I worked on many problems related to data, and very often they come down to quickly and parallelly launching many implementations of intensive processor algorithms, such as Hamming / Change distance. Is that what CUDA would be useful for?
What data processing problems did you solve? Is there any real improvement over Intel's standard quad-core processor?
Chris
, . , CUDA/OpenCL . CUDA DSP (FFT, FIR) . - . , MKL OpenMP, , CUDA/OpenCL .
CUDA
-, SIGGRAPH '09 CUDA Vray Maya. 20 200 ? , .
, CUDA. , :
, . , , . , .
CUDA , FASTRA, , ( - !), .
, , - .
: , , . , :). , ...
parallelism: parallelism parallelism. GPU . , CPU , , (, 4 ). , GPU ALU. , parallelism (, , ), GPU ALU . - GPU, , .
Source: https://habr.com/ru/post/1747977/More articles:Sending an image to a faceless email address - objective-cDetect if two pngs are different - .netWhat is the difference between these codes? - javascriptWPF and very large images - imageGetting started with workflows in SharePoint 2010 - workflowPHP website source code - phpProblem behavior of Linq Union? - c #Why won't the warning method be detected? - objective-cAlternative for PHP mail - httpSometimes a big delay when using PHP mail () - httpAll Articles