On Tue, Feb 21, 2012 at 4:58 AM, Toon Knapen <toon.knapen@gmail.com> wrote:
On Tue, Feb 21, 2012 at 7:04 AM, Ben Robinson <cppmaven@gmail.com> wrote:
<snip> Does this pique anyone's interest?
Can you elaborate on how this affects performance for integer arithmetic ?
There are two components to the library, overflow detection, and overflow handling. Overflow handling is policy dependent, but overflow detection is built-in to the library, and a key feature.
The overflow detection routines were meta-programmed with respect to the data types involved in the operation, in order to maximize performance. Template specializations are provided for each detection routines, which eliminate unnecessary operations based on both the signededness, and the ranges on both data types involved in the operation. For example, detect_addition_overflow performs only a single subtract and a single compare if both data types are unsigned. If both data types are signed, it will perform a maximum of one subtract, and either two or three compares. Because the detection routines are built into the overloaded math operators, and does not require one to rewrite existing code to make use of the detection, a developer could use these routines in a debug build, and then simply supply the ignore_overflow policy for a release build, which will inline out to zero overhead. Or even typedef to the library data types for debug, and typedef to PODs for release. The usage is exactly that of PODs. In short, performance was a key consideration in the design and implementation. Thank you, Ben Robinson
thanks, toon
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost