Emulate “double” using 2 “float”s

by

double-float is a technique that uses pairs of single-precision numbers to achieve almost twice the precision of single precision arithmetic accompanied by a slight reduction of the single precision exponent range (due to intermediate underflow and overflow at the far ends of the range). The basic algorithms were developed by T.J. Dekker and William Kahan in the 1970s. Below I list two fairly recent papers that show how these techniques can be adapted to GPUs, however much of the material covered in these papers is applicable independent of platform so should be useful for the task at hand.

https://hal.archives-ouvertes.fr/hal-00021443
Guillaume Da Graça, David Defour
Implementation of float-float operators on graphics hardware,
7th conference on Real Numbers and Computers, RNC7.

http://andrewthall.org/papers/df64_qf128.pdf
Andrew Thall
Extended-Precision Floating-Point Numbers for GPU Computation.

More Related Contents:

  1. What is the most effective way for float and double comparison?
  2. How do I print a double value with full precision using cout?
  3. What is the difference between float and double?
  4. Efficient implementation of log2(__m256d) in AVX2
  5. C++ floating point precision [duplicate]
  6. In which order should floats be added to get the most precise result?
  7. Set back default floating point print precision in C++
  8. Matlab vs C++ Double Precision
  9. Is there a functional difference between “2.00” and “2.00f”?
  10. How to find number of words in a phrase with spaces removed by checking v. dictionary
  11. What is the efficient way to count set bits at a position or lower?
  12. Fast exact bigint factorial
  13. Generating combinations in c++
  14. Finding duplicates in O(n) time and O(1) space
  15. Floating point format for std::ostream
  16. How to get a random element from a C++ container?
  17. Why did Microsoft abandon long double data type? [closed]
  18. LRU cache design
  19. c++ sort keeping track of indices [duplicate]
  20. How do you print the EXACT value of a floating point number?
  21. How to detect double precision floating point overflow and underflow?
  22. Determining if a number is prime
  23. C- Floating point precision
  24. How to add two numbers without using ++ or + or another arithmetic operator
  25. Are there any better methods to do permutation of string?
  26. Reorder vector using a vector of indices [duplicate]
  27. Converting epoch time to “real” date/time
  28. What is the safe way to fill multidimensional array using std::fill?
  29. Is it more efficient to copy a vector by reserving and copying, or by creating and swapping? [duplicate]
  30. Extra leading zeros when printing float using printf?

Leave a Comment