Hi all !
I have a program that connects to a DB and recovers many registers from it.
I have to check for every row of the result set that all the fields have the correct data. I think I could do this using parallelism.
Can I achieve a validation algorithm with CUDA?
For example:
| Field 1 | Field 2 |Field 3 | Field 4 | Field 5 | Field 6 | Field 7 | … | Field n |
| 99 | 2 | 123 | 1 | 2 | 3 | 3 | … | X |
| 99 | 1 | 200 | 1 | 2 | 3 | 3 | … | X |
| 99 | 2 | 230 | 1 | 2 | 4 | 3 | … | X |
| 99 | 1 | 100 | 2 | 2 | 4 | 3 | … | X |
…
…
…
So I would like to check the first row (all its fields) to see if the values are in the correct range, then second row, and so on.
I was thinking using CUDA and each thread check one entire row… what do you think? I mean my result set is about 2 million registers, of course I could not copy all of them in the memory at once but I can do this a little set and then another part and so on…
Any advise is very welcome :D