CUDA Parallel Programming/Learn CUDA and Implementation with ANOVA
Learn CUDA and Implementation with ANOVA
CUDA Example 1
On this example bb.txt file is 5 MB and it contains total of 1649996 characters to be searched. Through CUDA I implement my search time as 184 ms. Related to size of the file efficient Block number is 4, and Thread number is 512. Also all found location of search string is stored in an Array for analyze further steps.
CUDA Example 2
This example shows that how a data divided into pieces to search the SearchString. Divided data is makes the improvement of string search rather than search on whole data, each code runs on divided data. On the example it finds the found number of strings in each piece of data.
CUDA Example 3
On this example b.txt file is 24 MB and it contains total of 24269185 characters to be searched. Through CUDA I implement my search time as 617 ms. These search time is takes much more time in C.
Related to size of the file efficient Block number is 48, and Thread number is 512. Also all found location of search string is stored in an Array for analyze further steps. On the Example 1, block number was 4, when the data size increase as parallel my number of threads as increase for to handle data size.
There are 2378185 number of search string is found and those searched strings location(index thats is located on the data text is stored in a file).
CUDA Example 4
These example runs on functions of ANOVA formula as parallel. It is efficient because data set that contains information related to each object increase CUDA make the implementation faster.
Our data members can be:
int data1_store[MEMBERNUM] = {7, 4, 6, 8, 6, 6, 2, 9 …...};
int data2_store[MEMBERNUM] = {5, 5, 3, 4, 4, 7, 2, 2 …...};
int data3_store[MEMBERNUM] = {2, 4, 7, 1, 2, 1, 5, 5 …...};
Our result find the relation between each member . By looking F-Table Value we can conclude those members interaction with each other.
CUDA Example 5
This example shows the found index location of search string in whole data set. On the example find locations are not in order because CUDA runs program as parallel that switch between search data set in each step. When I checked from the original dataset found index values are all correct.
