Search results

Why "academia and government" are not legit? Anyway... I can speak from my personal experience only. I recently completed a project for large Oil&Gas company, with critical parts of it written in Fortran - and it was not legacy code, but instead an implementation from scratch of numerical...

It was not mentioned that the person in question, that completed his master degree in computational math, is going to work in the quant finance domain, so your advice is not much meaningful - on the contrary, if he's going to specialize in any sort of applied math domain (like molecular...

Please find attached functions for Newton method (newton.m) and composite Simpson method (compositeSimpson.m), as well as corresponding test files (newton_test.m and compositeSimpson_test.m). As mentioned in comments, these simple scripts are written according to Burden & Faires "Numerical...

I guess previous poster had MATLAB Coder tool on his mind. I used this one sparingly. As with any source-to-source translator, it's hard to get maximum performance out of the generated code, and going for better performance is typically the main reason for considering translation. Still, MATLAB...

Oh, and: This formula is applicable for calculating speedups between configurations with one or more processors of the same type. Between completely different processors, it doesn't make sense - one just have to use absolute timings ratio; but I agree that in these cases speedup numbers are...

I suggest you get in touch with the article author for all the clarifications, I'm not particularly familiar with GPU implementation of ADI method. I cited the article mainly as you questioned the possibility of an ADI implementation on GPU, basing your reasoning only on SIMD/MIMD/...

Daniel Egloff, GPUs in Financial Computing Part III: ADI Solvers on GPUs with Application to Stochastic Volatiliy, Wilmott Magazine, March 2011, 51-53

When I mentioned that Pfister's blog post is rather old, it was sort of apology on his behalf - I guess he understood CUDA architecture better in the meantime, and that these days he wouldn't make factual errors like one I pointed in the excerpt you cited. Maybe I should have been more clear...

Have Mira chair for 4 years, I think. Don't know if you're interested in some specific details, but overall it's great chair, and I particularly like mesh material and design of the seat itself. Also, I'd say user reviews on Amazon page you linked above are spot on. (On the other side, my...

Right, and this is exactly why I mentioned the publication year in my previous post: even 8 years time ago, many parallel programming practitioners considered Flynn categorization obsolete. I claim that for GPU programming in particular, this categorization is not helpful in clarifying...

Indeed, thanks for the correction. I guess it was subconscious typo, as I find Michael J. Quinn work more useful exactly in the context that is discussed above - for example his book is great both as tutorial and reference for MPI (and, to some lesser extent, for OpenMP), but also surprisingly...

Never claimed that they are not important, I just said that they are completely and utterly irrelevant for serious number crunching work. Well: "Quant finance is a niche all to its self. It demands special skills and disposition. You have quant developer/analyst/whatever !=...

When I mentioned "re-hashing the obvious" above, I meant on patterns for parallel programming, not on design patterns in general. I remember when first reading GoF book, there was lots of new and valuable stuff for me in-there, but when skimming through Mattson book, it was all the time just...

Well, Laplace transform of given function is: (F(s)=\int_{0}^{\infty}\frac{e^{-st}}{(t+3)^2}dt) Through partial integration with: (u=e^{-s t},dv=\frac{1}{(t+3)^2}dt) this gives: (F(s)=\left.-\frac{e^{-s t}}{t+3}\right|_{0}^{\infty}-s\int_{0}^{\infty}\frac{e^{-s...

Let me Google Wolfram Alpha that for you: here.

Great. Why don't you implement on an accelerator architecture, and profit? My point: I was asked to implement fastest possible FD solver on GPU, so I did it (and, as an aside, sort of re-assured myself that C++ is crap for any sort of numerical work - and this is why I mentioned this whole...

(Sorry for another distraction from the thread subject.) I should have been more specific: In this particular case, it was about solving PDE using finite differences approach, implicit variation; which in turn boils down to coming up with fast sparse solver implementation. There are loads of...

That's probably because you're in touch only with people from this narrow field of quant finance, and for whatever reason people in this field have strong bias towards C/C++... Many people doing some serious work (Oil&Gas, CFD, CEM, and any other scientific computing discipline) are happily...

Note sure about this particular situation, but Glassdoor may be your friend.

