1 |
Am Sat, 25 Jun 2011 13:41:11 -0400 |
2 |
schrieb Frank Peters <frank.peters@×××××××.net>: |
3 |
|
4 |
> Hello, |
5 |
> |
6 |
[background information] |
7 |
|
8 |
Huh, I was unaware that GNU (since they're the authors of GCC/GLIBC) ever had |
9 |
problems in this area. Since I use my machine for numerical tasks, too, I was |
10 |
interested in seeing whether I would see the same problem. |
11 |
|
12 |
> The code will take several minutes to compile, execute, and complete. While |
13 |
> the test is running it will spit out its progress to stdout. Upon completion |
14 |
> a brief summary is given: |
15 |
> |
16 |
> UCBFAIL indicates problems! |
17 |
> /tmp/fp-test-results/ccos_DP.output: ucbtest UCBFAIL in COS(X) at line 444 for generic |
18 |
> /tmp/fp-test-results/clib_DP.output: ucbtest UCBFAIL in cabsd at line 701 for double |
19 |
> /tmp/fp-test-results/clib_DP.output: ucbtest UCBFAIL in powd at line 701 for double |
20 |
> /tmp/fp-test-results/clib_DP.output:UCBFAIL clib_DP.output , 25 out of 25 tests completed |
21 |
> /tmp/fp-test-results/csin_DP.output: ucbtest UCBFAIL in SIN(X) at line 444 for generic |
22 |
> |
23 |
> only 11 out of 14 show UCBPASS! |
24 |
|
25 |
I ran the tests without any modifications to ucbREADME/linux.sh and don't get |
26 |
the cabsd failure: |
27 |
|
28 |
UCBFAIL indicates problems! |
29 |
/tmp/fp-test-results/ccos_DP.output: ucbtest UCBFAIL in COS(X) at line 444 for generic |
30 |
/tmp/fp-test-results/clib_DP.output: ucbtest UCBFAIL in powd at line 701 for double |
31 |
/tmp/fp-test-results/clib_DP.output:UCBFAIL clib_DP.output , 25 out of 25 tests completed |
32 |
/tmp/fp-test-results/csin_DP.output: ucbtest UCBFAIL in SIN(X) at line 444 for generic |
33 |
|
34 |
only 11 out of 14 show UCBPASS! |
35 |
|
36 |
Out of curiosity, I ran the test two more times and got the same results. |
37 |
|
38 |
This machine runs on stable, with a slightly older Athlon64 X2. |
39 |
|
40 |
> There are expected failures in the trigonometric tests since trigonometric |
41 |
> performance is not specified in IEEE754/854. The failure in powd (double power |
42 |
> function) is also expected. But everything else should have passed. |
43 |
|
44 |
Then according to the results above my machine is apparently OK. (I guess I |
45 |
really should finally read about IEEE floating point more thoroughly, I've been |
46 |
planning to do that for a while now.) |
47 |
|
48 |
[...] |
49 |
> Frank Peters |
50 |
|
51 |
HTH |
52 |
-- |
53 |
Marc Joliet |
54 |
-- |
55 |
"People who think they know everything really annoy those of us who know we |
56 |
don't" - Bjarne Stroustrup |