- edited description
UFLACS fails on Holzapfel-Ogden
Issue #133
resolved
$ ffc -v -r uflacs ~/Downloads/holzapfel.ufl
This is FFC, the FEniCS Form Compiler, version 2016.2.0.dev0.
UFC backend version 2016.2.0.dev0, signature 59227d849f669930c2cf84f1ef02c21153a5b091.
For further information, visit https://bitbucket.org/fenics-project/ffc/.
Python 2.7.12 (default, Jul 1 2016, 15:12:24)
[GCC 5.4.0 20160609] on linux2
Compiling form holzapfel
Compiler stage 1: Analyzing form(s)
-----------------------------------
Geometric dimension: 3
Number of cell subdomains: 0
Rank: 2
Arguments: '(v_0, v_1)'
Number of coefficients: 12
Coefficients: '[w_0, w_1, w_2, w_3, w_4, w_5, w_6, w_7, w_8, w_9, w_{
10}, w_{11}]'
Unique elements: 'Vector<3 x CG1(?)>, R0(?), Vector<3 x R0(?)>'
Unique sub elements: 'Vector<3 x CG1(?)>, R0(?), Vector<3 x R0(?)>, CG1(?)'
representation: uflacs
quadrature_rule: auto --> default
quadrature_degree: auto --> 0
quadrature_degree: 0
Compiler stage 1 finished in 0.499413 seconds.
Compiler stage 2: Computing intermediate representation
-------------------------------------------------------
Computing representation of 4 elements
Computing representation of 4 dofmaps
Computing representation of 1 coordinate mappings
Computing representation of integrals
Computing uflacs representation
Computing representation of forms
Compiler stage 2 finished in 5.9483 seconds.
Compiler stage 3: Optimizing intermediate representation
--------------------------------------------------------
Skipping optimizations, add -O to optimize
Compiler stage 3 finished in 0.000283957 seconds.
Compiler stage 4: Generating code
---------------------------------
Generating code for 4 element(s)
Generating code for 4 dofmap(s)
Generating code for 1 coordinate_mapping(s)
Generating code for integrals
Generating code from ffc.uflacs representation
Traceback (most recent call last):
File "/home/mh1714/ffc-tsfc/bin/ffc", line 11, in <module>
load_entry_point('FFC', 'console_scripts', 'ffc')()
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/__main__.py", line 237, in main
code_h, code_c = compile_ufl_data(ufd, prefix, parameters)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/__main__.py", line 97, in compile_ufl_data
parameters=parameters)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/compiler.py", line 141, in compile_form
prefix, parameters, jit)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/compiler.py", line 198, in compile_ufl_objects
code = generate_code(oir, parameters)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/codegeneration.py", line 86, in generate_code
for ir in ir_integrals]
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/codegeneration.py", line 319, in _generate_integral_code
code = r.generate_integral_code(ir, ir["prefix"], parameters)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/uflacs/uflacsgenerator.py", line 39, in generate_integral_code
uflacs_code = generate_tabulate_tensor_code(ir, prefix, parameters)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/uflacs/uflacsgenerator.py", line 63, in generate_tabulate_tensor_code
parts = ig.generate()
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/uflacs/generation/integralgenerator.py", line 103, in generate
body += self.generate_unstructured_partition(num_points, "piecewise")
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/uflacs/generation/integralgenerator.py", line 409, in generate_unstructured_partition
num_points)
File "/home/mh1714/ffc-tsfc/src/ffc/ffc/uflacs/generation/integralgenerator.py", line 351, in generate_partition
vops = [vaccesses[op] for op in v.ufl_operands]
KeyError
^C: <exception str() failed>
Comments (10)
-
reporter -
- changed milestone to 2017.1
- changed component to uflacs-repr
-
assigned issue to
-
Thanks for the report. The failing code has been rewritten some in a branch awaiting the release before merging, but fails at the equivalent line. I don't understand how it can happen, so will have to do some digging when I get time.
-
Does it work with tsfc through ffc?
-
reporter Yes, it does work with TSFC through FFC. IIRC it also used to work with UFLACS.
-
reporter git bisect blames 2e09973f344c8cbd3aa908218d12d3163cfff832.
-
Thanks. Although that commit doesn't seem to change anything...
-
- marked as blocker
-
-
- changed status to resolved
- Log in to comment