qmcpack
qmcpack copied to clipboard
Unit test failure ReferencePoints
Failure happened on Summit.
String comparison is not reliable. See below f1m comparing -0 and 0
ReferencePoints::Description
-------------------------------------------------------------------------------
/ccs/home/yeluo/opt/qmcpack/src/Estimators/tests/test_ReferencePoints.cpp:209
...............................................................................
/ccs/home/yeluo/opt/qmcpack/src/Estimators/tests/test_ReferencePoints.cpp:273: FAILED:
CHECK( ostr_stream.str() == expected_description )
with expansion:
" reference_points
a1: 3.373161077 3.373161077 0
a2: 0 3.373161077 3.373161077
a3: 3.373161077 0 3.373161077
cmmm: -3.373161077 -3.373161077 -3.373161077
cmmp: 0 -3.373161077 0
cmpm: -3.373161077 0 0
cmpp: 0 0 3.373161077
cpmm: 0 0 -3.373161077
cpmp: 3.373161077 0 0
cppm: 0 3.373161077 0
cppp: 3.373161077 3.373161077 3.373161077
f1m: -1.686580539 -1.686580539 -0
f1p: 1.686580539 1.686580539 0
f2m: -0 -1.686580539 -1.686580539
f2p: 0 1.686580539 1.686580539
f3m: -1.686580539 -0 -1.686580539
f3p: 1.686580539 0 1.686580539
ion1: 0 0 0
ion2: 1.686580539 1.686580539 1.686580539
r1: 3.373161077 3.373161077 0
r2: 0 3.373161077 3.373161077
r3: 3.373161077 0 3.373161077
zero: 0 0 0
end reference_points
"
==
" reference_points
a1: 3.373161077 3.373161077 0
a2: 0 3.373161077 3.373161077
a3: 3.373161077 0 3.373161077
cmmm: -3.373161077 -3.373161077 -3.373161077
cmmp: 0 -3.373161077 0
cmpm: -3.373161077 0 0
cmpp: 0 0 3.373161077
cpmm: 0 0 -3.373161077
cpmp: 3.373161077 0 0
cppm: 0 3.373161077 0
cppp: 3.373161077 3.373161077 3.373161077
f1m: -1.686580539 -1.686580539 0
f1p: 1.686580539 1.686580539 0
f2m: 0 -1.686580539 -1.686580539
f2p: 0 1.686580539 1.686580539
f3m: -1.686580539 0 -1.686580539
f3p: 1.686580539 0 1.686580539
ion1: 0 0 0
ion2: 1.686580539 1.686580539 1.686580539
r1: 3.373161077 3.373161077 0
r2: 0 3.373161077 3.373161077
r3: 3.373161077 0 3.373161077
zero: 0 0 0
end reference_points
"
Should I just drop unit testing for the reports? I don't think its worth writing a parser to check this output.
Trim the first two rows to compare and also do the last row?
Still failing and f1m still differ as it was.
On Summit?
Yes
I'm going to go reproduce that you're probably right that matching so much output is a fools errand.