def find_undefined_variables(self, statements):
'''
Find identifiers that are not in ``self.variables`` dictionary.
Brian does not save the ``_lio_`` variables it uses anywhere. This is
problematic for our GSL implementation because we save the lio variables
in the ``_dataholder`` struct (for which we need the datatype of the
variables). This function adds the left hand side variables that are
used in the vector code to the variable dictionary as
`AuxiliaryVariable`\ s (all we need later is the datatype).
Parameters
----------
statements : list
list of statement objects (need to have the dtype attribute)
Notes
-----
I keep ``self.variables`` and ``other_variables`` separate so I can
distinguish what variables are in the Brian namespace and which ones are
defined in the code itself.
'''
variables = self.variables
other_variables = {}
for statement in statements:
var, op, expr, comment = (statement.var, statement.op,
statement.expr, statement.comment)
if var not in variables:
other_variables[var] = AuxiliaryVariable(var,
dtype=statement.dtype)
return other_variables
def render_node(self, node):
'''
Assumes that the node has already been fully processed by BrianASTRenderer
'''
# can we pull this out?
if node.scalar and node.complexity > 0:
expr = self.node_renderer.render_node(
self.arithmetic_simplifier.render_node(node))
if expr in self.loop_invariants:
name = self.loop_invariants[expr]
else:
self.n += 1
name = '_lio_' + self.extra_lio_prefix + str(self.n)
self.loop_invariants[expr] = name
self.loop_invariant_dtypes[name] = node.dtype
numpy_dtype = {
'boolean': bool,
'integer': int,
'float': prefs.core.default_float_dtype
}[node.dtype]
self.variables[name] = AuxiliaryVariable(name,
dtype=numpy_dtype,
scalar=True)
# None is the expression context, we don't use it so we just set to None
newnode = ast.Name(name, None)
newnode.scalar = True
newnode.dtype = node.dtype
newnode.complexity = 0
newnode.stateless = node.stateless
return newnode
# otherwise, render node as usual
return super(Simplifier, self).render_node(node)
def add_gsl_variables_as_non_scalar(self, diff_vars):
'''
Add _gsl variables as non-scalar.
In `GSLStateUpdater` the differential equation variables are substituted
with GSL tags that describe the information needed to translate the
conventional Brian code to GSL compatible code. This function tells
Brian that the variables that contain these tags should always be vector
variables. If we don't do this, Brian renders the tag-variables as
scalar if no vector variables are used in the right hand side of the
expression.
Parameters
----------
diff_vars : dict
dictionary with variables as keys and differential equation index as
value
'''
for var, ind in diff_vars.items():
name = '_gsl_{var}_f{ind}'.format(var=var, ind=ind)
self.variables[name] = AuxiliaryVariable(var, scalar=False)
def make_statements(code, variables, dtype, optimise=True, blockname=''):
'''
Turn a series of abstract code statements into Statement objects, inferring
whether each line is a set/declare operation, whether the variables are
constant or not, and handling the cacheing of subexpressions.
Parameters
----------
code : str
A (multi-line) string of statements.
variables : dict-like
A dictionary of with `Variable` and `Function` objects for every
identifier used in the `code`.
dtype : `dtype`
The data type to use for temporary variables
optimise : bool, optional
Whether to optimise expressions, including
pulling out loop invariant expressions and putting them in new
scalar constants. Defaults to ``False``, since this function is also
used just to in contexts where we are not interested by this kind of
optimisation. For the main code generation stage, its value is set by
the `codegen.loop_invariant_optimisations` preference.
blockname : str, optional
A name for the block (used to name intermediate variables to avoid
name clashes when multiple blocks are used together)
Returns
-------
scalar_statements, vector_statements : (list of `Statement`, list of `Statement`)
Lists with statements that are to be executed once and statements that
are to be executed once for every neuron/synapse/... (or in a vectorised
way)
Notes
-----
If ``optimise`` is ``True``, then the
``scalar_statements`` may include newly introduced scalar constants that
have been identified as loop-invariant and have therefore been pulled out
of the vector statements. The resulting statements will also use augmented
assignments where possible, i.e. a statement such as ``w = w + 1`` will be
replaced by ``w += 1``. Also, statements involving booleans will have
additional information added to them (see `Statement` for details)
describing how the statement can be reformulated as a sequence of if/then
statements. Calls `~brian2.codegen.optimisation.optimise_statements`.
