%21Queries%20%0A+%20%5B%5BAll%20Pages%7CSpecial%3A%3AAll%20Pages%5D%5D%3A%20a%20list%20of%20all%20pages%20in%20this%20wiki.%20%0A+%20%5B%5BPages%20not%20yet%20created%7CSpecial%3A%3APages%20not%20yet%20created%5D%5D%3A%20a%20list%20of%20pages%20that%20have%20links%20to%20them%2C%20but%20were%20still%20not%20created.%20%0A+%20%5B%5BOrphaned%20Pages%7CSpecial%3A%3AOrphaned%20Pages%5D%5D%3A%20a%20list%20of%20pages%20that%20once%20were%20linked%20to%2C%20but%20now%20no%20page%20links%20to%20them%20anymore.%0A%0A%21Visual%20%0A+%20%5B%5BEdit%20CSS%7CSpecial%3A%3AEdit%20CSS%5D%5D%20%0A+%20%5B%5BBlock/Unblock%20edits%7CSpecial%3A%3ABlock%20Edits%5D%5D%3A%20hides%20the%20edit%20button%2C%20to%20be%20used%20when%20this%20wiki%20goes%20public%20%28when%20used%20as%20a%20webpage%2C%20for%20example%29%0A%0A%21Import%20%0A+%20%5B%5BImport%20wiki%7CSpecial%3A%3AImport%20wiki%5D%5D%20from%20an%20early%20%28or%20current%29%20version.%0A%0A%21Info%0A+%20This%20wiki%20was%20last%20modified%20in%20%3Cspan%20id%3D%22lastDate%22%3E%5Berror%5D%3C/span%3E.
Import%20a%20wiki%20from%20an%20early%20%28or%20current%29%20version%3A%0A%0AFile%3A%20%3Cinput%20id%3D%22filename_%22%20type%3D%22file%22%3E%3Cinput%20type%3D%22button%22%20value%3D%22Import%22%20onclick%3D%22import_wiki%28%29%22%3E%0A%0A*Important%3A*%20CSS%20will%20%3Cu%3Enot%3C/u%3E%20be%20imported%2C%20as%20it%20might%20break%20some%20funcionality%20of%20the%20current%20version.
%2126.%20October%202007%0ALong%20time%20no%20entry.%20But%20I%20have%20not%20been%20%28too%29%20lazy.%20I%20implemented%20a%20local%20vola%20simulation%20for%20the%20G80%20and%20started%20writing%20up%20the%20stuff%20I%20did.%20Paper%20and%20source%20code%20will%20be%20available%20form%20this%20site%20soon.%20Meanwhile%20have%20a%20look%20at%20my%20two%20famous%20videos%20on%20the%20net%3A%0A%5B%5BStringDucky%7Chttp%3A//discovermagazine.com/twominutesorless%3Fbcpid%3D716091875%26bclid%3D686943766%26bctid%3D687029421%5D%5D%20and%20%5B%5BWhat%20we%20are%7Chttp%3A//www.youtube.com/watch%3Fv%3Da15KgyXBX24%5D%5D%0A%0ABTW.%20The%20PS3%27s%20price%20finally%20dropped%20below%20400%u20AC.%20Will%20buy%20one%20soon%2C%20latest%20for%20Christmas.%0A%0A%0A%2126.%20June%202007%0AStripped%20down%20the%20code%20for%20the%20LCG%20random%20number%20generator%20to%20a%20minimum.%20Now%20it%20is%20using%20a%20mix%20of%20constant%20and%20shared%20memory.%0A%0A%212.%20June%202007%0APorted%20two%20parallel%20random%20number%20generators%20to%20CUDA.%0A%0A%2128.%20May%202007%0ANot%20happening%20much%20on%20the%20Main%20Page.%20But%20progress%20with%20the%208800GTS%20project...%0A%0A%2115.%20April%202007%0AI%20have%20to%20confess%2C%20I%20did%20not%20update%20my%20site%20for%20a%20while.%20Now%20I%20found%20a%20nice%20and%20simple%20layout%20and%20promise%20to%20update%20more%20recently.%20Main%20updates%20will%20go%20to%20the%20%5B%5BBooks%20and%20Papers%5D%5D%20section%20where%20I%20want%20to%20list%20all%20papers%20I%20hav%20filed%20with%20some%20comments.%20%0A%0AThis%20page%20by%20the%20way%20is%20a%20%5B%5BStickWiki%7Chttp%3A//stickwiki.sourceforge.net/%5D%5D.%20Very%20simple%2C%20very%20easy...%0A%0A%2127.%20December%202005%0AUntil%20now%20I%20did%20not%20feel%20entitled%20to%20create%20a%20homepage.%20But%20with%20my%20installation%20of%20the%20NSLU2%20as%20control%20center%20for%20my%20power%20socket%20this%20has%20changed.%20I%20would%20like%20to%20show%20this%20simple%20and%20cheap%20way%20to%20more%20people.%20You%20will%20find%20details%20at%20%5B%5BNSLU2%20controls%20power%20socket%5D%5D.%0A%0A