I find that these lengthy rants about which programming language is good, and which is not, are typically pointless (and also typically mostly written by people without too much actual practical experience, thus useless). Namely, in real world these things work in very simple manner: either you...

These are pretty broad statements, true for some problems, untrue for many others... In any case: CUBLAS is BLAS implementation delivered along with CUDA toolkit. There are also several high-quality libraries implementing LAPACK-type stuff, like CULA or MAGMA. So dense matrix operations are...

I hope, for your future employers sake, that with this sort of attitude you are never allowed to touch any piece of performance critical code...

There should be "Cygwin-X" group in your system menu, and "XWin Server" item there - run this one, and then either run emacs from the shell window that will pop-up, or run "export DISPLAY=:0.0" and then "emacs &" from your Cygwin shell window. Alternatively, you may wish to install an Emacs...

It's because of rounding to default precision on output. You'll need to add in something like std::setprecision(16) into the mix, and you'll see that value j is actually less than 1 at the end of this piece of code.

OP complained about the compilation problem, and putting the block around object construction and usage code is a solution; everything further is speculation without looking into the rest of the code. Still, it's pretty much obvious that this piece of code is driven by user input on terminal, so...

Put initialization, and then usage of the object, into a block, like: case 1: cin >> x >> y >> r; { ObjectA a(x, y, r); // do something with a here } break; and it will compile then. Then, read about scopes from your C++ book.

I'm not sure I understand properly what you're trying to do, but you may wish to organize your Matlab code so that you have script GenPaths.m with GenPaths() function definition inside: function SPaths = GenPaths(S0, mu, sigma, T, NSteps, NRepl) dt = T / NSteps; % Time increments nudt = (mu...

See here.

For TeX related stuff, you're probably much better if you could use tools under Linux. Still, you may wish to try TeX Live - I know MiKTeX was first choise of TeX user under Windows for long time, but TeX Live has reputation of the most vibrant TeX distribution nowadays.

So this should be some sort of excuse for this sort of stupid "tests"? If recruiters are not technical, then why they bother interviewing/selecting technical people? I was not aware there exist other scams alike to Brainbench... And you're right that Brainbench questions are at least...

Brainbench is the most brain damaged C++ test ever. @OP: consider yourself lucky that you get rejected, you certainly would not want to work for morons using something that stupid to select among candidates.

It is actually possible to take individual courses from UW applied math program - see about "nonmatriculated student status" at this page.

Hard to tell, as I don't know which courses are offered at your department... But, try to browse other alike threads on this forum, you'll find much advice on the preparatory materials/courses for doing quantitative finance related work from other (much more knowledgeable in that than myself)...

I know you can't :) Joking aside, my point was that these days, the most appropriate languages for learning numerical algorithms are ones with built-in support for vector operations, like Matlab or Fortran - this is both in order to be able to concisely express individual steps for many...

Burden & Faires is fantastic Numerical Analysis book; so if possible for you, then by all means go for NA class offered by your math department - these guys obviously know what they are doing. Then, complement it with some sort of parallel programming class from your CompSci department and...

...and plain C is much better of all of them, and then Matlab is much more appropriate of anyone mentioned, and finally Fortran rules them all.

My follow-ups to your posts actually are not directed to you - it's pretty much obvious you're quite stubborn in doing things the way you think they should be done, and that's certainly fine with me. However, it is also a fact you're flooding this forum with all kind of opinions and thoughts...

I just don't see what sort of deeper understanding could be gained from designing options classes hierarchy - it's not any sort of rocket science and one way or another, you just have to end up with something very similar to what QuantLib guys did, so why bother. I'd say there exist many other...

Why not just checking how QuantLib guys built their hierarchy of classes (hint: their Option class has LazyObject as its base class, and inherits directly from Instrument class)?

You could run as many threads as your operating system allow, however too many of them are just going to thrash. So one would usually check, in the specific processor documentation, how many threads could be executing simultaneously, and then profile the code around that number to come up with...

Yes - this is how people are doing matrix multiplication for the last 40 years, or so.

Sure, QuantLib is first coming to my mind.

It is not about luck - it is about the quality of your code, and the effort you put into making your work on this code noticeable. So, in that regard it's the same thing as when you do your internship in a company. The difference is that the company won't allow you to show this code to anyone...