'''
code = strip_empty_lines(deindent(code))
lines = re.split(r'[;\n]', code)
lines = [LineInfo(code=line) for line in lines if len(line)]
# Do a copy so we can add stuff without altering the original dict
variables = dict(variables)
# we will do inference to work out which lines are := and which are =
defined = set(k for k, v in variables.iteritems()
if not isinstance(v, AuxiliaryVariable))
for line in lines:
statement = None
# parse statement into "var op expr"
var, op, expr, comment = parse_statement(line.code)
if op == '=':
if var not in defined:
op = ':='
defined.add(var)
if var not in variables:
is_scalar = is_scalar_expression(expr, variables)
new_var = AuxiliaryVariable(var,
dtype=dtype,
scalar=is_scalar)
variables[var] = new_var
elif not variables[var].is_boolean:
sympy_expr = str_to_sympy(expr, variables)
sympy_var = sympy.Symbol(var, real=True)
try:
collected = sympy.collect(sympy_expr,
sympy_var,
exact=True,
evaluate=False)
except AttributeError:
# If something goes wrong during collection, e.g. collect
# does not work for logical expressions
collected = {1: sympy_expr}
if (len(collected) == 2
and set(collected.keys()) == {1, sympy_var}
and collected[sympy_var] == 1):
# We can replace this statement by a += assignment
statement = Statement(var,
'+=',
sympy_to_str(collected[1]),
comment,
dtype=variables[var].dtype,
scalar=variables[var].scalar)
elif len(collected) == 1 and sympy_var in collected:
# We can replace this statement by a *= assignment
statement = Statement(var,
'*=',
sympy_to_str(collected[sympy_var]),
comment,
dtype=variables[var].dtype,
scalar=variables[var].scalar)
if statement is None:
statement = Statement(var,
op,
expr,
comment,
dtype=variables[var].dtype,
scalar=variables[var].scalar)
line.statement = statement
# for each line will give the variable being written to
line.write = var
# each line will give a set of variables which are read
line.read = get_identifiers_recursively([expr], variables)
# All writes to scalar variables must happen before writes to vector
# variables
scalar_write_done = False
for line in lines:
stmt = line.statement
if stmt.op != ':=' and variables[
stmt.var].scalar and scalar_write_done:
raise SyntaxError(
('All writes to scalar variables in a code block '
'have to be made before writes to vector '
'variables. Illegal write to %s.') % line.write)
elif not variables[stmt.var].scalar:
scalar_write_done = True
# all variables which are written to at some point in the code block
# used to determine whether they should be const or not
all_write = set(line.write for line in lines)
# backwards compute whether or not variables will be read again
# note that will_read for a line gives the set of variables it will read
# on the current line or subsequent ones. will_write gives the set of
# variables that will be written after the current line
will_read = set()
will_write = set()
for line in lines[::-1]:
will_read = will_read.union(line.read)
line.will_read = will_read.copy()
line.will_write = will_write.copy()
will_write.add(line.write)
subexpressions = dict((name, val) for name, val in variables.items()
if isinstance(val, Subexpression))
# sort subexpressions into an order so that subexpressions that don't depend
# on other subexpressions are first
subexpr_deps = dict((name, [dep for dep in subexpr.identifiers if dep in subexpressions]) for \
name, subexpr in subexpressions.items())
sorted_subexpr_vars = topsort(subexpr_deps)
statements = []
# none are yet defined (or declared)
subdefined = dict((name, False) for name in subexpressions.keys())
for line in lines:
stmt = line.statement
read = line.read
write = line.write
will_read = line.will_read
will_write = line.will_write
# update/define all subexpressions needed by this statement
for var in sorted_subexpr_vars:
if var not in read:
continue
subexpression = subexpressions[var]
# if already defined/declared
if subdefined[var]:
op = '='
constant = False
else:
op = ':='
subdefined[var] = True
# set to constant only if we will not write to it again
constant = var not in will_write
# check all subvariables are not written to again as well
if constant:
ids = subexpression.identifiers
constant = all(v not in will_write for v in ids)
statement = Statement(var,
op,
subexpression.expr,
comment='',
dtype=variables[var].dtype,
constant=constant,
subexpression=True,
scalar=variables[var].scalar)
statements.append(statement)
var, op, expr, comment = stmt.var, stmt.op, stmt.expr, stmt.comment
# constant only if we are declaring a new variable and we will not
# write to it again
constant = op == ':=' and var not in will_write
statement = Statement(var,
op,
expr,
comment,
dtype=variables[var].dtype,
constant=constant,
scalar=variables[var].scalar)
statements.append(statement)
scalar_statements = [s for s in statements if s.scalar]
vector_statements = [s for s in statements if not s.scalar]
if optimise and prefs.codegen.loop_invariant_optimisations:
scalar_statements, vector_statements = optimise_statements(
scalar_statements,
vector_statements,
variables,
blockname=blockname)
return scalar_statements, vector_statements
def make_statements(code, variables, dtype):
'''
Turn a series of abstract code statements into Statement objects, inferring
whether each line is a set/declare operation, whether the variables are
constant or not, and handling the cacheing of subexpressions.