%21News%0ASince%20my%20gateway%20down%20in%20the%20basement%20now%20runs%20all%20the%20time%20I%20can%20use%20it%20for%20switching%20stuff.%20So%20the%20NSLU2%20became%20obsolete.%0AThe%20gateway%20runs%20%5B%5Bsispmctl-2.4b%7Chttp%3A//sispmctl.sourceforge.net/%5D%5D%20which%20should%20also%20work%20on%20the%20NSLU2.%0A%0A%21Summary%20%2827.%20December%202007%29%0A%0ASince%20some%20time%20I%20have%20been%20looking%20for%20a%20simple%20solution%20to%20switch%20a%20power%20socket%20through%20my%20LAN.%20There%20are%20some%20vendors%20offering%20such%20solutions%20but%20usually%20at%20fairly%20high%20rates.%0A%0AI%20considered%20a%20serially%20attached%20custom%20solution%20but%20quickly%20discarded%20this%20idea.%20My%20electroni%20skills%20are%20rudimentyry%20and%20playing%20around%20with%20230V%20is%20not%20always%20healthy.%0A%0ADuring%20my%20search%20I%20found%20a%20multiple%20socket%20outlet%20being%20sold%20by%20the%20German%20distributer%20Pearl%20named%20Intelli-Plug.%20The%20producer%20%5B%5BGembird%7Chttp%3A//www.gmb-online.nl%5D%5D%20distributes%20theses%20devices%20under%20the%20name%20%5B%5BSIS-PM%7Chttp%3A//www.gmb-online.nl/Reference.asp%3FReference%3D2755%5D%5D.%20At%20Pearl%20I%20paid%2030Euro%2C%20meanwhile%20I%20found%20cheaper%20offers%20down%20to%2020Euro.%20The%20SIS-PM%20is%20controlled%20via%20USB%20and%20drivers%20for%20Windows%20and%20Linux%20exist.%20So%20this%20is%20the%20USB-switchable%20power%20socket%20I%20have%20been%20looking%20for.%20Only%20thing%20missing%20is%20the%20control%20center.%0A%0A%3Cimg%20src%3D%22SIS-PM.jpg%22%3E%0A%0AMy%20gateway%20in%20the%20basement%20could%20have%20done%20the%20job%20but%20one%20switchig%20task%20was%20to%20switch%20on%20exactly%20this%20gateway%20while%20I%20am%20in%20my%20home%20office.%20Furthermore%20the%20power%20consumption%20of%20this%20gateway%20is%20higher%20than%20I%20expected%20the%20control%20center%27s%20to%20be.%0A%0AAfter%20some%20time%20I%20found%20a%20link%20to%20the%20NSLU2.%20It%27s%20original%20purpose%20is%20a%20network%20attached%20storage%20%28NAS%29%20but%20it%20is%20easily%20extended.%20Thanks%20to%20OpenSlug%20%28and%20others%29.%20The%20installation%20is%20not%20compliated%20an%20the%20Linux%20driver%20for%20the%20SIS-PM%20compiles%20flawlessly%20after%20insatlling%20some%20development%20libraries%20%28see%20below%29.%0A%09%0A%3Cimg%20src%3D%22linksysnslu2.jpg%22%3E%0A%0AConclusion%3A%20for%20about%20110%20to%20120%20Euro%20one%20gets%20the%20possibility%20to%20switch%204%20power%20sockets%20through%20the%20LAN...%0A%0A%21Installing%20NSLU2%0A%0AThe%20installation%20of%20the%20NSLU2%20is%20pretty%20easy.%20You%20will%20find%20all%20necessary%20information%20on%20%5B%5Bhttp%3A//www.nslu2-linux.org%5D%5D.%0A%0AIn%20addition%20to%20the%20standard%20%5B%5BOpenSlug%7Chttp%3A//www.nslu2-linux.org/wiki/OpenSlug/HomePage%5D%5D%20nach%20installation%20I%20use%20the%20following%20packages%0A%0Avim%2C%20less%2C%20ncftp%2C%20rsync%2C%20thttpd%2C%20wget%2C%20samba%2C%20ntpdate%2C%20libusb-dev%2C%20libusb%2C%20usbutils%2C%20samba%2C%20openslug-native%2C%20kernel-source%2C%20atftpd%2C%20vsftpd%2C%20e2fsprogs%2C%20busybox%0A%0AThe%20following%20stuff%20had%20to%20be%20recompiled%20by%20myself%3A%0A%0Atunes2fs%2C%20rtools%0A%0AEverything%20went%20smoothly.