\(\newcommand{\W}[1]{ \; #1 \; }\) \(\newcommand{\R}[1]{ {\rm #1} }\) \(\newcommand{\B}[1]{ {\bf #1} }\) \(\newcommand{\D}[2]{ \frac{\partial #1}{\partial #2} }\) \(\newcommand{\DD}[3]{ \frac{\partial^2 #1}{\partial #2 \partial #3} }\) \(\newcommand{\Dpow}[2]{ \frac{\partial^{#1}}{\partial {#2}^{#1}} }\) \(\newcommand{\dpow}[2]{ \frac{ {\rm d}^{#1}}{{\rm d}\, {#2}^{#1}} }\)

speed_cppadcg

View page source

Speed Test Derivatives Using Cppadcg

Purpose

CppAD has a set of speed tests that are used to compare Cppadcg with other AD packages. This section links to the source code the Cppadcg speed tests (any suggestions to make the Cppadcg results faster are welcome).

Building Tests

Starting in the Build Directory , execute the following commands

      cd build/speed/cppadcg

      ./ speed_cppadcg test seed option_list

A message saying that the sizes are incorrect will be printed. In addition, it will say that source code with the correct sizes has been created. If you then execute

make check_speed_cppadcg VERBOSE =1

It will build speed_cppadcg with the proper sizes and run its correctness tests.

Running Tests

You can then run the corresponding speed tests with the either of the following commands:

      ./ speed_cppadcg speed seed onetape

      ./ speed_cppadcg speed seed onetape optimize

where seed is a positive integer. See speed_main for more options.

Contents

• cppadcg Speed: Gradient of Determinant by Minor Expansion

• Cppadcg Speed: Gradient of Determinant Using Lu Factorization

• Cppadcg Speed: Matrix Multiplication

• Cppadcg Speed: Ode

• Cppadcg Speed: Second Derivative of a Polynomial

• Cppadcg Speed: Sparse Hessian

• Cppadcg Speed: Sparse Jacobian