JDK-6351144 : JCK14a Some api/java_lang StrictMath tests fail with 142_11b1 64bit JDK on Itanium2
  • Type: Bug
  • Component: core-libs
  • Sub-Component: java.lang
  • Affected Version: 1.4.2_11
  • Priority: P1
  • Status: Closed
  • Resolution: Fixed
  • OS: linux_redhat_3.0
  • CPU: itanium
  • Submitted: 2005-11-16
  • Updated: 2010-04-03
  • Resolved: 2005-12-09
The Version table provides details related to the release that this issue/RFE will be addressed.

Unresolved : Release in which this issue/RFE will be addressed.
Resolved: Release in which this issue/RFE has been resolved.
Fixed : Release in which this issue/RFE has been fixed. The release containing this fix may be available for download as an Early Access Release or a General Availability Release.

To download the current JDK release, click here.
Other
1.4.2_11 b02Fixed
Related Reports
Relates :  
JDK-4880538 - REGRESSION: 6 JCK14a api/java_lang/StrictMath tests fail on tiger
Relates :  
JDK-2131990 - REGRESSION: 6 JCK14a api/java_lang/StrictMath tests fail on tiger
Relates :  
JDK-6345520 - JCK14a Some api/java_lang StrictMath tests fail with 142_11b1 64bit JDK on Windows x86_64
Description
The following tests fail on Windows 2003 ES and RHEL3.0 AS on Itanium 2 system with java version 1.4.2_11-ea-b01 [ia64 bundles].

Bundles tested:
JDK win bundle: j2sdk1.4.2_11-windows-ia64.bin
JDK bundle: j2sdk-1_4_2_11-linux-ia64.bin
TestSuite: JCK-runtime-14a

Failed tests -
api/java_lang/StrictMath/desc.html#acos
api/java_lang/StrictMath/desc.html#asin
api/java_lang/StrictMath/desc.html#cos
api/java_lang/StrictMath/desc.html#log
api/java_lang/StrictMath/desc.html#sin
api/java_lang/StrictMath/desc.html#tan

NOTE: These tests pass with same configuration using 142_10b3 j2sdk ia64 bundle.
I also tested on Solaris_AMD with 142_11b1 64bit bundle and 64bit libraries - tests pass ok.
This is probably related to the 64 bit backport in 142_11.

Output of desc_acos.jtr

home/JDKS/142_11/j2sdk1.4.2_11/bin/java -Djava.awt.headless=false -Xfuture -Djava.security.policy=/.automount/carlow.ireland/root/export/home4/jck/JCK-runtime-14a/lib/jck.policy javasoft.sqe.tests.api.java.lang.StrictMath.acos_test
----------ref:(0/0)----------
----------log:(19/1435)----------
acos failed for 3fd91eb851eb8525 Expected 3ff2adde24ebedf7 Got: 3ff2adde24ebedf8
acos failed for 3fd9f3b645a1cac7 Expected 3ff273cbd844a711 Got: 3ff273cbd844a710
acos failed for 3fde9fbe76c8b441 Expected 3ff1264ce316587c Got: 3ff1264ce316587d
acos failed for 3fe1e353f7ced8e0 Expected 3fef48a38b561d62 Got: 3fef48a38b561d63
acos failed for 3fe1e76c8b43954a Expected 3fef43b2ac9dd368 Got: 3fef43b2ac9dd369
acos failed for 3fe24189374bc666 Expected 3feed67b3ade4070 Got: 3feed67b3ade4071
acos failed for 3fe30624dd2f1a46 Expected 3fede4941291c674 Got: 3fede4941291c673
acos failed for 3fe35c28f5c28ef8 Expected 3fed7914deb3b8dc Got: 3fed7914deb3b8dd
acos failed for 3fe3645a1cac07cc Expected 3fed6ec9be891e57 Got: 3fed6ec9be891e56
acos failed for 3fe3db22d0e55fce Expected 3fecd866c7d231fa Got: 3fecd866c7d231f9
acos failed for 3fe456041893743a Expected 3fec3a7e7e9ab1e6 Got: 3fec3a7e7e9ab1e5
acos failed for 3fe47ae147ae13f4 Expected 3fec0aa23b4885a5 Got: 3fec0aa23b4885a6
acos failed for 3fe6083126e9781e Expected 3fe9f6ce28d39dc8 Got: 3fe9f6ce28d39dc9
acos failed for 3fe649ba5e353ebe Expected 3fe99bf5d1b38113 Got: 3fe99bf5d1b38112
acos failed for 3fe672b020c49ae2 Expected 3fe962aeccd89ecd Got: 3fe962aeccd89ece
acos failed for 3fe7e353f7ced826 Expected 3fe74bdb9ddeb82b Got: 3fe74bdb9ddeb82c
acos failed for 3fe94395810623c2 Expected 3fe5254b1ddfcd52 Got: 3fe5254b1ddfcd53
acos failed 17 out of 2108 tests
STATUS:Failed. Check log
result: Failed. Check log
test result: Failed. Check log
Comments
EVALUATION The fix for bug# 4880538 works. As stated by ###@###.### "Generally the compiler options for the math library have to have a lower optimization level or other options to get the proper floating-point semantics implemented. My fix for 4880538 removes optimizations for all linux platforms. This should always be a safe change; although it won't help windows of course. IIRC, we had to fiddle with the windows Itanium options as well."
22-11-2005
SUGGESTED FIX See suggested fix section for bug# 4880538.
22-11-2005
EVALUATION As stated in the comments by ###@###.###, bug# 4880538 seems to identify a potential solution.
21-11-2005