%20The%20only%20thing%20puzzling%20me%20was%20the%20slow%20transfer%20rate%20to%20the%20attached%20diskdrive%20with%20only%20300kByte/s%20via%20%27scp%27%20and%20%27rsync%27.%20My%20first%20guess%20was%20that%20the%20external%20USB-case%20for%20my%20IDE%20drive%20was%20the%20bottleneck.%20Pluging%20it%20to%20my%20desktop%20showed%20a%20transfer%20rate%20of%20750kByte/s%20which%20still%20is%20too%20small%2C%20so%20I%20saw%20my%20first%20guess%20confirmed.%20Eventually%20the%20slow%20rate%20on%20my%20desktop%20was%20explained%20by%20the%20use%20of%20the%20wrong%20driver%20%28USB1.1%20instead%20of%202.0%29.%20The%20slow%20rate%20at%20the%20NSLU2%20using%20%27scp%27%20and%20%22rsync%27%20was%20explained%20by%20the%20small%20computing%20power%20of%20the%20NSLU2%20%28I%20consider%20de-underclocking%20the%20device%20but%20did%20not%20do%20so%20up%20to%20now%29.%20Looking%20at%20%27top%27%20during%20data%20transfer%20shows%20100%25%20CPU-usage%20by%20%27scpd%27.%20Using%20%27rcp%27%20and%20%27rsync%20-e%20rcp%27%20instead%20everything%20works%20as%20expected.%20I%20did%20not%20find%20%27rcp%27%20in%20any%20precompiled%20package%2C%20so%20I%20compiled%20it%20myself.%20%0A%0A
%21Papers%0A+%20%5B%5Bpapers%5D%5D%20read%20recently%0A+%20%5B%5BFinance%20papers%5D%5D%20sorted%20by%20topic%0A+%20%5B%5BComputing%20papers%5D%5D%20sorted%20by%20topic%0A%0A%21Books%0A+%20%5B%5BFinance%20books%5D%5D%0A+%20%5B%5BComputing%20book%5D%5D%0A+%20%5B%5BNumerics%20books%5D%5D%0A%0A
%21Scientific%20computing%20on%20the%20PS3%20%2820.%20April%202007%29%0AAnother%20aspired%20fun%20project%20is%20to%20try%20the%20PS3%20for%20scientific%20computing.%20So%20I%20plan%20to%20purchase%20a%20PS3%20with%2020GB%20harddrive%20as%20soon%20as%20it%20is%20out%20and%20available%20for%20less%20than%20400%u20AC.%20Since%20this%20wont%20be%20the%20case%20for%20the%20next%20few%20months%2C%20I%20will%20concentrate%20on%20my%20%5B%5BCUDA%7CScientific%20computing%20on%20the%20GTS8800%5D%5D%20project.%0A%0ASome%20links%3A%0A%5B%5BIBM1%7Chttp%3A//www-306.ibm.com/chips/techlib/techlib.nsf/products/Cell_Broadband_Engine%5D%5D%0A%5B%5BIBM2%7Chttp%3A//www-128.ibm.com/developerworks/power/cell/docs_articles.html%5D%5D%0A
%21Scientific%20computing%20on%20the%208800GTS%0A%0ALinks%3A%0A+%20See%20for%20example%20the%20original%20Nvidia%20documentation%20on%20%5B%5BCUDA%7Chttp%3A//developer.nvidia.com/object/cuda.html%5D%5D.%0A+%20And%20a%20recent%20%28as%20time%20of%20writing%20this%20in%20April%202007%29%20%5B%5Blecture%7Chttp%3A//courses.ece.uiuc.edu/ece498/al%5D%5D.%0A+%20Nvidia%20%5B%5BForums%7Chttp%3A//forums.nvidia.com/index.php%3Fact%3Didx%5D%5D%0A+%20%5B%5BGPGPU%7Chttp%3A//www.gpgpu.org/sc2006/%5D%5D%0A%0A%0A%21History%0A%21%21%2826.%20June%202007%29%0AAfter%20I%20upgraded%20to%20CUDA%200.9%20last%20weekend%2C%20today%20CUDA%201.0%20has%20been%20released.%20I%20upgraded%20immediately%20after%20I%20found%20the%20announcement.%20First%20good%20news%3A%20the%20timeouts%20after%20about%207.5secs%20are%20gone.%20Calculations%20can%20now%20take%20much%2C%20much%20longer.%20Stuff%20is%20working%20seamlessly%20on%20my%20Debian%20Etch/Lenny/whatever%20mix.