public inbox for glibc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug libc/2445] New: Bug in powl or expl?
@ 2006-03-11 19:13 john at johnmaddock dot co dot uk
2006-04-25 20:33 ` [Bug libc/2445] " drepper at redhat dot com
2010-06-01 3:31 ` [Bug math/2445] " pasky at suse dot cz
0 siblings, 2 replies; 3+ messages in thread
From: john at johnmaddock dot co dot uk @ 2006-03-11 19:13 UTC (permalink / raw)
To: glibc-bugs
I have a strange problem with powl that's very specific to AMD64 SMP machines
(Suse Linux 9 on HP 2x Opteron, the actual machine is one of the HP testdrive
machines: td190.testdrive.hp.com, so you can even try it on the same machine if
you need to). Basically I have an expression that should underflow to zero, but
which yields a non-zero result *only* on the second or subsequent execution of
the expression. The gcc version info is:
maddock@td190:~/libs/math/test> g++ -v
Reading specs from /usr/lib64/gcc-lib/x86_64-suse-linux/3.3.3/specs
Configured with: ../configure --enable-threads=posix --prefix=/usr --with-local-
prefix=/usr/local --infodir=/usr/share/info --mandir=/usr/share/man --enable-lan
guages=c,c++,f77,objc,java,ada --disable-checking --libdir=/usr/lib64 --enable-l
ibgcj --with-gxx-include-dir=/usr/include/g++ --with-slibdir=/lib64 --with-syste
m-zlib --enable-shared --enable-__cxa_atexit x86_64-suse-linux
Thread model: posix
gcc version 3.3.3 (SuSE Linux)
glibc version is 20040214, not the latest by any means, but unfortunately I'm
unable to test a newer version on this machine.
The code to reproduce the issue is:
#include <math.h>
#include <iostream>
template <class T, class L>
T regularised_gamma_prefix(T a, T z, const L&)
{
using namespace std;
T agh = a + 4.6875L - 0.5L;
//T prefix;
//T d = ((z - a) - 4.6875L + 0.5L) / agh;
return powl((z * expl((a-z)/a)) / agh, a);
}
template <class T>
void test(T)
{
T a = 1858.277099609375;
T z = 185827.703125;
T r;
for(int i = 0; i < 2; ++i)
{
r = regularised_gamma_prefix(
a, z, int());
std::cout << r << " " << i << std::endl;
}
}
int main()
{
test((long double)(0));
return 0;
}
which when run yields:
0 0
4.00731e-560 1
and on every other platform I've tried yields as expected:
0 0
0 1
My apologies if this has been fixed or reported already, but I haven't been able
to find anything similar reported.
Thanks in advance for any insights you may have,
Regards,
John Maddock.
--
Summary: Bug in powl or expl?
Product: glibc
Version: unspecified
Status: NEW
Severity: normal
Priority: P2
Component: libc
AssignedTo: drepper at redhat dot com
ReportedBy: john at johnmaddock dot co dot uk
CC: glibc-bugs at sources dot redhat dot com
GCC build triplet: Linux 2.6.5-7.252-smp x86_64
GCC host triplet: Linux 2.6.5-7.252-smp x86_64
GCC target triplet: Linux 2.6.5-7.252-smp x86_64
http://sourceware.org/bugzilla/show_bug.cgi?id=2445
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug libc/2445] Bug in powl or expl?
2006-03-11 19:13 [Bug libc/2445] New: Bug in powl or expl? john at johnmaddock dot co dot uk
@ 2006-04-25 20:33 ` drepper at redhat dot com
2010-06-01 3:31 ` [Bug math/2445] " pasky at suse dot cz
1 sibling, 0 replies; 3+ messages in thread
From: drepper at redhat dot com @ 2006-04-25 20:33 UTC (permalink / raw)
To: glibc-bugs
------- Additional Comments From drepper at redhat dot com 2006-04-25 20:33 -------
If you have accuracy problems with libm, send a patch. Until then the bugs is
suspended.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |SUSPENDED
http://sourceware.org/bugzilla/show_bug.cgi?id=2445
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug math/2445] Bug in powl or expl?
2006-03-11 19:13 [Bug libc/2445] New: Bug in powl or expl? john at johnmaddock dot co dot uk
2006-04-25 20:33 ` [Bug libc/2445] " drepper at redhat dot com
@ 2010-06-01 3:31 ` pasky at suse dot cz
1 sibling, 0 replies; 3+ messages in thread
From: pasky at suse dot cz @ 2010-06-01 3:31 UTC (permalink / raw)
To: glibc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
Component|libc |math
http://sourceware.org/bugzilla/show_bug.cgi?id=2445
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2010-06-01 3:31 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2006-03-11 19:13 [Bug libc/2445] New: Bug in powl or expl? john at johnmaddock dot co dot uk
2006-04-25 20:33 ` [Bug libc/2445] " drepper at redhat dot com
2010-06-01 3:31 ` [Bug math/2445] " pasky at suse dot cz
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).