This is completely wrong. When working on an open-source project, most of the time you are in contact with end users, so your work is much more recognized than when writing software as a corporate drone. Sometimes, these contacts with end users could even land you a paid job, on a same project...

Care to point to Web page of announcement/add or something alike? University of Washington has rather respectable online applied math MS program (and also as mentioned here on the forum they're now into Computational Finance & Risk Management online MS program), so these guys have experience...

The last argument of "lindex" command could be actually a list of indices, feature that not that much languages supported back at a time; still I can imagine (I never used TCL much) the pain... Perl is on the other extreme of the spectrum: so many syntactic sugar, so that you type minimally...

You're right: indeed EDA is another niche where Perl still stands strong, albeit I'd say TCL use is even stronger in that domain. Related note: for alike scripting languages, as well as many general purpose programming languages, it is often very important to be able to wrap lower-level C/C++...

Perl and Fortran communities are quite different. Perl is mainly popular among older Unix users - it was practically first (development started in late eighties) scripting language in widespread use, with text processing facilities as its main strength (especially its excellent built-in regexp...

Yes - Fortran is indeed still popular in specific branches of science (and academia, to the some extent), mostly in the domain of applied math.

Indeed, seems like everyone and their grandmas have accounts and use Skype these days - as mentioned above, Skype guys did great job on grabbing the market. Not much: Asterisk is for cases you want to switch over your company phone infra-structure to SIP - I'm using SIP just as an end-user, so...

Skype is evil, as they use proprietary protocol and their call-out (to call from Skype app to regular phones) rates are rather expensive. There exist many providers for this sort of communication using standard SIP protocol, and most of smartphones have native support for SIP, together with...

Emacs 23.3.1 (on Linux): Press Ctrl+space at the start of region of text I intend to copy, press Ctrl+v several times to select several screens of text, press Alt+w to copy text, move cursor to where I want to insert text, press Ctrl+y to paste - all fine. Don't really know what sort of problem...

"Unix is user-friendly. It just isn't promiscuous about which users it's friendly with." – Steven King I won't even start on this one, I'll just say I had a good laugh...

I've used TBB to the some extent (but not on a quant project). Don't know about "Task Parallel Lib", so I cannot compare here. Basically, TBB is OK for expressing task-level parallelism in situations when one want/has to go with C++; however, most people in HPC community are sticking with C or...

It's very hard to generalize, as we're practically talking all the world except US, Canada and Western Europe here. My message was sort of reply to @Alexei Smirnov comments, as he was particularly vocal about how the perspective is pessimistic in Russia, and I'm also living in an Eastern...

Obviously, I was not talking about this type of jobs, as these are not topic of this discussion - I was talking about jobs related to products design/development, or even to research (to some extent).

I don't think that the perspective is that bleak, there exist number of factors not mentioned in the discussion above that are counter-balancing, to the some extent, the fact that workforce mobility is these days more easier than ever. For example, lots of people are just staying in their home...

There exist many types of programming work. Certainly a well-defined process is to be followed when writing code for pacemaker, or anything alike. But not that many programmers are in position to work on this type of software. Still, for many of them, some sort of process is needed, simply...

What on Earth this has to do with this thread? As for your question: JFGI.

Interesting that throughout all of my work experience, whenever I was part of an organization trying to push developers beyond level 1, it ended up as a disaster... On the other side, level 1 always served me very well, so I guess I'm just going to stay there.

Actually, current Matlab version is 2011a, but indeed from version 2010b there is some sort of CUDA support. However, MathWorks guys are lagging significantly behind AccelerEyes guys in that regard - Jacket plugin by AccelerEyes has much broader scope, and faster implementation, regarding...

With all due respect to your son, your comparison is sort of insulting for my intelligence - do you really find it appropriate? Again: all of my comments were about the context of doing programming work. I readily admit what I wrote above is over the top, but it's just that I find these needs...

I'd rather move my fingers (Alt+Tab) than my neck. I didn't get that part - which tabs and graphs? As a matter of fact, recently I've tried one of those "netbook" variations of the desktop environment, where there is always single full-screen top-level window active only (and you do Alt+Tab...

