How to reduce CPU time in file IO operations in OpenMP

Basically I get the same performance for both sereial and parallel computation. I do certian file read operations in the parallel processing.
Will the file reading be parallel or will the file be read serially eventhough it is in the parallel region.
Note: The same file is read by both the threads(processors).

It doesn’t sound to me like what you are doing is valid. OpenMP programs have a single address space. Each thread is doing the same i/o operations, so at a process level, you are doubling the amount of i/o. I would move the i/o operations out of the parallel regions. Or, manually break up the i/o using conditionals based on thread num, but then you may introduce race conditions on stdin/stdout, etc.