Computational Economics ( > vol.32, no.4 > 1st article) Michael Creel: Using Parallelization to Solve a Macroeconomic Model: A Parallel Parameterized Expectations Algorithm (PDF) Computational Economics, Volume 32, Number 4, November 2008, pp.343--352 ISSN: 0927-7099 (Paper) 1572-9974 (Online) DOI: 10.1007/s10614-008-9142-6 ICCS'06 is ref.#5, cited in p.4/10. MPITB is also cited in pp.8,10. Nice remark in the acknowledgements. Nice highlighting that if the simple case (toy example) scales well, the serious example (with more computing load) will scale better. Impressive Table 1.
		Vadose Zone Journal ( > May 2008, v.7, iss.2 > last paper, pp.843-864) Jasper A. Vrugt, Philip H. Stauffer, Th. Wöhling, Bruce A. Robinson, and Velimir V. Vesselinov: Inverse Modeling of Subsurface Flow and Transport Properties: A Review with New Developments  (PDF/full text no longer free) (PDF draft from Stauffer) Published online 1 May 2008 Vadose Zone J 7:843-864 (2008) DOI: 10.2136/vzj2007.0078 Google found this reference to MPITB: "An example implementation of these functions for Markov chain Monte Carlo simulation using the MPITBtoolbox developed by Fernández et al. ..." In author's draft there are nice remarks about MPITB in pages 17-19. WRT results, they state that "The parallel implementation of AMALGAM is presented in Figure 2. [...] Various case studies presented in Vrugt et al. (2006a) have demonstrated that this setup results in an almost linear speed up for more complex simulation models [...]". For spanish-fluent audience: see location names in Figs.4-5 (Cañada del Buey, Cañón Pajarito). There are also some Cerros del Río in Fig.4. Fig.2 is an aerial areal photo of the MDA@LANL (see figure caption p.69).
		Journal of Applied Econometrics ( > Issues > Jan.Feb.2007 > Vol.22.iss.1 > pp.215-223 ) Michael Creel: I ran four million probits last night: HPC clustering with ParallelKnoppix (html, pdf) Journal of Applied Econometrics, Volume 22, Issue 1, Jan./Feb. 2007, pp.215-223 doi: 10.1002/jae.945 MPITB is mentioned in p.220. VecPar'04 paper referenced.
		Computers & Geosciences ( > Volumes/Issues > Vol.32 iss.8 >  Paper 12) Jasper A. Vrugt, Breanndán Ó Nualláin, Bruce A. Robinson, Willem Bouten, Stefan C. Dekker and Peter M.A. Sloot: Application of parallel computing to stochastic parameter estimation in environmental models Computers & Geosciences, Volume 32, Issue 8 , October 2006, Pages 1139-1155 doi:10.1016/j.cageo.2005.10.015 MPITB (web/VecPar'04) is footnotes [2-3] in p.2. Discussion on use of MPITB is started in Section 4, p.5. The parallel SCEM-UA algorithm is used in 3 case studies. Fig.6 shows almost linear speedup for up to 25 computers, Fig.8 up to 10 computers. Breanndán is system-integrator for Nimrod & MPITB/Octave in VL-E.
		J.A. Vrugt et al. / Journal of Hydrology 325 (2006) 288–307 Application of stochastic parameter optimization to the Sacramento Soil Moisture Accounting model (PDF)  doi:10.1016/j.jhydrol.2005.10.041 They mention the "LAM/MPI interface for the Octave programming environment (Vrugt et al, in press)" in p.291 and p.305. They are using it. Runtimes of 5min for 60.000 model evaluations on a 25 PC Pentium 4 3.4GHz cluster.
		Journal of Hydrology ( > vol. 332, iss.3/4 > paper #4 ) L. Feyen; J.A. Vrugt; B. Ó Nualláin; J. van der Knijff and A. De Roo: Parameter optimisation and uncertainty assessment for large-scale streamflow simulation with the LISFLOOD model (DOI: 10.1016/j.jhydrol.2006.07.004) (alt) Journal of Hydrology, vol. 332, nr 3-4 pp. 276-289. Elsevier, January 2007. MPITB (Vecpar'02 & '04) is references #11 & #12. They used 11 Pentium4 PCs @ 3.4GHz during 35h to calibrate the parameters for the SCEM-UA model for a 3-year simulation period with daily timestep and 1E4 parameter combinations.
		AMS Journals > Journal of HydroMeteorology > Vol.7 > no.3 > last paper: Jasper A. Vrugt, Hoshin V. Gupta, Breanndán Ó Nualláin and Willem Bouten: Real-Time Data Assimilation for Operational Ensemble Streamflow Forecasting DOI: 10.1175/JHM504.1 Journal of Hydrometeorology: Vol. 7, No. 3, pp. 548–565, June 2006 Vrugt's presentation (PDF) from the 2nd HEPEX Workshop Wednesday 10:30 Our library has not online access for this one.
		Computational Economics ( > Volume 26, Number 2 > 1st paper) Creel, M.: User-Friendly Parallel Computations with Econometric Examples (PDF) (alt) Computational Economics, Volume 26, Number 2, October 2005, pp.107--128 ISSN: 0927-7099 (Paper) 1572-9974 (Online) DOI: 10.1007/s10614-005-6868-2 (ACM Portal) (EconPapers) (Ideas) (RecerCat) MPITB is cites [8-10] (VecPar'02, '04 and web) in p.109. Five example problems (Montecarlo, Bootstrap, MLE, GMM, Kernel) are parallelized  using MPITB, starting at Section 4, p.117. See results in Fig.1 p.119. Almost linear speedup up to 12 computers, except for the GMM example (hard one, saturates at speedup 4.75).

