# Review for the Final Exam

### Assignment :

Study prepare questions for Friday's review session

### New Fortran:

none

What to study for the exam? Start by looking at the review summaries for the first and
second exams. If you understand, all listed commands and concepts, you will be in very good
shape. I will not ask any questions on Fortran statements or Unix commands not listed here or in
the previous summaries. I might ask general questions on the related concepts of pipelining, code
testing, timing, and debugging.

I have constructed the exam based, entirely on the summary lists below, and related
material in the class notes and examples on the Web. I have not included questions based on any
material in the text not covered in the class notes.

The following is material covered since the last exam:

### Fortran Summary

Multidimensional Arrays, index notation and order of storage
EQUIVALENCE

INTERFACE structure and use to create vector valued functions and generic functions.

INTENT, intent(in), intent(out), intent(inout)

MATMUL intrinsic function

SIZE intrinsic function and its use to set array dimensions

DIM optional argument to array functions. Be able to use SIZE, SUM, MINVAL, or MAXVAL with DIM

SYSTEM_CLOCK function for elapsed **wall clock** time, and cautions in use

STATEMENT functions, inline substitution, knowledge of local variables

CONTAINS statement and internal subprograms , knowledge of local variables

COMMON blocks, INCLUDE and BLOCK DATA, what replaces them?

MODULE program units can also contain INTERFACES, SUBROUTINES, and FUNCTIONS

POINTER and TARGET attributes, => operator, relation of POINTER to EQUIVALENCE

### Unix

What is a shell script
-O option on f77

### Computational Applications

LU Factorization (what is it, but not how you do it): operation count for decomposition versus that for back
substitution
Matrix Notation: Storage of coefficients for a system of linear equations

Least Squares Fits: Know how to obtain the least squares linear equations, used to get coefficients of a "best
fit" polynomial. What is the error function? How do you find its minimum?

NETLIB: LINPACK, LAPACK

Numerical integration with Trapezoidal and Simpson's methods. Which is generally more accurate?

Numerical derivatives, first and second derivatives based on quadratic fit to 3 points.

Numerical solution of ODE, Definition and application of Crank-Nicholson, Forward (Explicit) and
Backward (Implicit) Euler methods. Names of fourth order accurate methods: Runge-Kutta, Adams-Bashford, Adams-Moulton. Error and stability behavior of all listed methods.

Secant method

**
**

Written and Maintained by John Mahaffy : jhm@psu.edu