%20Fine%20new%20stuff%20on%20my%20random%20number%20generator%20porting%20of%20the%20%5B%5BLCG%7CLCG%20success%5D%5D%20combined%20with%20some%20simple%20Monte%20Carlo%20option%20pricing.%0A%0A%21%21%282.%20June%202007%29%0AAfter%20some%20pain%20with%20the%20%5B%5BMLFG%7CMLFG%20pain%5D%5D%20from%20%5B%5BSPRNG%7Chttp%3A//sprng.cs.fsu.edu/%5D%5D%20I%20move%20to%20the%20%5B%5BLCG%7CLCG%20success%5D%5D%20from%20the%20same%20library.%20Find%20reports%20and%20downloads%20follwing%20the%20above%20links.%0A%0A%21%21%2828.%20May%202007%29%0AAll%20computer%20up%20and%20runings.%20Found%20time%20to%20do%20some%20first%20steps%20in%20CUDA%20programming%20and%20started%20porting%20the%20Parallel%20Random%20Number%20Generators%20Library%20%5B%5BSPRNG%7Chttp%3A//sprng.cs.fsu.edu/%5D%5D%20to%20CUDA.%20The%20Modified%20Lagged%20Fibonacci%20Generator%20%28MLFG%29%20is%20working%20%28just%20have%20to%20sort%20out%20one%20issue%20about%20not%20completely%20agreeing%20sequences%20on%20host%20and%20device%29.%20As%20soon%20as%20this%20issue%20is%20soved%2C%20I%20will%20make%20the%20code%20available.%0A%0A%21%21%2817.%20May%202007%29%0AAscension%20Day.%20Got%20up%20at%206%3A15%20in%20the%20morning%20and%20finally%20installed%20and%20tested%20CUDA.%20First%20thing%20to%20mention%3A%20do%20not%20use%20the%20latest%20driver.%20It%20does%20not%20support%20CUDA.%20Use%20*driver%201.0-9751*%20instead.%20%0A%0A%21%21%2814.%20May%202007%29%0AFound%20time%20to%20quickly%20install%20%5B%5BXubuntu%7Chttp%3A//www.xubuntu.org%5D%5D%20and%20the%20latest%20%281.0-9755%29%20%5Bbad%20idea%21%5D%20Nvidia%20drivers.%20After%20finding%20the%20%22irqpoll%22%20commandline%20switch%20everything%20went%20like%20a%20charm.%20The%20Nvidia-card%20could%20be%20configure%20on%20the%20desktop%20to%20my%20monitor%27s%20native%20solution%20of%201920x1200.%20Last%20big%20install%20was%20some%20years%20ago%20with%20a%20plain%20Debian%20and%20that%20needed%20more%20manual%20work.%0A%0A%21%21%2811.%20May%202007%29%0AFetched%20the%20computer%20from%20the%20post%20office%20just%20before%20traveling%20to%20my%20parents%20over%20the%20weekend.%20The%20vendor%20used%20a%20different%20case%20than%20I%20expected.%20Now%20I%20have%20a%20case-modder%27s%20case%20with%20one%20side%20made%20of%20Plexiglas.%20That%20was%20_not_%20intended%20%3A%29%0A%0A%21%21%2801.%20May%202007%29%0AI%20eventually%20orderd%20a%20new%20PC%20including%20the%20wanted%20Nvidia%208800GTS.%20The%20remaining%20part%20of%20the%20system%20is%20straight%20forward%3A%0A+%20Motherboard%20Abit%20AB9%20%0A+%20Core2Duo%20E6600%20%282x2.4GHz%29%0A+%202048MB%20DDR2%20PC800%0A+%20320MB%208800GTS%0A+%20160GB%20Harddrive%0ATime%20to%20delivery%20up%20to%209%20days...%0A%0A%21%21%2820.%20April%202007%29%0AI%20think%20about%20ordering%20a%20new%20PC%20including%20a%208800GTS%20to%20play%20with%2C%20understand%20and%20exploit%20the%20power%20of%20the%20graphics%20processor...%0A%0A%0A%0A%0A
