gives the last element of CUDAFoldList[f,x,list].
Details and Options
- The CUDALink application must be loaded using Needs["CUDALink`"].
- CUDAFold works only on vectors.
- CUDAFold does not work on fixed vector structure types like "Float", "Integer32", ….
- CUDAFold does not work on symbolic input.
- CUDAFold is equivalent to Fold.
- The value for f can be either Max, Min, Plus, Minus, or Times.
Basic Examples (3)
This folds an input list with the Plus operator and initial value 0:
The above is equivalent to Fold:
CUDAFold operates on input CUDAMemory:
This folds the CUDAMemory as before:
Note that the output type is the same as the input—that is, if the input is a Real valued list, then the output is a real number.
Compared to Fold, only a few operations are supported by CUDAFold:
Wolfram Research (2010), CUDAFold, Wolfram Language function, https://reference.wolfram.com/language/CUDALink/ref/CUDAFold.html.
Wolfram Language. 2010. "CUDAFold." Wolfram Language & System Documentation Center. Wolfram Research. https://reference.wolfram.com/language/CUDALink/ref/CUDAFold.html.
Wolfram Language. (2010). CUDAFold. Wolfram Language & System Documentation Center. Retrieved from https://reference.wolfram.com/language/CUDALink/ref/CUDAFold.html