Re: double to string causes divide by zero!

• Subject: Re: double to string causes divide by zero!
• From: Doug Zwick <email@hidden>
• Date: Fri, 27 Feb 2004 16:27:31 -0700

This shouldn't make any difference, as explained above. Nor would
declaring the 'double' as volatile, nor any other Java-language tricks.

The problem exhibited by the original code is not that there's a double or
any other apparent object being shared by two threads. Hence, any attempt
to fix the problem by synchronizing the threads is not addressing the real
problem.

The problem appears to be that something in the JVM/JITC or other
implementation is incorrectly sharing some data between two threads, even
though it shouldn't be. That implies a bug in the JVM or JITC, and since
the bug only appears on a G5, which uses 64-bit opcodes for long division
(see the archives), that's where I'd place my money.

The failure is due to corruption of the hardware's floating point control
word (FPCW). The Java implementation requires that all floating point
exceptions (such as overflow, underflow and division by zero) be
disabled. If the FPCW gets changed to enable some of those exceptions,
Java floating point operations can fail. Moreover, they can fail in
surprising places, such as in converting a floating point value to a
String for display. (That one is because the Java conversion code
determines the sign of a value to be converted by dividing it by zero,
since that is one of the few ways to distinguish +0 from -0 when using
IEEE arithmetic.)

• Prev by Date: Re: Java 1.4.2 and Wacom tablets?
• Next by Date: Re: Java 1.4.2 and Wacom tablets?
• Previous by thread: Re: double to string causes divide by zero!
• Next by thread: Re: double to string causes divide by zero!
• Index(es):
  • Date
  • Thread