%21CMS%20-%20Convexity%20adjustments%20and%20spreads%0A+%20Convexity%20Conundrums%3A%20Pricing%20CMS-Swap%2C%20Caps%20and%20Floors%20-%20*Pat%20Hagan*%20-%20Wilmott%20Magazin%0A+%20Towards%20a%20multi-stochastic%20volatility%20model%20for%20CMS%20spread%20exotic%20-%20*Vladimir%20Piterbarg*%20-%20Global%20Derivatives%202007%0A+%20Notes%20on%20convexity%20and%20quanto%20adjustments%20for%20interest%20rate%20and%20related%20options%20-%20*Wolfram%20Boenkost%2C%20Wolfgang%20Schmidt*%20-%20HfB%0A+%20Interest%20Rate%20Convexity%20and%20teh%20Volatility%20Smile%20-%20*Wolfram%20Boenkost%2C%20Wolfgang%20Schmidt*%20-%20HfB%0A+%20CMS-Derivatives%2C%20Convexity%20Adjustments%20and%20Smile%20-%20*Massimo%20Morini*%20-%20Fixed%20Income%20Conference%202006%0A+%20Smiling%20at%20Convexity%20-%20*Fabio%20Mercurio%2C%20Andrea%20Pallavicini*%20-%20May%2025.%202006%0A+%20Swaption%20skews%20and%20convexity%20adjustments%20-%20*Fabio%20Mercurio%2C%20Andrea%20Pallavicini*%20-%20July%2021.%202006%0A+%20Buy%20a%20dollar%20for%2080cents%20-%20*Lehman%20Brothers*%0A%0A%21MFM%0A+%20Markov-functional%20interest%20Rate%20Models%20-%20*Hunt%2C%20Kennedy%2C%20Pelsser*%20-%20Finance%20and%20Stochastics%202000%20%0A+%20Markov%20interest%20rate%20models%20-%20*Hagan%2C%20Woodward*%20-%20October%2027.%201997%0A+%20A%20comparison%20of%20Markov-Functional%20and%20Market%20Models%20-%20The%201d%20Case%20-%20*Bennett%2C%20Kennedy*%20-%20July%2025.%202005%0A+%20A%20comparison%20of%20Single%20Factor%20Markov-functional%20and%20Multi%20factor%20Market%20Models%20-%20*Pietersz%2C%20Pelsser*%20-%0A+%20Market%20Models%20vs%20Markov%20Functional%20Models%20-%20*F%F6rger*%20-%20Thesis%20Oxford%0A+%20Cross%20Currency%20and%20Hybrid%20MFM%20-%20*Fries%2C%20Rott*%20-%20May%202004%0A+%20A%20Two-Factor%20MFM%20-%20*Gernsheimer*%20-%20Thesis%20October%202003%0A%0A%21Shortrate%20Models%20papers%0A%21Libor%20Market%20Model%20papers%0A%0A%21Local%20Vol%0A%0A%21Stochastic%20Volatility%20papers%0A%21Monte%20Carlo%20papers%0A%21PDE%20papers%0A%0A%0A%21miscellaneous%20papers%0A
%0A%21General%3A%0ADeriVatiVes%3A%20Paul%20Wilmott%0AOptionFuturesAndOtherDerivatives%20%3A%20John%20Hull%0ADynamicHedging%20%3A%20Taleb%0AMarketModels%20%3A%20Alexander%0A%0A%21Numerical%20methods%20and%20Mathematics%3A%0AStochastic%20CalculusForFinance%20%3A%20Steve%20Shreve%0AImplementing%20Derivatives%20Models%20%3A%20Clewlow%20/%20Strickland%0AFinancial%20Engineering%20With%20FEM%20%3A%20Topper%0AFinancialInstrumentPricingUsingC++%20%3A%20Daniel%20Duffy%20%28there%20are%20better%20books%29%0AModelingDerivativesInC++%20%3A%20Justin%20London%20%28worst%20book%20ever%21%29%0AMonteCarlo%20%3A%20Jaeckel%0AMonteCarloMethodsInFinancialEng%20%3A%20Glasserman%0ANumericalSolutionsToSDE%20%3A%20Kloeden%20/%20Platen%0AStochasticDifferentialEquations%20%3A%20Oksendal%0AOptionValuationUnderSV%20%3A%20Alan%20Lewis%0A%0A%21%21Interest%20Rate%20Derivatives%3A%0AInterest%20Rate%20Models%20%3A%20Brigo%20/%20Mercurio%0AModern%20Pricing%20Of%20IRD%20%3A%20Rebonato%0AEfficient%20Methods%20For%20Pricing%20IRD%20%3A%20Pelser%0A%0A%21%21Credit%20Derivatives%3A%0ACreditDerivativesAndSyntheticStructures%20%3A%20Tavakoli%0ACreditDerivatives%20%3A%20Chaplin%0A%0A
%21%21C%26%2343%26%2343%3A%0AThis%20is%20a%20list%20of%20my%20famous%20C%26%2343%26%2343-books.%20For%20some%20books%20I%20added%20a%20comment.%0A+%20C%26%2343%26%2343%20Programming%20Language%20%3A%20Bjarne%20Stroustrp%0A+%20Effective%20C%26%2343%26%2343%20%3A%20Scott%20Meyers%20%283.%20edition%29%0A+%20C%26%2343%26%2343%20Standard%20Library%20%3A%20Nicolai%20Josuttis%0A+%20Beyond%20The%20C%26%2343%26%2343%20Standard%20Library%20%3A%20Bj%F6rn%20Karlson%0A+%20C%26%2343%26%2343%20Coding%20Standards%20%3A%20Sutter%20/%20Alexandrescu%0A+%20C%26%2343%26%2343%20Template%20Meta%20Programming%20%3A%20Abrahms%20/%20Gurtovovy%0A+%20Modern%20C%26%2343%26%2343%20Design%20%3A%20Alexandrescu%0A+%20%5B%5BExcel%20Add%20In%20Development%5D%5D%20%3A%20Steve%20Dalton%0A%0A%21%21Parallel%20Computing%0A+%20Parallel%20Programming%20in%20C%20with%20MPI%20and%20OpenMP%20%3A%20Michael%20Quinn%3Cbr%3E%20%5BPretty%20nice%20introduction%20focussing%20mainly%20on%20MPI%20and%20less%20on%20OpenMP.