Monitors are not that costly these days, but still it would be silly to buy one just for the sake of "trying" something that I know up front I don't need (even borrowing it from someone for this purpose would be cumbersome)... I was asking for clarification regarding the type of work out of...

"A must" - for what type of work? I'm sitting in front of single monitor, doing programming work, for more than 20 years now (last 15 or so in front of laptop screen actually), and I find no use whatsoever for more - I have xterm open on the left side of my screen, emacs on the right side, and...

Latest Recruitement & Hiring Survey from Notices of the AMS.

If you think about coding matrix multiplication from scratch to be executed on GPU, then indeed there is better chance your could fare well this way - GPU architectures are actually typically simpler regarding achieving decent level of speedup for the parallel implementation of given algorithm...

This question was answered for you, some months ago, on this thread: Matlab is faster because it probably uses optimized version of BLAS library for matrix multiplication. You cannot come even close (if you didn't get it - I was ironic in my previous post about "impressive" speeds of C# codes...

No, I don't think so. For example: if I have some sort of code using BLAS/LAPACK, or say some sort of Monte Carlo solver code, I see no way on how this code could benefit from an effort like kgpu. And most of the code interesting at least to community around this forum is of this type...

I can't see exactly how something alike would be possible in general. Kernel could speed-up execution of a program only if program requests a service from kernel, through a syscall, and kernel has this syscall implementation improved. However, the scope of syscalls that GPU could help in...

Maybe I didn't understood your point properly here, but I don't see kgpu as something "moving GPU closer to end users". The whitepaper explains clearly what is the intention behind kgpu - it is an interesting development, but solely intended for making it possible for Linux kernel code to...

Amazing indeed: managed to extract whopping 100 MFLOPS out of machine that is probably north of 50 GFLOPS peak.

Strang books are certainly not perfect. On the other side if you know about the subject, then most of the time you'll find something you think it's important missing from any kind of book - for example, Lipschutz book you mentioned above is not discussing (as far as I can see from book index on...

Video lectures cover "Introduction to Linear Algebra" book almost completely, and mostly in order (only applications chapter material is scattered throughout the course) - you could check for book sections matching video lectures in Readings section of OCW 18.06 site. The book itself is rather...

If going with Strang for Linear Algebra, then instead of this one, I'd recommend Linear Algebra And Its Applications. Strang is brilliant teacher, but his books are infamous for its "conversational" style, which makes Introduction to Linear Algebra book rather unreadable (even if reader is...

OUP is excellent choice too, but I think you don't have to be so dismissive towards OCW either. Lots of courses, including 18.03, offer lecture notes, as well as homework (i.e. psets) and midterm and final exams, oftentimes together with solutions. Moreover, usually it is possible to find...

As far as ODE concerned: I think if you grab a copy of Edwards and Penney, and watch Prof. Mattuck lectures at MIT, working on psets and exams along the way - that you should be fine.

OT, but gotta say +1 for mentioning these guys: I don't know of any research position more elite than working there. On the other side, their existence is based in having money thrown from hedge fund earnings into what they are doing - it may all pay off someday, but certainly this is not a...

I never said something alike... It's certainly not my intention here to claim that any platform is "better" than other one, or that it's "stupid" to engage in development work on any platform - obviously Unix programming works better for me, but whatever gets the job done for anyone else is...

I've mentioned it above: correct me if I'm wrong, but I thought this whole latest exchange of messages started on the topic of comparing native platform programming interfaces: POSIX etc. for Unix vs. Win32 and everything else that was considered as native API for Windows programming. Qt and...

Once again: Qt and Gtk are not relevant for the discussion here - albeit primarily written on Linux, these are multi-platform toolkits. Motif is Unix native toolkit, and it is not replacing Xlib, but complementing it instead - Xlib has no functionality included for creating GUI widgets (Xlib is...

Each of these options is equally valid/usable for Windows as it is for Linux, and each of them came from trying to achieve bigger goal - multi-platform compatibility. For tasks I've mentioned above, and in the context of single platform native APIs (compatibility of which is, if I understood it...

I was mostly having MFC vs. .Net API on my mind as an example, as both cover most of the functionality needed to write desktop applications: like building GUI, and then maybe connecting to database, communicate over the network, or doing some 2D graphics - I'd say all of this functionality get...

