# source:rtems/testsuites/samples/paranoia/paranoia.scn@a136fb6

4.11
Last change on this file since a136fb6 was a136fb6, checked in by Joel Sherrill <joel.sherrill@…>, on Nov 15, 2013 at 5:53:36 PM

• Property mode set to `100644`
File size: 5.3 KB
Line
1
2
3*** PARANOIA TEST ***
4paranoia version 1.1 [cygnus]
5Program is now RUNNING tests on small integers:
6TEST: 0+0 != 0, 1-1 != 0, 1 <= 0, or 1+1 != 2
7PASS: 0+0 != 0, 1-1 != 0, 1 <= 0, or 1+1 != 2
8TEST: 3 != 2+1, 4 != 3+1, 4+2*(-2) != 0, or 4-3-1 != 0
9PASS: 3 != 2+1, 4 != 3+1, 4+2*(-2) != 0, or 4-3-1 != 0
10TEST: -1+1 != 0, (-1)+abs(1) != 0, or -1+(-1)*(-1) != 0
11PASS: -1+1 != 0, (-1)+abs(1) != 0, or -1+(-1)*(-1) != 0
12TEST: 1/2 + (-1) + 1/2 != 0
13PASS: 1/2 + (-1) + 1/2 != 0
14TEST: 9 != 3*3, 27 != 9*3, 32 != 8*4, or 32-27-4-1 != 0
15PASS: 9 != 3*3, 27 != 9*3, 32 != 8*4, or 32-27-4-1 != 0
16TEST: 5 != 4+1, 240/3 != 80, 240/4 != 60, or 240/5 != 48
17PASS: 5 != 4+1, 240/3 != 80, 240/4 != 60, or 240/5 != 48
18-1, 0, 1/2, 1, 2, 3, 4, 5, 9, 27, 32 & 240 are O.K.
19
22Closest relative separation found is U1 = 1.1102230e-16 .
23
25 confirms closest relative separation U1 .
27TEST: Radix is too big: roundoff problems
28PASS: Radix is too big: roundoff problems
29TEST: Radix is not as good as 2 or 10
30PASS: Radix is not as good as 2 or 10
31TEST: (1-U1)-1/2 < 1/2 is FALSE, prog. fails?
32PASS: (1-U1)-1/2 < 1/2 is FALSE, prog. fails?
33TEST: Comparison is fuzzy,X=1 but X-1/2-1/2 != 0
34PASS: Comparison is fuzzy,X=1 but X-1/2-1/2 != 0
35The number of significant digits of the Radix is 53.000000 .
36TEST: Precision worse than 5 decimal figures
37PASS: Precision worse than 5 decimal figures
38TEST: Subtraction is not normalized X=Y,X+Z != Y+Z!
39PASS: Subtraction is not normalized X=Y,X+Z != Y+Z!
40Subtraction appears to be normalized, as it should be.
41Checking for guard digit in *, /, and -.
42TEST: * gets too many final digits wrong.
43
44PASS: * gets too many final digits wrong.
45
46TEST: Division lacks a Guard Digit, so error can exceed 1 ulp
47or  1/3  and  3/9  and  9/27 may disagree
48PASS: Division lacks a Guard Digit, so error can exceed 1 ulp
49or  1/3  and  3/9  and  9/27 may disagree
50TEST: Computed value of 1/1.000..1 >= 1
51PASS: Computed value of 1/1.000..1 >= 1
52TEST: * and/or / gets too many last digits wrong
53PASS: * and/or / gets too many last digits wrong
54     *, /, and - appear to have guard digits, as they should.
55Checking rounding on multiply, divide and add/subtract.
56TEST: X * (1/X) differs from 1
57PASS: X * (1/X) differs from 1
58* is neither chopped nor correctly rounded.
59/ is neither chopped nor correctly rounded.
65Sticky bit used incorrectly or not at all.
66TEST: lack(s) of guard digits or failure(s) to correctly round or chop
67(noted above) count as one flaw in the final tally below
68ERROR: Severity: FLAW:  lack(s) of guard digits or failure(s) to correctly round or chop
69(noted above) count as one flaw in the final tally below.
70PASS: lack(s) of guard digits or failure(s) to correctly round or chop
71(noted above) count as one flaw in the final tally below
72
73Does Multiplication commute?  Testing on 20 random pairs.
74     No failures found in 20 integer pairs.
75
76Running test of square root(x).
77TEST: Square root of 0.0, -0.0 or 1.0 wrong
78PASS: Square root of 0.0, -0.0 or 1.0 wrong
79Testing if sqrt(X * X) == X for 20 Integers X.
80Test for sqrt monotonicity.
81sqrt has passed a test for Monotonicity.
82Testing whether sqrt is rounded or chopped.
83Square root is neither chopped nor correctly rounded.
84Observed errors run from -5.0000000e-01 to 5.0000000e-01 ulps.
85TEST: sqrt gets too many last digits wrong
86PASS: sqrt gets too many last digits wrong
87Testing powers Z^i for small Integers Z and i.
88... no discrepancies found.
89
90Seeking Underflow thresholds UfThold and E0.
91Smallest strictly positive number found is E0 = 4.94066e-324 .
92Since comparison denies Z = 0, evaluating (Z + Z) / Z should be safe.
93What the machine gets for (Z + Z) / Z is  2.00000000000000000e+00 .
94This is O.K., provided Over/Underflow has NOT just been signaled.
95Underflow is gradual; it incurs Absolute Error =
96(roundoff in UfThold) < E0.
97The Underflow threshold is 2.22507385850720188e-308,  below which
98calculation may suffer larger Relative error than merely roundoff.
99Since underflow occurs below the threshold
100UfThold = (2.00000000000000000e+00) ^ (-1.02200000000000000e+03)
101only underflow should afflict the expression
102        (2.00000000000000000e+00) ^ (-2.04400000000000000e+03);
103actually calculating yields: 0.00000000000000000e+00 .
104This computed value is O.K.
105
106Testing X^((X + 1) / (X - 1)) vs. exp(2) = 7.38905609893065218e+00 as X -> 1.
108Testing powers Z^Q at four nearly extreme values.
109 ... no discrepancies found.
110
111Searching for Overflow threshold:
112This may generate an error.
113Can `Z = -Y' overflow?
114Trying it on Y = -inf .
115Seems O.K.
116Overflow threshold is V  = 1.79769313486231571e+308 .
117Overflow saturates at V0 = inf .
118No Overflow should be signaled for V * 1 = 1.79769313486231571e+308
119                           nor for V / 1 = 1.79769313486231571e+308 .
120Any overflow signal separating this * from the one
121above is a DEFECT.
122
123
124What message and/or values does Division by Zero produce?
125    Trying to compute 1 / 0 produces ...  inf .
126
127    Trying to compute 0 / 0 produces ...  nan .
128
129The number of  FLAWs  discovered =           1.
130
131The arithmetic diagnosed seems Satisfactory though flawed.
132END OF TEST.
133*** END OF PARANOIA TEST ***
Note: See TracBrowser for help on using the repository browser.