ieee-754

Yes I'm aware of the IEEE-754 half-precision standard, and yes I'm aware of the work done in the field. Put very simply, I'm trying to save a simple floating point number (like 52.1 , or 1.25 ) in just 2 bytes. I've tried some implementations in Java and in C# but they ruin the input value by decoding a different number. You feed in 32.1 and after encode-decode you get 32.0985 . Is there ANY way I can store floating point numbers in just 16-bits without ruining the input value? Thanks very much. You could store three digits in BCD and use the remaining four bits for the decimal point position:

I have the following code, which writes 6 floats to disk in binary form and reads them back: #include <iostream> #include <cstdio> int main() { int numSegs = 2; int numVars = 3; float * data = new float[numSegs * numVars]; for (int i = 0; i < numVars * numSegs; ++i) { data[i] = i * .23; std::cout << data[i] << std::endl; } FILE * handle = std::fopen("./sandbox.out", "wb"); long elementsWritten = std::fwrite(data, sizeof(float), numVars*numSegs, handle); if (elementsWritten != numVars*numSegs){ std::cout << "Error" << std::endl; } fclose(handle); handle = fopen("./sandbox.out", "rb"); float *

I am wondering if the max float represented in IEEE 754 is: (1.11111111111111111111111)_b*2^[(11111111)_b-127] Here _b means binary representation. But that value is 3.403201383*10^38 , which is different from 3.402823669*10^38 , which is (1.0)_b*2^[(11111111)_b-127] and given by for example c++ <limits> . Isn't (1.11111111111111111111111)_b*2^[(11111111)_b-127] representable and larger in the framework? Does anybody know why? Thank you. The exponent 11111111 b is reserved for infinities and NaNs, so your number cannot be represented. The greatest value that can be represented in single

Does there exist an IEEE double x>0 such that sqrt(x*x) ≠ x , under the condition that the computation x*x does not overflow or underflow to Inf , 0 , or a denormal number? This is given that sqrt returns the nearest representable result, and so does x*x (both as mandated by the IEEE standard, "square root operation be calculated as if in infinite precision, and then rounded to one of the two nearest floating-point numbers of the specified precision that surround the infinitely precise result"). Under the assumption that if such doubles would exist, then there are probably examples close to 1,

I need to encode and decode IEEE 754 floats and doubles from binary in node.js to parse a network protocol. Are there any existing libraries that do this, or do I have to read the spec and implement it myself? Or should I write a C module to do it? Dobes Vandermeer Note that as of node 0.6 this functionality is included in the core library, so that is the new best way to do it. See http://nodejs.org/docs/latest/api/buffer.html for details. If you are reading/writing binary data structures you might consider using a friendly wrapper around this functionality to make things easier to read and

The ECMAScript specification for Math.pow has the following peculiar rule: If x < 0 and x is finite and y is finite and y is not an integer, the result is NaN. ( http://es5.github.com/#x15.8.2.13 ) As a result Math.pow(-8, 1 / 3) gives NaN rather than -2 What is the reason for this rule? Is there some sort of broader computer science or IEEEish reason for this rule, or is it just a choice TC39/Eich made once upon a time? Update Thanks to Amadan's exchanges with me, I think I understand the reasoning now. I would like to expand upon our discussion for the sake of posterity. Let's take the