From 77c56f647c05bdf0d1718f9995761fe6cbb62643 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ren=C3=A9=20He=C3=9F?= <rene.hess@iwr.uni-heidelberg.de>
Date: Thu, 7 Sep 2017 11:22:43 +0200
Subject: [PATCH] Do not generate code for stage 1 sumfact kernels that don't
get used
Save all stage 1 sum factorization kernels that are used in
accumulation expression in the cache during the dry run. Discard all
inactive sum factorization kernels in decide_vetorization_strategy.
---
python/dune/perftool/sumfact/accumulation.py | 23 +++++++++++++
python/dune/perftool/sumfact/basis.py | 10 ++++++
python/dune/perftool/sumfact/realization.py | 1 -
python/dune/perftool/sumfact/vectorization.py | 33 ++++++++++++++-----
4 files changed, 57 insertions(+), 10 deletions(-)
diff --git a/python/dune/perftool/sumfact/accumulation.py b/python/dune/perftool/sumfact/accumulation.py
index 50f4d169..c9c32b48 100644
--- a/python/dune/perftool/sumfact/accumulation.py
+++ b/python/dune/perftool/sumfact/accumulation.py
@@ -6,6 +6,7 @@ from dune.perftool.pdelab.argument import (name_accumulation_variable,
from dune.perftool.generation import (backend,
domain,
dump_accumulate_timer,
+ generator_factory,
get_counted_variable,
get_counter,
iname,
@@ -40,10 +41,29 @@ from pytools import ImmutableRecord
import loopy as lp
import numpy as np
import pymbolic.primitives as prim
+from pymbolic.mapper import WalkMapper
import ufl.classes as uc
from ufl import FiniteElement, MixedElement, TensorProductElement
+basis_sf_kernels = generator_factory(item_tags=("basis_sf_kernels",), context_tags='kernel', no_deco=True)
+
+
+class SumfactCollectMapper(WalkMapper):
+ def map_tagged_variable(self, expr, *args, **kwargs):
+ self.visit(expr, *args, **kwargs)
+ self.post_visit(expr, *args, **kwargs)
+
+ def map_variable(self, expr, *args, **kwargs):
+ self.visit(expr, *args, **kwargs)
+ self.post_visit(expr, *args, **kwargs)
+
+ def map_sumfact_kernel(self, expr, *args, **kwargs):
+ basis_sf_kernels(expr)
+ self.visit(expr, *args, **kwargs)
+ self.post_visit(expr, *args, **kwargs)
+
+
@iname
def _sumfact_iname(bound, _type, count):
name = "sf_{}_{}".format(_type, str(count))
@@ -199,6 +219,9 @@ def generate_accumulation_instruction(expr, visitor):
test_info = visitor.test_info
trial_info = visitor.trial_info
+ # Cache all stage 1 sum factorization kernels used in this expression
+ SumfactCollectMapper()(expr)
+
# Number of basis functions per direction
leaf_element = test_info.element
from ufl import MixedElement
diff --git a/python/dune/perftool/sumfact/basis.py b/python/dune/perftool/sumfact/basis.py
index 8e745c7d..00ef2966 100644
--- a/python/dune/perftool/sumfact/basis.py
+++ b/python/dune/perftool/sumfact/basis.py
@@ -146,6 +146,11 @@ def pymbolic_coefficient_gradient(element, restriction, index, coeff_func, visit
from dune.perftool.sumfact.vectorization import attach_vectorization_info
vsf = attach_vectorization_info(sf)
+ # If this sum factorization kernel was not used in the dry run we
+ # just return 0
+ if vsf == 0:
+ return 0, None
+
from dune.perftool.sumfact.realization import realize_sum_factorization_kernel
var, insn_dep = realize_sum_factorization_kernel(vsf)
@@ -182,6 +187,11 @@ def pymbolic_coefficient(element, restriction, index, coeff_func, visitor_indice
from dune.perftool.sumfact.vectorization import attach_vectorization_info
vsf = attach_vectorization_info(sf)
+ # If this sum factorization kernel was not used in the dry run we
+ # just return 0
+ if vsf == 0:
+ return 0, None
+