 

		Ingram, B., Cornford, D. Parallel geostatistics for sparse and dense datasets geoENV 2008, 8-10 September 2008, Southampton, UK. MPITB (VecPar'04) is reference #5. mentioned in p.3. Pseudo-code in pp.5 & 7. They obtain some of speedup for one of the algorithms proposed (parallel Vecchia), but far from linear. Seems to be a difficult-to-parallelize application?
		Yun Guan, Jan Verschelde: Parallel Implementation of a Subsystem-by-Subsystem Solver (PDF) Proceedings of the HPCS'08, pp. 117-123,  2008 MPITB is cite#4 (ICCS'06), briefly mentioned in Conclusions (p.122).
		Creel, M: Creating and Using a Non-Dedicated HPC Cluster with ParallelKnoppix 12th International Conference on Computing in Economics and Finance, no.202, Cyprus, Jun 22-24, 2006 MPITB is cite [4] (VecPar'04) in p.8. Context: MPITB is included in ParallelKnoppix.
		J. Fernández, M. Anguita, E. Ros, J.L. Bernier: SCE Toolboxes for the development of high-level parallel applications 6th International Conference Computational Science - ICCS 2006, Reading, United Kingdom, May 28-31, 2006. Proceedings of the... Lecture Notes in Computer Science, vol.3992, pp.518-525. ISBN: 3-540-34381-4. ISSN: 0302-9743 DOI: 10.1007/11758525_70 Please use this paper to cite MPITB for Octave. Thanks!
		Creel, M: User-Friendly Parallel Computations with Econometric Examples 11th International Conference on Computing in Economics and Finance, no.445, 2005, Jun 23-25, Washington DC MPITB is cites [8-10] (VecPar'02, '04 and web). Nice acknowledgement, thanks a lot.
		J. Fernández, M. Anguita, S. Mota, A. Cañas, E. Ortigosa, F.J. Rojas: Parallel Programming Toolboxes for Octave VecPar'04, Valencia, Spain, June 28-30 2004.

