:
ret.update(i._arg_names)
try:
ret.update(i.grid._arg_names)
except AttributeError:
pass
for d in self.dimensions:
ret.update(d._arg_names)
ret.update(p.name for p in self.parameters)
return frozenset(ret)
def _autotune(self, args, setup):
"""Auto-tuning to improve runtime performance."""
return args
[docs]
def arguments(self, **kwargs):
"""Arguments to run the Operator."""
args = self._prepare_arguments(**kwargs)
# Check all arguments are present
for p in self.parameters:
if args.get(p.name) is None:
raise ValueError("No value found for parameter %s" % p.name)
return args
# Code generation and JIT compilation
@cached_property
def _soname(self):
"""A unique name for the shared object resulting from JIT compilation."""
return Signer._digest(self, configuration)
@cached_property
def ccode(self):
try:
return self._ccode_handler(compiler=self._compiler).visit(self)
except (AttributeError, TypeError):
from devito.ir.iet.visitors import CGen
return CGen(compiler=self._compiler).visit(self)
def _jit_compile(self):
"""
JIT-compile the C code generated by the Operator.
It is ensured that JIT compilation will only be performed once per
Operator, reagardless of how many times this method is invoked.
"""
if self._lib is None:
with self._profiler.timer_on('jit-compile'):
recompiled, src_file = self._compiler.jit_compile(self._soname,
str(self.ccode))
elapsed = self._profiler.py_timers['jit-compile']
if recompiled:
perf("Operator `%s` jit-compiled `%s` in %.2f s with `%s`" %
(self.name, src_file, elapsed, self._compiler))
else:
perf("Operator `%s` fetched `%s` in %.2f s from jit-cache" %
(self.name, src_file, elapsed))
@property
def cfunction(self):
"""The JIT-compiled C function as a ctypes.FuncPtr object."""
if self._lib is None:
self._jit_compile()
self._lib = self._compiler.load(self._soname)
self._lib.name = self._soname
if self._cfunction is None:
self._cfunction = getattr(self._lib, self.name)
# Associate a C type to each argument for runtime type check
self._cfunction.argtypes = [i._C_ctype for i in self.parameters]
return self._cfunction
[docs]
def cinterface(self, force=False):
"""
Generate two files under the prescribed temporary directory:
* `X.c` (or `X.cpp`): the code generated for this Operator;
* `X.h`: an header file representing the interface of `X.c`.
Where `X=self.name`.
Parameters
----------
force : bool, optional
Overwrite any existing files. Defaults to False.
"""
dest = self._compiler.get_jit_dir()
name = dest.joinpath(self.name)
cfile = name.with_suffix(".%s" % self._compiler.src_ext)
hfile = name.with_suffix('.h')
# Generate the .c and .h code
ccode, hcode = CInterface().visit(self)
for f, code in [(cfile, ccode), (hfile, hcode)]:
if not force and f.is_file():
debug("`%s` was not saved in `%s` as it already exists" % (f.name, dest))
else:
with open(str(f), 'w') as ff:
ff.write(str(code))
debug("`%s` successfully saved in `%s`" % (f.name, dest))
return ccode, hcode
# Execution
def __call__(self, **kwargs):
return self.apply(**kwargs)
[docs]
def apply(self, **kwargs):
"""
Execute the Operator.
With no arguments provided, the Operator runs using the data carried by the
objects appearing in the input expressions -- these are referred to as the
"default arguments".
Optionally, any of the Operator default arguments may be replaced by passing
suitable key-value arguments. Given ``apply(k=v, ...)``, ``(k, v)`` may be
used to:
* replace a Constant. In this case, ``k`` is the name of the Constant,
``v`` is either a Constant or a scalar value.
* replace a Function (SparseFunction). Here, ``k`` is the name of the
Function, ``v`` is either a Function or a numpy.ndarray.
* alter the iteration interval along a Dimension. Consider a generic
Dimension ``d`` iterated over by the Operator. By default, the Operator
runs over all iterations within the compact interval ``[d_m, d_M]``,
where ``d_m`` and ``d_M`` are, respectively, the smallest and largest
integers not causing out-of-bounds memory accesses (for the Grid
Dimensions, this typically implies iterating over the entire physical
domain). So now ``k`` can be either ``d_m`` or ``d_M``, while ``v``
is an integer value.
