This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/18370] New: csqrt missing underflows
- From: "jsm28 at gcc dot gnu.org" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sourceware dot org
- Date: Sat, 02 May 2015 20:43:42 +0000
- Subject: [Bug math/18370] New: csqrt missing underflows
- Auto-submitted: auto-generated
https://sourceware.org/bugzilla/show_bug.cgi?id=18370
Bug ID: 18370
Summary: csqrt missing underflows
Product: glibc
Version: 2.21
Status: NEW
Severity: normal
Priority: P2
Component: math
Assignee: unassigned at sourceware dot org
Reporter: jsm28 at gcc dot gnu.org
Target Milestone: ---
The csqrt implementations in glibc can miss underflow exceptions when the
result becomes tiny in the course of multiplying by 0.5 (or other scaling down)
and that scaling happens to be exact (but the mathematical result is not
exact). E.g., for float on x86_64:
Failure: Real part of: csqrt_downward (-0x1.08e754p+16 - 0x7.ede03p-120 i):
Exception "Underflow" not set
Failure: Real part of: csqrt_downward (0xb.6554ep+0 + 0x1.247482p-124 i):
Exception "Underflow" not set
Failure: Real part of: csqrt_towardzero (0xb.6554ep+0 + 0x1.247482p-124 i):
Exception "Underflow" not set
--
You are receiving this mail because:
You are on the CC list for the bug.