CppADCodeGen Versions Save

Some of the highlights of this release:

• performance improvements in equation pattern detection
• bugfix in adding operations pointed by alias operations to the operation queue
• bugfix in printing floats/doubles in the source code when they can also be represented as integers
• bugfix in the generated forward_one and reverse_one functions when there are loops involved
• other minor improvements

Some of the highlights of this release:

• implemented random index patterns
• improved performance of the generated code for models with loops and atomics
• improved source code generation performance for models without loops and atomics
• minor improvements in printouts
• bugfix in hessian generation for models with loops when there is no second order information
• bugfix in loop tape creation (avoids potential NaNs)
• bugfix in equation pattern detection when there are several equations that do not match any pattern

Some of the highlights of this release:

• Added support for equation pattern detection (equations inside loops)
• Dynamic library creation can take advantage of equation patterns
• Added a new namespace (extra) for CppAD generic functions

Some of the highlights of this release:

• Atomic function support
• Improved Hessian sparsity evaluation for multiple equations/dependents at once