 

		Irish Centre for High-End Computing (ICHEC -> Training & Support -> Tutorials -> Last guide) Porting Applications to the Blue Gene Platform by Honoré Tapamo, date unknown (near Sept 2008?) (see this thread Jul/08) The LoadLeveler submit script at bottom p.4 reads -env "MPITB_HOME=/ichec/home/staff/htapamo/octave/mpitb"   [...] -exe /ichec/home/staff/htapamo/octave/octave-3.0.1b/bin/octave which indirectly indicates Dr. Tapamo is efectively using MPITB (and perhaps Nils' patches or some variation of them) with Octave-3.0.1.
		University of Minnesota > Institute for Mathematics and its Applications (IMA) > IMA Thematic Year on Applications of Algebraic Geometry (September 2006 - June 2007) > Workshop : Software for algebraic geometry (October 23-27, 2006) Yun Guan, Jan Verschelde: "PHClab: A MATLAB/Octave interface to PHCpack" (pdf) (poster) Poster #1, see schedule. Link #27, see software list. URL: PHCpack. MPITB (ICCS'06) is reference #2 in page 16/21. See MPITB code and results in p.11. They got a rather (dynamically) balanced workload, between 12-17 tasks per slave for 13nodes +360s (6min) -> sequential solution would have taken 1:18'. This was published in Springer: The IMA Volumes in Mathematics and its Applications vol.148, 2008, IX, 176 p., Hardcover, ISBN: 978-0-387-78132-7 Software for Algebraic Geometry TOC > paper#2 > PhClab: A MATLAB/Octave interface to PHCpack (pdf) doi:10.1007/978-0-387-78133-4 May 29, 2008. pp.15-32 "Solving Polynomial Systems using PHCpack". Geometry of Mechanism Science (GeMS'07). University of Notre Dame, 1 - 3 March 2007. slides of the talk (see pp.17-18/44, ICCS'06 paper referenced) Guan managed to overcome the limitations of a cluster where ssh was installed setgid root, xterms didn't inherit LD_LIBRARY_PATH settings, slave nodes didn't resolve the master node name... she managed to fix all that and parallelize her code in just 2 months!
		University of Oxford (> E-Research Centre > oerc > ) "Supercomputing @ Oxford"  (PDF) Seminar by the OERC, presented by Dr. Anne Trefethen, Oct  26th 2006 MPITB mentioned in slide 38/45, in the context of Choy's parallel MATLAB survey BTW, Moler was there too Parallel MATLAB, Cleve Moler seminar, Tuesday 9th January, 2007 Same name as the other Cleve's Corner (see the list) Parallel MATLAB, Cleve's Corner June 2007
		RecerCat ( > Autors > C > Creel, Michael D. > 9 May 2006) Creel, M: "A note on parallelizing the Parameterized Expectations Algorithm" (PDF) Working paper, Unitat de Fonaments de l'Anàlisi Econòmica; Institut d'Anàlisi Econòmica; Universitat Autònoma de Barcelona. Other works (Tutorials) by Michael from RecerCat: "Creating and using a non-dedicated HPC cluster with ParallelKnoppix" "ParallelKnoppix tutorial" (alt) "ParallelKnoppix - Rapid deployment of a linux cluster for MPI parallel processing using non-dedicated computers"
		VL-E > News > 7 April 2006 VL-e workshop > Presentations > Virtual lab methodology I) Breanndán Ó Nualláin: Interactive Problem Solving Environments Presentations of the VL-e workshop 7 April 2006. MPITB is slide 10/14

 

