I use an engineering modeling application which is able to run (using Argonne National Labs MPI-2) on up to 8 parallel processes (the vendor has limited model execution to 8 cores or threads). My question is whether an 8-core CPU, such as a Xeon E5-2687 (8 cores/16 threads) would run this application faster than a 4 core/8 thread Intel i5 CPU or 6 core/12 thread i7, assuming each CPU option is operating at the same clock speed. i.e., will having an actual 8 core CPU be faster than running less physical cores, but with hyperthreading?
I5 is a 4 core/4 thread processor, you seem to have that confused.
The program should run better on the 8 core xeon if it uses 8 threads, having 8 real cores is going to perform better than a 4 core with hyperthreading. Evidence of this can be seen where the 4 core I5 out performs the I3 which has 2 cores/4 threads.