FastSquareRoot分析和总结分析和总结.pdfVIP

Fast Square Root From ILabWiki Jump to: navigation , search Contents [ hide ] 1 C++Source Codefor Computing Fast Square Root Approximations with Floating and Double Precision Floating Points o 1.1 Heading o 1.2 Method using Log Base 2 Approximation o 1.3 Method using Log Base 2 Approximation With One Extra Babylonian Steps o 1.4 Method using Log Base 2 Approximation With Two Extra Babylonian Steps o 1.5 Method using The Quake 3 Walsh Method o 1.6 Setting a Default Function 2 Comments? Suggestions? C++ Source Code for Computing Fast Square Root Approximations with Floating and Double Precision Floating Points Here are listed a few different methods for computing a square root faster than the standard methods used with gcc and libm. Many of these methods are discussed in greater detail on the Wikipedia in Methods of computing square roots . Heading An optional header can be put in place if you implement all of these so that you can describe them to users who look at your source. // ##################################################################### # // ##### Fast Square Root functions // ##################################################################### # /* These are several methods for computing a quick square root approximation. The order is by decreasing speed / increasing accuracy Name : opperation time relative to sqrt() fastSqrt_2 : 1/5 fastSqrt_Bab : 1/2 fastSqrt_Q3 : 1/2 fastSqrt_Bab_2 : 2/3 Details: fastSqrt_2 - This is a simple log base 2 approximation using a union to int and 3 simple arith operations. fastSqrt_Bab - This is the fastSqrt_2 with a Babylonian step a