%20Also%20touches%20parallel%20programming%20concepts.%5D%0A+%20Parallel%20Programming%20in%20OpenMP%20%3A%20Chandra%20et%20al.%3Cbr%3E%20%5BDetailed%20introduction%20to%20OpenMP%5D%0A
%21%21Maths%3A%0A+%20Introduction%20To%20Time%20Series%20and%20Forecasting%20%3A%20Brockwell%20/%20Davis%0A+%20Numerical%20Recipes%20%3A%20Press%20et.al.%0A+%20Probability%20and%20Statistics%20%3A%20DeGroot%0A
%21General%20Information%0ADr.%20Michael%20Dirkmann%0Aborn%2028.April%201970%20in%20Marl%2C%20Germany%0A%0AContact%20me%20by%0A+%20eMail%3A%20michael%20AT%20dirkmann.org%0A+%20ICQ%3A%20354866309%0A%0A%21Degrees%0A2003%20MSc%20in%20Computational%20Finance%20from%20Carnegy%20Mellon%20University%20%28CMU%29%0A1998%20PhD%20in%20Physics%3A%20Experimental%20High%20Energy%20Physics%20%3A%20University%20of%20Dortmund%20and%20DESY%20in%20Hamburg%0A1995%20Diplom%20in%20Physics%20%3A%20University%20of%20Dortmund%0A1989%20Abitur%0A%0A%21Employments%0A10/2006%20-%20present%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Sal.Oppenheim%20Frankfurt%20%3A%20Quantitative%20Analyst%0A09/2005%20-%2009/2006%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20DZ%20Bank%20%3A%20Finacial%20Engineer%0A06/2004%20-%2008/2005%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20DrKW%20%3A%20Quantitative%20Researcher%0A06/2002%20-%2005/2004%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20DrKW%20%3A%20Project%20Manager%20%28software%20development%29%0A06/1999%20-%2005/2002%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20DrKW%20%3A%20Senior%20Software%20Engineer%0A01/1999%20-%2005/1999%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Mummert+Partner%20%3A%20Consultant%20%0A03/1998%20-%2012/1998%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Military%20Service%0A11/1997%20-%2012/1998%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Scientific%20Staff%20at%20University%20of%20Dortmund%0A03/1995%20-%2011/1997%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Scholarship%20%3A%20Graduiertenkolleg%20%22Erzeugung%20und%20Zerf%E4lle%20von%20Elementarteilchen%22%0A09/1993%20-%2012/1994%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Assistent%20and%20Scientific%20Staff%20at%20University%20of%20Dortmund%0A01/1993%20-%2004/1993%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Assistent%20at%20Institute%20for%20Accelerator%20Physics%20at%20the%20University%20of%20Dortmund%0A05/1992%20-%2012/1992%20%26nbsp%3B%26nbsp%3B%26nbsp%3B%20Assistent%20at%20Institute%20for%20Robot%20Research%20at%20the%20University%20of%20Dortmund%0A%0A%21Current%20and%20previous%20works%0A%21%21Interest%20Rate%20Modelling%0A+%20Markov%20Functional%20Model%0A+%20Cheyette%20Model%20%28multi-factor%20and%20stochastic%20volatility%3A%20see%20my%20%5B%5Btalk%7Chttp%3A//www.dirkmann.org/hfb_2006.pdf%5D%5D%20at%20Frankfurt%20School%20of%20Finance%29%0A+%20SABR-%20and%20SVI-Model%0A+%20convexity%20adjustments%0A+%20extended%20Vasicek%20%28multi%20factor%29%0A+%20Interest%20Rate-Equity-Hybrid%20model%20