Fortran - Open MP

Files
file	mp_elec.f90
	Electron model from COPS test set.
file	mp_elec2.f90
	Electron model from COPS test set, this time with a Hessian.
file	mp_leastsq.f90
	We solve the following nonlinear least squares model:
file	mp_leastsq11.f90
	The model is similar to leastsq, but we have added non-binding bounds on the res-variables so they cannot be eliminated and moved to the post-triangle. The efficiency bottlenecks have therefore been moved to other parts of CONOPT.
file	mp_leastsq12.f90
	The model is similar to leastsq, but we have added non-binding bounds on the res-variables so they cannot be eliminated and moved to the post-triangle. The efficiency bottlenecks have therefore been moved to other parts of CONOPT.
file	mp_leastsq13.f90
	The model is similar to leastsq, but we have added non-binding bounds on the res-variables so they cannot be eliminated and moved to the post-triangle. The efficiency bottlenecks have therefore been moved to other parts of CONOPT.
file	mp_leastsq14.f90
	This model is an extension of leastsq.f90 in which we test that the solution with 1, 2, and 4 threads give the same solution when threadc is defined as 4.
file	mp_leastsq2.f90
	This model is similar to leastsq. The key difference is that we supply a callback routine that can compute 2nd derivatives of the model. However, we only include part of the 2nd derivatives corresponding to the direct objective terms, res(i)**2. The terms from b(i,j)*x(j)**2 are ignored in the 2nd derivatives. CONOPT will not notice the incorrect derivatives but it may converge more slowly.
file	mp_leastsq5.f90
	This model is similar to leastsq, but this time we define 2nd order information in the form of a 2DDir routine.
file	mp_lincns.f90
	Large Linear Dense CNS model. Used to test the inversion routine for a very dense model.
file	mp_lincns2.f90
	Large Linear Dense CNS model. Used to test the inversion routine for a very dense model. The model is similar to Lincns.f90, but many nonzeros are so small that they are removed by the factorization routine. The factorization will therefore have to handle many fairly large blocks with many pivots and streamlining the updates is a challenge.
file	mp_pinthread4.f90
	This is a CONOPT implementation of the Pindyck model from the GAMS model library. The implementation is similar to the one in pindyck.f90, but this time we first solve the model for a number of periods in parallel, using periods between 21 and 40. We have three loops over periods: First loop is done sequentially The second loop is done in parallel using a static schedule and The third loop is done in parallel using a dynamic schedule.
file	mp_polygon.f90
	Polygon model from COPS test set.
file	mp_polygon2.f90
	Polygon model from COPS test set. The model is run sequentially and with various number of threads but with ThreadC defined. The models with various number of threads should produce identical solutions and we test this.
file	mp_qpbandb.f90
	QPBandB: Multi Threaded implementation of QP4 MINLP model The implementation shows how to implement a parallel Branch & Bound procedure that still is deterministic independent of the order in which the sub-models are solved (within the framework defined here with Work-pools).
file	mp_rosex.f90
	Extended Rosenbrock function.

Detailed Description

Examples written in Fortran that use Open MP for multithreading.