I'll comment on this one only, as it is typical for most of the items you mentioned (except for embedded related stuff - I don't care about these, so I don't know much about it). The situation is rather clear with memcpy(): it was never supposed to allow overlapping between source and...

Now when this year old thread is bumped: can't resist to comment that OSX, albeit based on good old Unix-type kernel (FreeBSD) diverged quite a lot from a Unix system (just look at those silly file system hierarchy decision), so I'd say it feels rather alien to experienced Unix users -...

The discussion was about compatibility in the context of programming interface, and I'd say your statement is wrong in that regard. For Unix systems, POSIX is mature and well-maintained programming API. Also, Unix programmers have choice between number of tools, like GNU autotools or CMake...

Well it does in a sort of way, as Unix is such of a platform that is forcing its users to learn about the system in more details, instead of just clicking around. Although I have to admit that with this new generation of Linux users, coming to use only distributions that are created to be as...

This argument is ages old, and rebutted number of times: see for pointers to some discussions here, but basically Unix machines are much less virus susceptible because Unix has sane system architecture regarding the security, and because Unix users are in general knowledgeable enough not to rush...

No, that statement is completely wrong: most supercomputer machines run Linux (see here), that is in turn written in C; Fortran is the best language for numerical programming, but it is not used for system programming - C rules there.

I completely disagree: I think in case OP intend to get involved into any sort of numerical programming, choosing to learn Fortran is very good decision. With all of the recent updates of the standard, Fortran is now a modern imperative programming language, and having built-in, high-performance...

You could check it yourself here.

If you're on Linux (or Mac, with Xcode installed) then, as said by previous poster, GCC is already there for you - to call Fortran compiler front-end, use "gfortran" command from the shell. If you're on Windows, and used to Visual Studio, then FTN95 may be interesting for you. Furthermore...

I guess my background is atypical (software developer, just dabbling in the domain of quantitative finance related work from time to time, and also doing work off-site so I actually very rarely get involved in any sort of classical job interview), but I'll never forget two occasions: Once some...

It's not my intention to be picky here, and it's indeed hard to find details regarding specs and pricing of these machines now, but: if I remember it correctly ThinkPad W701 was again in the same price range as previous generation of 17 inch MacBook Pro machines, but offering much better...

I don't think so. Basic 15" MacBook Pro model, with 2.0 GHz quad-core processor, is $1799, and I guess one would add at least $100 to escape from silly 1440x900 resolution to 1680x1050. Same money plus $30, gives you ThinkPad W520, with 2.2 GHz quad-core processor, with 1920x1080 screen...

In case someone would really need something alike - why not just access Ideone.com, that these guys seem to be using as back-end, from any browser on any phone? Now, if someone would write really good basketball stats smartphone app, that would be something that would really get me excited...

Living outside US, I had excellent experience with ForwardIt service for ordering all stuff (including laptops) from US. Their service is rather inexpensive (you pay $10 per package to them, and then postal costs from US to your country), and with current dollar exchange rate, buying all kind...

But in turn it will instantly make your code obfuscated...

That's rather old thread, but the main point is still valid: CUDA practically has nothing to do with SLI. On the other side, depending on what you are intending to do, there were some improvements with recent toolkit releases in handling multiple GPUs - for example, CUDA 4.0 introduced...

@Novak: If you're still into job search, try with these guys. It's not exactly quant work, but their development team is based in Paris, and seems to me they could put your knowledge to some good use.

3816547290.

Yes, you move to Win HPC in this case as you have no other choice: what you described is exactly the situation that MS is trying to put their customers into - lock-in.

We're not talking about this type of person here, we're talking about person that is just willing to learn some basics of SQL. On the other side: with most of commonly used client/server RDBMS already mentioned in this thread, allow me to mention also DB2 Express-C here - one could not get more...

I'm not sure Joshi books is good place to start - the book is not teaching much about basics of quantitative finance, and also is insisting on peculiarities of only one of possible coding approaches here (to fully comprehend it, I'd say very good knowledge of C++ is needed, and at least GoF book...

Corrected simulation code from above: #include <stdbool.h> #include <stdio.h> #include <stdlib.h> #define N 100000000 int main(void) { double sum = 0; for (int i = 0; i < N; ++i) { double curr = (double) random() / RAND_MAX; double next...

