From 896cb5e429cad84b1cec870cb815106c20abf6a4 Mon Sep 17 00:00:00 2001
From: Dominic Kempf <dominic.kempf@iwr.uni-heidelberg.de>
Date: Thu, 15 Feb 2018 09:15:35 +0100
Subject: [PATCH] Add a temporary solution for extraction of vectorized output
properties
---
python/dune/perftool/sumfact/realization.py | 36 ++++++++----
python/dune/perftool/sumfact/symbolic.py | 62 +++++++++------------
2 files changed, 50 insertions(+), 48 deletions(-)
diff --git a/python/dune/perftool/sumfact/realization.py b/python/dune/perftool/sumfact/realization.py
index cc6704af..20949b20 100644
--- a/python/dune/perftool/sumfact/realization.py
+++ b/python/dune/perftool/sumfact/realization.py
@@ -193,37 +193,51 @@ def _realize_sum_factorization_kernel(sf):
# of the Sumfactorization into some global data structure.
if l == len(matrix_sequence) - 1 and get_form_option('fastdg') and sf.stage == 3:
ft = get_global_context_value("form_type")
- if sf.test_element_index is None:
- direct_output = "{}_access".format(sf.accumvar)
+ # TODO This one will break super-hard!
+ if sf.vectorized:
+ test_element = sf.output.outputs[0].test_element
+ test_element_index = sf.output.outputs[0].test_element_index
+ trial_element = sf.output.outputs[0].trial_element
+ trial_element_index = sf.output.outputs[0].trial_element_index
+ accumvar = sf.output.outputs[0].accumvar
else:
- direct_output = "{}_access_comp{}".format(sf.accumvar, sf.test_element_index)
+ test_element = sf.output.test_element
+ test_element_index = sf.output.test_element_index
+ trial_element = sf.output.trial_element
+ trial_element_index = sf.output.trial_element_index
+ accumvar = sf.output.accumvar
+
+ if test_element_index is None:
+ direct_output = "{}_access".format(accumvar)
+ else:
+ direct_output = "{}_access_comp{}".format(accumvar, test_element_index)
if ft == 'residual' or ft == 'jacobian_apply':
globalarg(direct_output,
shape=output_shape,
dim_tags=novec_ftags,
- offset=_dof_offset(sf.test_element, sf.test_element_index),
+ offset=_dof_offset(test_element, test_element_index),
)
- alias_data_array(direct_output, sf.accumvar)
+ alias_data_array(direct_output, accumvar)
assignee = prim.Subscript(prim.Variable(direct_output), output_inames)
else:
assert ft == 'jacobian'
- direct_output = "{}x{}".format(direct_output, sf.trial_element_index)
- rowsize = sum(tuple(s for s in _local_sizes(sf.trial_element)))
- element = sf.trial_element
+ direct_output = "{}x{}".format(direct_output, trial_element_index)
+ rowsize = sum(tuple(s for s in _local_sizes(trial_element)))
+ element = trial_element
if isinstance(element, MixedElement):
- element = element.extract_component(sf.trial_element_index)[1]
+ element = element.extract_component(trial_element_index)[1]
other_shape = tuple(element.degree() + 1 for e in range(sf.length))
from pytools import product
manual_strides = tuple("stride:{}".format(rowsize * product(output_shape[:i])) for i in range(sf.length))
dim_tags = "{},{}".format(novec_ftags, ",".join(manual_strides))
globalarg(direct_output,
shape=other_shape + output_shape,
- offset=rowsize * _dof_offset(sf.test_element, sf.test_element_index) + _dof_offset(sf.trial_element, sf.trial_element_index),
+ offset=rowsize * _dof_offset(test_element, test_element_index) + _dof_offset(trial_element, trial_element_index),
dim_tags=dim_tags,
)
- alias_data_array(direct_output, sf.accumvar)
+ alias_data_array(direct_output, accumvar)
# TODO: It is at least questionnable, whether using the *order* of the inames in here
# for indexing is a good idea. Then again, it is hard to find an alternative.
_ansatz_inames = tuple(prim.Variable(i) for i in sf.within_inames)
diff --git a/python/dune/perftool/sumfact/symbolic.py b/python/dune/perftool/sumfact/symbolic.py
index eb9df931..322806fa 100644
--- a/python/dune/perftool/sumfact/symbolic.py
+++ b/python/dune/perftool/sumfact/symbolic.py
@@ -30,22 +30,6 @@ class SumfactKernelInputBase(object):
raise NotImplementedError
-class SumfactKernelOutputBase(object):
- @property
- def direct_output_is_possible(self):
- return False
-
- @property
- def within_inames(self):
- return ()
-
- def realize(self, sf, dep):
- return dep
-
- def realize_direct(self):
- raise NotImplementedError
-
-
class VectorSumfactKernelInput(SumfactKernelInputBase):
def __init__(self, inputs):
assert(isinstance(inputs, tuple))
@@ -88,6 +72,27 @@ class VectorSumfactKernelInput(SumfactKernelInputBase):
raise NotImplementedError("SIMD loads from scalars not implemented!")
+class SumfactKernelOutputBase(object):
+ @property
+ def direct_output_is_possible(self):
+ return False
+
+ @property
+ def within_inames(self):
+ return ()
+
+ def realize(self, sf, dep):
+ return dep
+
+ def realize_direct(self):
+ raise NotImplementedError
+
+
+class VectorSumfactKernelOutput(SumfactKernelOutputBase):
+ def __init__(self, outputs):
+ self.outputs = outputs
+
+
class SumfactKernelBase(object):
pass
@@ -508,31 +513,10 @@ class VectorizedSumfactKernel(SumfactKernelBase, ImmutableRecord, prim.Variable)
def within_inames(self):
return self.kernels[0].within_inames
- @property
- def test_element(self):
- return self.kernels[0].test_element
-
- @property
- def test_element_index(self):
- return self.kernels[0].test_element_index
-
- @property
- def trial_element(self):
- return self.kernels[0].trial_element
-
- @property
- def trial_element_index(self):
- return self.kernels[0].trial_element_index
-
@property
def predicates(self):
return self.kernels[0].predicates
- @property
- def accumvar(self):
- assert len(set(k.accumvar for k in self.kernels)) == 1
- return self.kernels[0].accumvar
-
@property
def transposed(self):
return self.kernels[0].transposed
@@ -556,6 +540,10 @@ class VectorizedSumfactKernel(SumfactKernelBase, ImmutableRecord, prim.Variable)
def input(self):
return VectorSumfactKernelInput(tuple(k.input for k in self.kernels))
+ @property
+ def output(self):
+ return VectorSumfactKernelOutput(tuple(k.output for k in self.kernels))
+
@property
def cache_key(self):
return (tuple(k.cache_key for k in self.kernels), self.buffer)
--
GitLab