Anith Gopal

Here are my notes from the second day of ISMP .  ( Here are my day 1 notes. )  I spent the whole day going to IPM talks: Meszaros : Unfortunately I missed the plenary talk by Friedrich Eisenbrand due to microbrews.  I attended an IPM session after that – first up was Csaba Meszaros – interesting to hear from him because BPMPD has been such an impressive code for so many years.  His talk was about recent improvements in BPMPD .  BPMPD has been in continuous development for 15 years or so.  His recent work appears to focus on a) QCQP b) exploiting “hypersparsity”.  He is using a primal-dual log barrier method  – not a primal-dual HSD method as we at Solver Foundation have chosen for our SOCP solver.  So his handling of starting points and infeasibility detection is different.  He went into the tradeoffs between using augmented and normal systems for the search direction, which affect both performance and numerical stability.  He has implemented both and has logic on top to determine which to use.  He also talked about QCQP presolve, which often whittles down the size of “real” models substantially.   An interesting bit of trivia: he showed a table that compared lines of code in various modules, “then and now”: 2009 1981 Ordering 5700 155 Cholesky 2825 100 Backsolve 820 45 I think the difference is indicative of both Csaba’s commitment to improving BPMPD as well as the number of practical issues you need to think about when building an industrial grade solver.

Read the original post:
Int’l Symposium on Mathematical Programming, Day 2