# Add a sum factorization kernel that implements the evaluation of
# the basis functions at quadrature points (stage 1)
from dune.perftool.sumfact.realization import realize_sum_factorization_kernel
diff --git a/python/dune/perftool/sumfact/realization.py b/python/dune/perftool/sumfact/realization.py
index 3f1afe9c..5c9e4a5b 100644
--- a/python/dune/perftool/sumfact/realization.py
+++ b/python/dune/perftool/sumfact/realization.py
@@ -24,7 +24,6 @@ from dune.perftool.sumfact.permutation import (sumfact_permutation_strategy,
permute_backward,
permute_forward,
)
-from dune.perftool.sumfact.vectorization import attach_vectorization_info
from dune.perftool.sumfact.accumulation import sumfact_iname
from dune.perftool.loopy.vcl import ExplicitVCLCast
diff --git a/python/dune/perftool/sumfact/vectorization.py b/python/dune/perftool/sumfact/vectorization.py
index 08e07782..3a044794 100644
--- a/python/dune/perftool/sumfact/vectorization.py
+++ b/python/dune/perftool/sumfact/vectorization.py
@@ -140,37 +140,52 @@ def decide_vectorization_strategy():
"""
logger = logging.getLogger(__name__)
+ # Retrieve all sum factorization kernels for stage 1 and 3
from dune.perftool.generation import retrieve_cache_items
- sumfacts = [i for i in retrieve_cache_items("kernel_default and sumfactnodes")]
+ all_sumfacts = [i for i in retrieve_cache_items("kernel_default and sumfactnodes")]
+
+ # Stage 1 sumfactorizations that were actually used
+ basis_sumfacts = [i for i in retrieve_cache_items('kernel_default and basis_sf_kernels')]
+
+ # This means we can have sum factorizations that will not get used
+ inactive_sumfacts = [i for i in all_sumfacts if i.stage == 1 and i not in basis_sumfacts]
+
+ # All sum factorization kernels that get used
+ active_sumfacts = [i for i in all_sumfacts if i.stage == 3 or i in basis_sumfacts]
+
+ # We map inacitve sum factorizatino kernels to 0
sfdict = {}
+ for sf in inactive_sumfacts:
+ sfdict[sf] = 0
- logger.debug("decide_vectorization_strategy: Found {} sum factorization nodes".format(len(sumfacts)))
+ logger.debug("decide_vectorization_strategy: Found {} active sum factorization nodes"
+ .format(len(active_sumfacts)))
if get_option("vectorize_grads"):
# Currently we base our idea here on the fact that we only group sum
# factorization kernels with the same input.
- inputkeys = set(sf.input_key for sf in sumfacts)
+ inputkeys = set(sf.input_key for sf in active_sumfacts)
for inputkey in inputkeys:
width = get_vcl_type_size(np.float64)
- sumfact_filter = [sf for sf in sumfacts if sf.input_key == inputkey]
+ sumfact_filter = [sf for sf in active_sumfacts if sf.input_key == inputkey]
for old, new in horizontal_vectorization_strategy(sumfact_filter, width).items():
sfdict[old] = new
elif get_option("vectorize_slice"):
- for sumfact in sumfacts:
+ for sumfact in active_sumfacts:
width = get_vcl_type_size(np.float64)
for old, new in vertical_vectorization_strategy(sumfact, width).items():
sfdict[old] = new
elif get_option("vectorize_diagonal"):
- inputkeys = set(sf.input_key for sf in sumfacts)
+ inputkeys = set(sf.input_key for sf in active_sumfacts)
for inputkey in inputkeys:
width = get_vcl_type_size(np.float64)
- sumfact_filter = [sf for sf in sumfacts if sf.input_key == inputkey]
+ sumfact_filter = [sf for sf in active_sumfacts if sf.input_key == inputkey]
for old, new in diagonal_vectorization_strategy(sumfact_filter, width).items():
sfdict[old] = new
else:
- for old, new in no_vectorization(sumfacts).items():
+ for old, new in no_vectorization(active_sumfacts).items():
sfdict[old] = new
# Register the results
- for sf in sumfacts:
+ for sf in all_sumfacts:
_cache_vectorization_info(sf, sfdict.get(sf, no_vec(sf)))
--
GitLab