|
From: Luigi B. <lui...@gm...> - 2007-11-06 10:26:47
|
On Tue, 2007-11-06 at 09:45 +0100, Bianchetti Marco wrote: > On the trunk using vc7 I get the following three errors in test suite: > > Testing Joint Calibration of an Heston Equity Processincl. Stochastic > Interest Rates via a Hull-White Model... > ./hybridhestonhullwhiteprocess.cpp(1418): error in > "HybridHestonHullWhiteProcessTest::testPseudoJointCalibration": Failed > to calibrate Heston Hull-White Model It happens depending on the evaluation date. The sample data are not very robust, I guess. Klaus? Any further insight? > Testing randomized lattice sequences, C up to dimension 30... > ./lowdiscrepancysequences.cpp(156): error in > "LowDiscrepancyTest::testRandomizedLattices": Lattice generatorC returns > a mean of 0.49985 with standard deviation -4.12168 in dimension 7 Strange. Not for the value as such (the check is for <= 4, I think, so it's not much far off) but for the negative sign. Mark, are you reading? Can you look into this? > Testing consistency of piecewise-spline forward-rate curve... > unknown location(0): fatal error in > "PiecewiseYieldCurveTest::testSplineForwardConsistency": std::exception: > > 1st iteration: could not bootstrap the 17th instrument, maturity > November 8th, 2019 > error message: root not bracketed: f[2.22045e-016,3] -> > [-3.371568e-003,-6.562049e-002] > .\lowdiscrepancysequences.cpp(156): last checkpoint This never worked as far as I know---the test case was not enabled in last release. Nando? Are you working on splines? How's it going? Luigi -- All generalizations are false, including this one. -- Mark Twain |