Gentoo Archives: gentoo-amd64

From: Mark Knecht <markknecht@×××××.com>
To: gentoo-amd64@l.g.o
Subject: Re: [gentoo-amd64] Possible Math Problem - Request For Verification [Solved]
Date: Sun, 26 Jun 2011 03:55:23
Message-Id: BANLkTim7kf_=YKBvvZ5w4Pq67rtCgVfXyw@mail.gmail.com
In Reply to: Re: [gentoo-amd64] Possible Math Problem - Request For Verification [Solved] by Frank Peters
1 On Sat, Jun 25, 2011 at 8:44 PM, Frank Peters <frank.peters@×××××××.net> wrote:
2 > On Sat, 25 Jun 2011 13:41:11 -0400
3 > Frank Peters <frank.peters@×××××××.net> wrote:
4 >
5 >>
6 >> /tmp/fp-test-results/clib_DP.output: ucbtest UCBFAIL in cabsd at line 701 for double
7 >>
8 >
9 > The culprit seems to be GCC optimization.  If I run the test with either "-O0"
10 > or "-O1" flags I can eliminate the cabsd failure.  Using "-O2" or "-O3" will
11 > result in the cabsd error.
12 >
13 > However, I've used "-O2" previously and had no problems with this test.  Possibly,
14 > some of these new LTO and GRAPHITE capabilities of GCC are to blame, even though
15 > I do not compile the ucbtest with either LTO or GRAPHITE enabled.  But GCC has itself
16 > been compiled using LTO and GRAPHITE.
17 >
18 > Anyway, thanks for all who actually ran the test on their machines.  I was
19 > thinking of filing bug reports with GLIBC and GCC and that would have turned
20 > out to be foolish.  I did check the Changelogs for GLIBC and there doesn't seem
21 > to have been any modification of the cabs() code over the last several versions.
22 >
23 > Frank Peters
24
25 I'm using -O2 here on all my machines. Certainly it isn't that option
26 that causes a problem for everyone. Sounds like something specific to
27 your processor revision.
28
29 Take care,
30 Mark

Replies

Subject Author
Re: [gentoo-amd64] Possible Math Problem - Request For Verification [Solved] Frank Peters <frank.peters@×××××××.net>