Re:Nearly All Binary Searches and Mergesorts are Broken (2006)
by wtallis on Thursday July 09, 2015 @05:22PM
(#9860778)
Yeah, it's pretty much a given that checking for overflow won't be the bottleneck when you're doing pointer arithmetic to traverse a large data structure. The memory accesses will dominate over the cost of the highly-predictable branches.
PS: There are several models you can test yourself. http://blog.regehr.org/archives/1154
edit: http://resources.sei.cmu.edu/asset_files/TechnicalNote/2010_... Has actual benchmarks in the 6-8% range.