Examples
--------
Consider the following Operator
>>> from devito import Eq, Grid, TimeFunction, Operator
>>> grid = Grid(shape=(3, 3))
>>> u = TimeFunction(name='u', grid=grid, save=3)
>>> op = Operator(Eq(u.forward, u + 1))
The Operator is run by calling ``apply``
>>> summary = op.apply()
The variable ``summary`` contains information about runtime performance.
As no key-value parameters are specified, the Operator runs with its
default arguments, namely ``u=u, x_m=0, x_M=2, y_m=0, y_M=2, time_m=0,
time_M=1``.
At this point, the same Operator can be used for a completely different
run, for example
>>> u2 = TimeFunction(name='u', grid=grid, save=5)
>>> summary = op.apply(u=u2, x_m=1, y_M=1)
Now, the Operator will run with a different set of arguments, namely
``u=u2, x_m=1, x_M=2, y_m=0, y_M=1, time_m=0, time_M=3``.
To run an Operator that only uses buffered TimeFunctions, the maximum
iteration point along the time dimension must be explicitly specified
(otherwise, the Operator wouldn't know how many iterations to run).
>>> u3 = TimeFunction(name='u', grid=grid)
>>> op = Operator(Eq(u3.forward, u3 + 1))
>>> summary = op.apply(time_M=10)
"""
# Build the arguments list to invoke the kernel function
with self._profiler.timer_on('arguments'):
args = self.arguments(**kwargs)
# Invoke kernel function with args
arg_values = [args[p.name] for p in self.parameters]
try:
cfunction = self.cfunction
with self._profiler.timer_on('apply', comm=args.comm):
cfunction(*arg_values)
except ctypes.ArgumentError as e:
if e.args[0].startswith("argument "):
argnum = int(e.args[0][9:].split(':')[0]) - 1
newmsg = "error in argument '%s' with value '%s': %s" % (
self.parameters[argnum].name,
arg_values[argnum],
e.args[0])
raise ctypes.ArgumentError(newmsg) from e
else:
raise
# Post-process runtime arguments
self._postprocess_arguments(args, **kwargs)
# Output summary of performance achieved
return self._emit_apply_profiling(args)
# Performance profiling
def _emit_build_profiling(self):
if not is_log_enabled_for('PERF'):
return
# Rounder to K decimal places
fround = lambda i, n=100: ceil(i * n) / n
timings = self._profiler.py_timers.copy()
tot = timings.pop('op-compile')
perf("Operator `%s` generated in %.2f s" % (self.name, fround(tot)))
max_hotspots = 3
threshold = 20.
def _emit_timings(timings, indent=''):
timings.pop('total', None)
entries = sorted(timings, key=lambda i: timings[i]['total'], reverse=True)
for i in entries[:max_hotspots]:
v = fround(timings[i]['total'])
perc = fround(v/tot*100, n=10)
if perc > threshold:
perf("%s%s: %.2f s (%.1f %%)" % (indent, i.lstrip('_'), v, perc))
_emit_timings(timings[i], ' '*len(indent) + ' * ')
_emit_timings(timings, ' * ')
if self._profiler._ops:
ops = ['%d --> %d' % i for i in self._profiler._ops]
perf("Flops reduction after symbolic optimization: [%s]" % ' ; '.join(ops))
def _emit_apply_profiling(self, args):
"""Produce a performance summary of the profiled sections."""