with%20multi-factor%20extended%20Vasicek%20IR-component%20%0A%0A%21%21Pricing%0A+%20Payoff%20language%20for%20Callable%20Interest%20Rate%20Exotics%20%28generic%20MC%20and%20regression%20based%20valuation%20of%20callrights%29%0A+%20Monte%20Carlo%20Simulation%0A+%20Partial%20Differential%20Equations%0A%0A%21Special%20Interests%20and%20Skills%0A%21%21Programming%0A+%20C%26%2343%26%2343%2C%20C%2C%20F77%0A+%20XLW%20for%20C%26%2343%26%2343-Excel%20Plugins%0A%0A%21%21Parallel%20Programming%20and%20Distributed%20Computing%0A+%20PVM%2C%20MPI%2C%20OpenMP%0A+%20Condor%0A+%20general%20purpose%20GPU%20computing%20%0A+%20DataSynapse%20Grid%20Server%20%28evaluation%20project%29%0A%3C%21--%20+%20Platform%20Symphony%20SDK%20%28to%20come%29%20--%3E%0A%0A%21%21Special%20Software%0A+%20R%0A+%20Matlab%0A+%20QuantLib%0A+%20LaTeX%0A%0A
+%20Some%20formatting%20for%20%5B%5BStickWiki%7CStickWiki%20Formatting%5D%5D%0A%0A%0A
%21Excel%20Add%20In%20Development%20%3A%20Steve%20Dalton%0A%0AThis%20book%20is%20only%20for%20deeper%20insight.%20If%20you%20just%20want%20an%20Excel%20plugin%20from%20your%20C%26%2343%26%2343-code%20you%20could%20just%0A+%20download%20XLW%0A+%20read%20the%20examples%0A+%20start%20coding%20%0A+%20have%20an%20Excel%20plugin%20within%20less%20than%20one%20hour.%0A%0A
%21Contact%20details%0A%0AName%3A%20Dr.%20Michael%20Dirkmann%20%0AeMail%3A%20michael%20AT%20dirkmann.org%0AICQ%3A%20%20354866309%0Afind%20more%20details%20in%20my%20%5B%5BCV%5D%5D
%21First%20attempt%3A%0ABased%20on%20the%20C-version%202.0%20of%20%5B%5BSPRNG%7Chttp%3A//sprng.cs.fsu.edu/%5D%5D%20I%20tried%20to%20modify%20the%20included%20MLFG%20generator%20to%20use%20it%20with%20CUDA.%20First%20problem%3A%20the%20implementation%20allocates%20some%20memory%20dynamically.%20Since%20this%20is%20not%20possible%20in%20a%20kernel-function.%20I%20created%20the%20random%20number%20genrators%20%28RNG%29%20from%20the%20host%20code%20in%20device%20memory%20and%20provided%20the%20kernel%20function%20with%20pointers%20to%20the%20gernerators.%20This%20is%20highly%20unsatisfactory.%20Imagine%20asaituation%20where%20you%20want%20to%20have%20thousands%20of%20RNGs.%20Creating%20the%20RNGs%20should%20be%20done%20in%20the%20thrad%20at%20not%20on%20the%20host.%0A%0A%21Second%20attempt%20%28%5B%5Bdownload%7Chttp%3A//www.dirkmann.org/mlfg2.tgz%5D%5D%29%3A%0ADue%20to%20the%20unsatisfactory%20result%20in%20the%20first%20attempt%20I%20removed%20all%20dynamic%20allocation%20from%20the%20MLFG%20and%20hence%20was%20able%20to%20create%20it%20for%20each%20thread%20individually%20within%20a%20kernel%20function.%20%0A%0AThe%20code%20works%20but%20there%20are%20differences%20between%20calculations%20done%20on%20the%20device%20compared%20to%20emulation%20mode.%20The%20algorithm%20includes%20some%20treatment%20of%20uint64%20which%20includes%20a%20multiplication%20with%202%5E-32.%20This%20seems%20to%20introduce%20the%20observed%20differences.%0A%0AI%20did%20not%20resolve%20this%20issue%20yet%20but%20moved%20to%20another%20RNG%20from%20the%20%5B%5BSPRNG%7Chttp%3A//sprng.cs.fsu.edu/%5D%5D%20which%20brought%20a%20feeling%20of%20%5B%5Bsuccess%7CLCG%20success%5D%5D.%0A%0AFind%20the%20current%20state%20of%20development%20%5B%5Bhere%7Chttp%3A//www.dirkmann.org/mlfg2.tgz%5D%5D.%20The%20code%20is%20not%20documented%20well.%20Just%20compile%20main.cu%20with%20nvcc%20and%20let%20it%20run...
