optimizing library functions like pow,exp

[pr_ate]

Hi all,
I'm using ifort & cc compilers for my application on the cluster.Through profiling I found out that function like log,pow,exp are taking a lot of time.I've downloaded Intel MKL cluster library.
Can I change my library to MKL library without changing source code?Or is there any other way to use optimized library functions other than the default one,with or without changing source code?
thanx in advance

[TimP]

I'm not clear on what you have in mind. If it related to MKL, which I doubt, you could post on the MKL forum. Recent versions of ifort and icc have incorporated SSE "vectorization" by automatically calling the short vector library, or in-lining for IPF. Options such as -vec_report3 or -opt_report will give the compilers reasons for not employing such optimizations.