# Rounder to 2 decimal places
fround = lambda i: ceil(i * 100) / 100
elapsed = fround(self._profiler.py_timers['apply'])
info("Operator `%s` ran in %.2f s" % (self.name, elapsed))
summary = self._profiler.summary(args, self._dtype, reduce_over=elapsed)
if not is_log_enabled_for('PERF'):
# Do not waste time
return summary
if summary.globals:
# Note that with MPI enabled, the global performance indicators
# represent "cross-rank" performance data
metrics = []
v = summary.globals.get('vanilla')
if v is not None:
metrics.append("OI=%.2f" % fround(v.oi))
metrics.append("%.2f GFlops/s" % fround(v.gflopss))
v = summary.globals.get('fdlike')
if v is not None:
metrics.append("%.2f GPts/s" % fround(v.gpointss))
if metrics:
perf("Global performance: [%s]" % ', '.join(metrics))
perf("Local performance:")
indent = " "*2
else:
indent = ""
if isinstance(self._profiler, AdvancedProfilerVerbose):
metrics = []
v = summary.globals.get('fdlike-nosetup')
if v is not None:
metrics.append("%.2f GPts/s" % fround(v.gpointss))
if metrics:
perf("Global performance <w/o setup>: [%s]" % ', '.join(metrics))
# Emit local, i.e. "per-rank" performance. Without MPI, this is the only
# thing that will be emitted
for k, v in summary.items():
rank = "[rank%d]" % k.rank if k.rank is not None else ""
if v.gflopss:
oi = "OI=%.2f" % fround(v.oi)
gflopss = "%.2f GFlops/s" % fround(v.gflopss)
gpointss = "%.2f GPts/s" % fround(v.gpointss)
metrics = "[%s]" % ", ".join([oi, gflopss, gpointss])
else:
metrics = ""
itershapes = [",".join(str(i) for i in its) for its in v.itershapes]
if len(itershapes) > 1:
itershapes = ",".join("<%s>" % i for i in itershapes)
elif len(itershapes) == 1:
itershapes = itershapes[0]
else:
itershapes = ""
name = "%s%s<%s>" % (k.name, rank, itershapes)
perf("%s* %s ran in %.2f s %s" % (indent, name, fround(v.time), metrics))
for n, time in summary.subsections.get(k.name, {}).items():
perf("%s+ %s ran in %.2f s [%.2f%%]" %
(indent*2, n, time, fround(time/v.time*100)))
# Emit performance mode and arguments
perf_args = {}
for i in self.input + self.dimensions:
if not i.is_PerfKnob:
continue
try:
perf_args[i.name] = args[i.name]
except KeyError:
# Try with the aliases
for a in i._arg_names:
if a in args:
perf_args[a] = args[a]
break
perf("Performance[mode=%s] arguments: %s" % (self._mode, perf_args))
return summary
# Pickling support
def __getstate__(self):
if self._lib:
state = dict(self.__dict__)
# The compiled shared-object will be pickled; upon unpickling, it
# will be restored into a potentially different temporary directory,
# so the entire process during which the shared-object is loaded and
# given to ctypes must be performed again
state['_lib'] = None
state['_cfunction'] = None
# Do not pickle the `args` used to construct the Operator. Not only
# would this be completely useless, but it might also lead to
# allocating additional memory upon unpickling, as the user-provided
# equations typically carry different instances of the same Function
# (e.g., f(t, x-1), f(t, x), f(t, x+1)), which are different objects
# with distinct `.data` fields
state['_args'] = None
with open(self._lib._name, 'rb') as f:
state['binary'] = f.read()
state['soname'] = self._soname
return state
else:
return self.__dict__
def __getnewargs_ex__(self):
return (None,), {}
def __setstate__(self, state):
soname = state.pop('soname', None)
binary = state.pop('binary', None)
for k, v in state.items():
setattr(self, k, v)
if soname is not None:
self._compiler.save(soname, binary)
self._lib = self._compiler.load(soname)
self._lib.name = soname
# Default action (perform or bypass) for selected compilation passes upon
# recursive compilation
# NOTE: it may not only be pointless to apply the following passes recursively
# (because once, during the main compilation phase, is simply enough), but also
# dangerous as some of them (the minority) might break in some circumstances
# if applied in cascade (e.g., `linearization` on top of `linearization`)
rcompile_registry = {
'mpi': False,
'linearize': False,
'place-transfers': False
}
def rcompile(expressions, kwargs=None):
"""
Perform recursive compilation on an ordered sequence of symbolic expressions.
"""
if not kwargs or 'options' not in kwargs:
kwargs = parse_kwargs(**kwargs)
cls = operator_selector(**kwargs)
kwargs = cls._normalize_kwargs(**kwargs)
else:
cls = operator_selector(**kwargs)
# Tweak the compilation kwargs
options = dict(kwargs['options'])
options.update(rcompile_registry)
kwargs['options'] = options
# Recursive profiling not supported -- would be a complete mess
kwargs.pop('profiler', None)
return cls._lower(expressions, **kwargs)
# Misc helpers
IRs = namedtuple('IRs', 'expressions clusters stree uiet iet')
class ArgumentsMap(dict):
def __init__(self, args, grid, op):
super().__init__(args)
self.grid = grid
self.allocator = op._allocator
self.platform = op._platform
self.language = op._language
self.compiler = op._compiler
self.options = op._options
@property
def comm(self):
"""The MPI communicator the arguments are collective over."""
return self.grid.comm if self.grid is not None else MPI.COMM_NULL
@property
def opkwargs(self):
temp_registry = {v: k for k, v in platform_registry.items()}
platform = temp_registry[self.platform]
temp_registry = {v: k for k, v in compiler_registry.items()}
compiler = temp_registry[self.compiler.__class__]
return {'platform': platform, 'compiler': compiler, 'language': self.language}
def parse_kwargs(**kwargs):
"""
Parse keyword arguments provided to an Operator.