%21LCG%20-%20Random%20Number%20generator%20%28%5B%5Bdownload%7Chttp%3A//www.dirkmann.org/lcg.tgz%5D%5D%29%3A%0A%0A%21%2126.%20June%20%3A%20update%20LCG%20for%20CUDA%0AI%20have%20a%20new%20version%20of%20the%20LCG%20available.%20The%20code%20has%20been%20stripped%20down%20significantly%3A%0A+%20No%20internal%20generation%20of%20primes%20anymore%3A%20The%20original%20code%20only%20stored%201000%20primes%20and%20generated%20more%20on%20the%20fly%20when%20needed.%20One%20prime%20is%20used%20for%20a%20new%20random%20number%20stream.%20I%20generated%20the%20first%2020000%20primes%20%20%20and%20copy%20them%20from%20host%20memory%20to%20constant%20device%20memory.%20Not%20all%2020000%20fit%20into%20the%20constant%20memory%20so%20there%20is%20a%20define%20to%20reduce%20the%20number%20used.%20Do%20not%20use%20more%20random%20number%20streams%20than%20primes%20are%20available.%20I%20intend%20to%20throw%20some%20errors%20as%20soon%20as%20I%20know%20how.%0A+%20The%20/struct%20rngen/%20does%20not%20contain%20any%20arrays%20any%20more%20and%20all%20the%20adminstrative%20variables%20have%20been%20removed.%0A+%20Many%20functions%20have%20been%20renamed.%0A%0AThe%20tar-archive%20also%20contains%20a%20simple%20Monte%20Carlo%20simulation%20of%20a%20geometric%20brownian%20motion%20to%20value%20a%20plain%20vanilla%20stock%20option.%20I%20alos%20will%20make%20code%20for%20the%20host%20avalibale%20soon.%20%0A%0AAdditioanlly%20I%20will%20post%20a%20standalone%20version%20of%20the%20LCG-code%20from%20SPRNG%202.0%20for%20comparison.%0A%0A%0A%21%212.%20June%20%3A%20first%20implementation%20of%20LCG%20for%20CUDA%0AThe%20LCG%20from%20version%202.0%20of%20%5B%5BSPRNG%7Chttp%3A//sprng.cs.fsu.edu/%5D%5D%20was%20easier%20to%20port%20to%20CUDA%20than%20the%20%5B%5BMLFG%7CMLFG%20pain%5D%5D%20which%20only%20was%20partly%20successful%20up%20to%20now.%20This%20may%20also%20be%20due%20to%20my%20slowling%20increasing%20expeience.%20I%20strip%20dynamic%20memory%20allocation%20and%20some%20functions%20from%20the%20original%20version%2C%20and%20made%20the%20functions%20to%20run%20on%20the%20device.%20More%20or%20less%20straight%20forward%20and%20done%20within%20few%20hours.%20The%20result%20is%20can%20be%20downloaded%20%5B%5Bhere%7Chttp%3A//www.dirkmann.org/lcg.tgz%5D%5D.%20Just%20compile%20using%20%27nvcc%20main.cuu%27.%20Generated%20numbers%20in%20the%20given%20setup%20agree%20between%20device%20code%2C%20emulated%20code%20and%20the%20original%20version.%20I%20did%20no%20extensive%20tests%20up%20to%20now%20though.%20%0A%0AThe%20usage%20will%20be%20clear%20when%20looking%20at%20lcgkernel.cu.%20Make%20sure%20to%20initialise%20the%20prime%20numbers%20before%20using%20it.%20This%20only%20needs%20to%20be%20done%20once%2C%20so%20this%20job%20is%20taken%20over%20by%20thread%20number%200.%0A%0AThere%20is%20plenty%20of%20space%20for%20improvements.%20One%20thing%20I%20would%20try%20is%20to%20store%20the%20prime-numbers%20in%20shared%20memory...
%21Some%20StickWiki%20formatting%20help%20%28just%20html%29%0A+%20quoting%20with%20%27blockquote%27%0A+%20centering%20text%20with%20%27center%27%0A+%20%3Ctt%3Etypewriter%3C/tt%3E%20with%20%27tt%27%0A+%20C%26%2343%26%2343%3A%20get%20the%20plus%20with%20%3A%20%26%20%23%2043%0A+%20%3Cs%3Estrike%20through%3C/s%3E%20with%20%27s%27%0A+%20comments%20%3C%20%21%20--%20Text%20%20--%3E%0A
%21Papers%20read%20recently%3A%0A%0A+%20Towards%20a%20multi-stochastic%20volatility%20model%20for%20CMS%20spread%20exotic%20-%20*Vladimir%20Piterbarg*%20-%20Global%20Derivatives%202007%0A