Let's say we have a very difficult task that can be effectively parallelized. How can we allocate all or almost all available processor power for this task?
The task can be varied, and the iterative generation of the Fibonacci number, which stores the recorded numbers, will be just one example.
You may have provided too little information about your target environment.
RTOS , , , , , ISR . , , .
RTOS ( ), , (.. , ), , . , Windows 100% , . , - , .
- , "" . ( ) .
. , , , ( "" Linux) .
, , .
:
1/ , , USB, CAN, SPI - , ( Processort), , . ( , ) . , (, ...)
2/ , ( , , ), , , !
, , , ... , "" !
, .
.
Disable each disk access and other IO lock caused by this or another process. If you need to do this, do it late, release it. Using multiple cores in parallel threads and establishing a dependency between the processor or communication between processors can help - the latter is likely to be OS specific.
Set the process priority so that it gets bigger.
Source: https://habr.com/ru/post/1746656/More articles:compile with -ansi -pedantic -Wall automatically with gcc - cAre Magento core questions related to core issues? - magentoFacebook multi-friend-selector + new javascript API = BROKEN? - facebookIs there a better way to get values ββfrom a table row? - jqueryHow to add a new page to the Lift frame - htmlSmart ways to implement various data structures in C structures and data that should be used more often - cHow to update MKAnnotation location with new coordinates? - iphonevisually show diff output in a java program - javaiPad translucent status bar? - ipadHow to capture live broadcast from Windows Media Server 2008 - c #All Articles