"""
# `dse` -- deprecated, dropped
dse = kwargs.pop("dse", None)
if dse is not None:
warning("The `dse` argument is deprecated. "
"The optimization level is now controlled via the `opt` argument")
# `dle` -- deprecated, replaced by `opt`
if 'dle' in kwargs:
warning("The `dle` argument is deprecated. "
"The optimization level is now controlled via the `opt` argument")
dle = kwargs.pop('dle')
if 'opt' in kwargs:
warning("Both `dle` and `opt` were passed; ignoring `dle` argument")
opt = kwargs.pop('opt')
else:
warning("Setting `opt=%s`" % str(dle))
opt = dle
elif 'opt' in kwargs:
opt = kwargs.pop('opt')
else:
opt = configuration['opt']
if not opt or isinstance(opt, str):
mode, options = opt, {}
elif isinstance(opt, tuple):
if len(opt) == 0:
mode, options = 'noop', {}
elif isinstance(opt[-1], dict):
if len(opt) == 2:
mode, options = opt
else:
mode, options = tuple(flatten(i.split(',') for i in opt[:-1])), opt[-1]
else:
mode, options = tuple(flatten(i.split(',') for i in opt)), {}
else:
raise InvalidOperator("Illegal `opt=%s`" % str(opt))
# `opt`, deprecated kwargs
kwopenmp = kwargs.get('openmp', options.get('openmp'))
if kwopenmp is None:
openmp = kwargs.get('language', configuration['language']) == 'openmp'
else:
openmp = kwopenmp
# `opt`, options
options = dict(options)
options.setdefault('openmp', openmp)
options.setdefault('mpi', configuration['mpi'])
for k, v in configuration['opt-options'].items():
options.setdefault(k, v)
# Handle deprecations
deprecated_options = ('cire-mincost-inv', 'cire-mincost-sops', 'cire-maxalias')
for i in deprecated_options:
try:
options.pop(i)
warning("Ignoring deprecated optimization option `%s`" % i)
except KeyError:
pass
kwargs['options'] = options
# `opt`, mode
if mode is None:
mode = 'noop'
kwargs['mode'] = mode
# `platform`
platform = kwargs.get('platform')
if platform is not None:
if not isinstance(platform, str):
raise ValueError("Argument `platform` should be a `str`")
if platform not in configuration._accepted['platform']:
raise InvalidOperator("Illegal `platform=%s`" % str(platform))
kwargs['platform'] = platform_registry[platform]()
else:
kwargs['platform'] = configuration['platform']
# `language`
language = kwargs.get('language')
if language is not None:
if not isinstance(language, str):
raise ValueError("Argument `language` should be a `str`")
if language not in configuration._accepted['language']:
raise InvalidOperator("Illegal `language=%s`" % str(language))
kwargs['language'] = language
elif kwopenmp is not None:
# Handle deprecated `openmp` kwarg for backward compatibility
kwargs['language'] = 'openmp' if openmp else 'C'
else:
kwargs['language'] = configuration['language']
# `compiler`
compiler = kwargs.get('compiler')
if compiler is not None:
if not isinstance(compiler, str):
raise ValueError("Argument `compiler` should be a `str`")
if compiler not in configuration._accepted['compiler']:
raise InvalidOperator("Illegal `compiler=%s`" % str(compiler))
kwargs['compiler'] = compiler_registry[compiler](platform=kwargs['platform'],
language=kwargs['language'],
mpi=configuration['mpi'])
elif any([platform, language]):
kwargs['compiler'] =\
configuration['compiler'].__new_with__(platform=kwargs['platform'],
language=kwargs['language'],
mpi=configuration['mpi'])
else:
kwargs['compiler'] = configuration['compiler'].__new_with__()
# `allocator`
kwargs['allocator'] = default_allocator(
'%s.%s.%s' % (kwargs['compiler'].name,
kwargs['language'],
kwargs['platform'])
)
return kwargs