Bug 2574

Summary:	Incorrect return from float log10 (0X1.118C6318C6319P+0)
Product:	glibc	Reporter:	H.J. Lu <hjl.tools>
Component:	math	Assignee:	Ulrich Drepper <drepper.fsp>
Status:	RESOLVED INVALID
Severity:	normal	CC:	glibc-bugs
Priority:	P2	Flags:	fweimer: security-
Version:	unspecified
Target Milestone:	---
Host:	x86_64-unknown-linux-gnu	Target:	x86_64-unknown-linux-gnu
Build:	x86_64-unknown-linux-gnu	Last reconfirmed:

Description H.J. Lu 2006-04-14 19:21:05 UTC

Failure: Test: log10 (0X1.118C6318C6319P+0) == 0X1.D7C3996E6179DP-6
Result:
 is:          2.87942159920930862427e-02   0x1.d7c3b200000000000000p-6
 should be:   2.87941936403512954712e-02   0x1.d7c39a00000000000000p-6
 difference:  2.23517417907714843750e-08   0x1.80000000000000000000p-26
 ulp       :  12.0000
 max.ulp   :  0.0000

Comment 1 Ulrich Drepper 2006-04-16 02:45:21 UTC

libm bugs remain suspended until somebody comes up with a patch.

Comment 2 Joseph Myers 2012-02-26 17:26:48 UTC

When the argument is rounded to float (0X1.118C64P+0), the correctly rounded result becomes 0X1.D7C3B0P-6, so the error is 1ulp not 12ulp, and since there's already a test of log10f in the testsuite with a 2ulp error recorded in libm-test-ulps for x86_64, I don't think it's useful to add this test as well or keep a bug open for this 1ulp error.