Parameters
----------
code : str
A (multi-line) string of statements.
variables : dict-like
A dictionary of with `Variable` and `Function` objects for every
identifier used in the `code`.
dtype : `dtype`
The data type to use for temporary variables
Returns
-------
scalar_statements, vector_statements : (list of `Statement`, list of `Statement`)
Lists with statements that are to be executed once and statements that
are to be executed once for every neuron/synapse/... (or in a vectorised
way)
Notes
-----
The `scalar_statements` may include newly introduced scalar constants that
have been identified as loop-invariant and have therefore been pulled out
of the vector statements. The resulting statements will also use augmented
assignments where possible, i.e. a statement such as ``w = w + 1`` will be
replaced by ``w += 1``.
'''
code = strip_empty_lines(deindent(code))
lines = re.split(r'[;\n]', code)
lines = [LineInfo(code=line) for line in lines if len(line)]
if DEBUG:
print 'INPUT CODE:'
print code
# Do a copy so we can add stuff without altering the original dict
variables = dict(variables)
# we will do inference to work out which lines are := and which are =
defined = set(k for k, v in variables.iteritems()
if not isinstance(v, AuxiliaryVariable))
for line in lines:
statement = None
# parse statement into "var op expr"
var, op, expr, comment = parse_statement(line.code)
if op == '=':
if var not in defined:
op = ':='
defined.add(var)
if var not in variables:
is_scalar = is_scalar_expression(expr, variables)
new_var = AuxiliaryVariable(
var,
Unit(1), # doesn't matter here
dtype=dtype,
scalar=is_scalar)
variables[var] = new_var
elif not variables[var].is_boolean:
sympy_expr = str_to_sympy(expr)
sympy_var = sympy.Symbol(var, real=True)
try:
collected = sympy.collect(sympy_expr,
sympy_var,
exact=True,
evaluate=False)
except AttributeError:
# If something goes wrong during collection, e.g. collect
# does not work for logical expressions
collected = {1: sympy_expr}
if (len(collected) == 2
and set(collected.keys()) == {1, sympy_var}
and collected[sympy_var] == 1):
# We can replace this statement by a += assignment
statement = Statement(var,
'+=',
sympy_to_str(collected[1]),
comment,
dtype=variables[var].dtype,
scalar=variables[var].scalar)
elif len(collected) == 1 and sympy_var in collected:
# We can replace this statement by a *= assignment
statement = Statement(var,
'*=',
sympy_to_str(collected[sympy_var]),
comment,
dtype=variables[var].dtype,
scalar=variables[var].scalar)
if statement is None:
statement = Statement(var,
op,
expr,
comment,
dtype=variables[var].dtype,
scalar=variables[var].scalar)
line.statement = statement
# for each line will give the variable being written to
line.write = var
# each line will give a set of variables which are read
line.read = get_identifiers_recursively([expr], variables)
# All writes to scalar variables must happen before writes to vector
# variables
scalar_write_done = False
for line in lines:
stmt = line.statement
if stmt.op != ':=' and variables[
stmt.var].scalar and scalar_write_done:
raise SyntaxError(
('All writes to scalar variables in a code block '
'have to be made before writes to vector '
'variables. Illegal write to %s.') % line.write)
elif not variables[stmt.var].scalar:
scalar_write_done = True
if DEBUG:
print 'PARSED STATEMENTS:'
for line in lines:
print line.statement, 'Read:' + str(
line.read), 'Write:' + line.write
# all variables which are written to at some point in the code block
# used to determine whether they should be const or not
all_write = set(line.write for line in lines)
if DEBUG:
print 'ALL WRITE:', all_write
# backwards compute whether or not variables will be read again
# note that will_read for a line gives the set of variables it will read
# on the current line or subsequent ones. will_write gives the set of
# variables that will be written after the current line
will_read = set()
will_write = set()
for line in lines[::-1]:
will_read = will_read.union(line.read)
line.will_read = will_read.copy()
line.will_write = will_write.copy()
will_write.add(line.write)
if DEBUG:
print 'WILL READ/WRITE:'
for line in lines:
print line.statement, 'Read:' + str(
line.will_read), 'Write:' + str(line.will_write)