		EconPapers ( >Authors > C > Creel, Michael > Working Paper 2005) IDEAS ( > Authors > C > Creel, Michael > Works > 637.05) UFAE (Unitat de Fonaments de l'Anàlisi Econòmica > Membres > Creel) M. Creel: (alt location) User-Friendly Parallel Econometric Computations: Monte Carlo, Maximum Likelihood, and GMM (Web Page / PDF) UFAE and IAE Working Papers from Unitat de Fonaments de l'Anàlisi Econòmica (UAB) and Institut d'Anàlisi Econòmica (CSIC) MPITB is cite [9-10] in p.4/25. Discussion on the parallel programs is started in Section 4 p.10/25. See speedup results in Fig.1 p.18/25. The other figures are also obtained with MPITB in the Parallel-Knoppix distribution introduced in this paper. Other works (Tutorials) by Michael from EconPapers: (note to self: some are conferences, must find/update links and add comments, perhaps merge with RecerCat) "Creating and Using a Non-Dedicated HPC Cluster with ParallelKnoppix" 2006 12th Computing in Economics and Finance no.202, 2006, Jun 22-24, Cyprus ([4] is VecPar'04) "A Note on Parallelizing the Parameterized Expectations Algorithm" 2005 ([4] is VecPar'04, nice Ack. on footnote, thanks!) See also this Computational Economics paper (DOI: 10.1007/s10614-008-9142-6) "ParallelKnoppix Tutorial" web page, nice acknowledgement. "User-Friendly Parallel Computations with Econometric Examples" 11th Computing in Economics and Finance no.445, 2005, Jun 23-25, Washington DC Computational Economics, Volume 26, Number 2, October 2005, pp.107--128 ([8-10] are VECPAR'02, '04 and web) "ParallelKnoppix - Rapid Deployment of a Linux Cluster for MPI Parallel Processing Using Non-Dedicated Computers" 2004 ([4] is VecPar'04)

 

		Linux on Itanium (Gelato > Community > Events > Apr'06 Technical presentations > General interest > "An evaluation of..." by Ashok Krishnamurthy > Presentation) Dalwani, Ludban, Hudak, Krishnamurthy: "High-Performance Parallel Octave on Itanium using ParaM" MPITB classified as "Other projects" at pp.21-22/38 (together with MultiMATLAB) J.W.Eaton listed as contributor at p.38/38 Previously from the Presto group, there was a Technical Report: R.Panuganti, M.M.Baskaran, D.Hudak, A.Krishnamurthy, J.Nieplocha, A.Rountev, and P. Sadayappan: "GAMMA: Global Arrays Meets MATLAB" Technical Report OSU-CISRC-1/06-TR15, Department of Computer Science and Engineering, Ohio State University, January 2006. MPITB (reference #[1], which is PVMTB web) was cited together with MultiMATLAB and MatlabMPI in Section 4.4 p.6/22, in the context of "message-passing is too low-level for MATLAB users". Their GAMMA software is explained to be part of a future ParaM software (p.8). Apparently, they moved from MATLAB to Octave in-between.
		John W. Eaton: "GNU Octave: History and Outlook for the Future" 2005 AIChE Meeting, Cincinnati Ohio, November 1, 2005 Just a mention in p.20/40, but in the context of bootable CDs... perhaps the idea was to cite ParallelKnoppix instead? Can't find the paper in the preliminary Conference program. There is at least a comment from the Editors, taken from the CACHE web, from where one can also reach to these good news. Congratulations! Aside: other 2 conferences, they don't cite us but are useful references. DSC'03 (link)(alt) John W. Eaton (alt) and James B. Rawlings. Ten years of Octave -- recent developments and plans for the future. Proceedings of the 3rd International Workshop on Distributed Statistical Computing, Vienna, Austria, March 2003. DSC'01 (link)(alt) John W. Eaton. Octave: Past, present and future. In Kurt Hornik and Fritz Leisch, editors, Proceedings of the 2nd International Workshop on Distributed Statistical Computing, March 15-17, 2001, Technische Universität Wien, Vienna, Austria, 2001.

 

		Linux-info-Tag: Jens Weiße: Cluster bauen für Anfänger (Cluster building for beginners) Oct'05 Mention to MPITB/Octave in p.33/49

 

		in spanish OSLUCA (Universidad de Cádiz > Oficina de Software Libre > II Jornadas (broken, cached@google, local copy) > Miércoles 6, 17:00) Javier Fernández: Programación científica en clusterscon Octave MPITB (OpenOffice Impress / PDF) II Jornadas de Software Libre de la Universidad de Cádiz Facultad de CC. Económicas y Empresariales 6-8 Abril 2005, Cádiz, Spain Otros enlaces: Universia 01/04/05 Universia 07/04/05 Noticias UCA 30/03/05 Noticias UCA 07/04/05 Gaditanos Linuxeros InfoCádiz