# generate cacheing statements for common subexpressions
# cached subexpressions need to be recomputed whenever they are to be used
# on the next line, and currently invalid (meaning that the current value
# stored in the subexpression variable is no longer accurate because one
# of the variables appearing in it has changed). All subexpressions start
# as invalid, and are invalidated whenever one of the variables appearing
# in the RHS changes value.
subexpressions = dict((name, val) for name, val in variables.items()
if isinstance(val, Subexpression))
# sort subexpressions into an order so that subexpressions that don't depend
# on other subexpressions are first
subexpr_deps = dict((name, [dep for dep in subexpr.identifiers if dep in subexpressions]) for \
name, subexpr in subexpressions.items())
sorted_subexpr_vars = topsort(subexpr_deps)
if DEBUG:
print 'SUBEXPRESSIONS:', subexpressions.keys()
statements = []
# all start as invalid
valid = dict((name, False) for name in subexpressions.keys())
# none are yet defined (or declared)
subdefined = dict((name, False) for name in subexpressions.keys())
for line in lines:
stmt = line.statement
read = line.read
write = line.write
will_read = line.will_read
will_write = line.will_write
# check that all subexpressions in expr are valid, and if not
# add a definition/set its value, and set it to be valid
# scan through in sorted order so that recursive subexpression dependencies
# are handled in the right order
for var in sorted_subexpr_vars:
if var not in read:
continue
# if subexpression, and invalid
if not valid.get(var, True): # all non-subexpressions are valid
subexpression = subexpressions[var]
# if already defined/declared
if subdefined[var]:
op = '='
constant = False
else:
op = ':='
subdefined[var] = True
# set to constant only if we will not write to it again
constant = var not in will_write
# check all subvariables are not written to again as well
if constant:
ids = subexpression.identifiers
constant = all(v not in will_write for v in ids)
valid[var] = True
statement = Statement(var,
op,
subexpression.expr,
comment='',
dtype=variables[var].dtype,
constant=constant,
subexpression=True,
scalar=variables[var].scalar)
statements.append(statement)
var, op, expr, comment = stmt.var, stmt.op, stmt.expr, stmt.comment
# invalidate any subexpressions including var, recursively
# we do this by having a set of variables that are invalid that we
# start with the changed var and increase by any subexpressions we
# find that have a dependency on something in the invalid set. We
# go through in sorted subexpression order so that the invalid set
# is increased in the right order
invalid = {var}
for subvar in sorted_subexpr_vars:
spec = subexpressions[subvar]
spec_ids = set(spec.identifiers)
if spec_ids.intersection(invalid):
valid[subvar] = False
invalid.add(subvar)
# constant only if we are declaring a new variable and we will not
# write to it again
constant = op == ':=' and var not in will_write
statement = Statement(var,
op,
expr,
comment,
dtype=variables[var].dtype,
constant=constant,
scalar=variables[var].scalar)
statements.append(statement)
if DEBUG:
print 'OUTPUT STATEMENTS:'
for stmt in statements:
print stmt
scalar_statements = [s for s in statements if s.scalar]
vector_statements = [s for s in statements if not s.scalar]
if prefs.codegen.loop_invariant_optimisations:
scalar_constants, vector_statements = apply_loop_invariant_optimisations(
vector_statements, variables, dtype)
scalar_statements.extend(scalar_constants)
return scalar_statements, vector_statements