I wouldn't, either - according to man page, the period of random() is approximately 16 * ((2^31) - 1). But it's readily available in C standard library, so it makes code short and does its job well here. On the other side, Mersenne Twister is not be-all end-all for simulations - oftentimes...

If array allocated, there is a limit (depending on the size of memory on your machine) on the number of simulations runs, while without arrays practically (the data type used in my code to hold counters and such should be changed for some very large runs) the number of runs is unlimited. Also...

Rather inefficient (why allocate array?)... Here is C code: #include <stdio.h> #include <stdlib.h> #define N 100000000 int main(void) { double sum = 0; for (int i = 0; i < N; ++i) { double curr = (double) random() / RAND_MAX; int j = 1...

Then check this: http://www.sqlite.org/famous.html. Also, Bloomberg is sponsoring SQLite development, so I guess they have some interest too... Now, SQLite certainly is not in the same league as other DMBS mentioned, it's much simpler and slower. But it has its place in some simple projects...

UW has very strong department for Applied Math, and they are offering well respected online master program in Applied Math from some time ago: http://www.amathonline.uw.edu/. As far as I can understand, this new program is also available online, so that may be selling point here too.

Just to clarify my previous post a bit: There are two aspects of CUDA/OpenCL programming. On one side, there are always pieces of code to be written that are going to be executed on CPU - this code is for copying input data into GPU memory, specifying GPU execution configuration, launching the...

I'd say you're quite wrong here. There exist lots of high quality numerical codes written in Fortran, still unsurpassed by anything written in C or any other language. And there are new developments in Fortran too, enough of a market that there are dozen or so companies fighting for it with...

Probably you either haven't compiler your code with debugging symbols included in or, on the other extreme, maybe you turned on such a level of optimization that the code you're trying to insert a breakpoint in is optimized out. But - it would be much better if you could create some small...

Frequencies of use, like one you posted above, are meaningless most of the time, and are always absolutely meaningless when there is no further info on what sort of sample is used to gather statistics. But that was not my point. My point is that oftentimes in this sort of discussions, people...

The statistics that would be really useful in this sort of discussions would be if people just state how many lines of code they actually wrote in programming languages they are talking about (usually in very abstract terms)...

Atlas is not hand optimized. On the contrary, the whole idea behind Atlas is to perform, for each machine it is going to be installed on (for optimum performance, Atlas should be always compiled from source - never use pre-built Atlas installation, as it is probably not tuned exactly for the...

Check out http://www.glassdoor.com/ too - different type of site, but I think extremely useful in learning what it takes to get through a job interview for a particular company and also, from employee reviews, how it feels actually working for it.

I've linked it in another thread on this forum (this one was also about C++/OOP, these discussions just seem to have to pop up over and over again), and I'll link it once again: http://www.faqs.org/docs/artu/unix_and_oo.html, from Eric Raymond's "Art of Unix Programming". This is the quote...

This is wrong, C is not "written in assembler" - most of the contemporary C and C++ compilers are written in C (recently, some in C++: for example, the hottest compiler project at the moment is LLVM http://llvm.org/ and it's written in C++), but this has nothing to do with the speed of the code...

To be honest, this is for the first time I've heard about Armadillo - from the description, it doesn't seem like C++ wrapper for BLAS to me, and also I really don't know about the performance. So I'd still suggest you go with BLAS implementation for your machine (either Intel MKL for Intel...

Changing IDE won't make your code more standard compliant - this is up to programmer, eventually with some help from compiler through corresponding flags that would check for the standard compliance etc. But indeed it's true that the community of programmers using VC++ is somehow least caring...

Very, very smart move, in my humble opinion.

If you just want to learn SQL syntax, in order to interact with existing databases, then "Learning SQL" book should serve you great. As far as choice of software to be used while learning concerned, I'd recommend SQLite (http://www.sqlite.org/) - all of Linux/Windows/Mac supported, database is...

As said above: Matlab does it in 2 secs because it uses some sort of vendor supplied BLAS under the hood. So your C++ or C# code should do it too. None of previous posts mention that any of you guys is doing this for the learning purpose. But in that case, indeed it makes sense to try; still...

For the sake of Christ: use your vendor supplied BLAS (http://en.wikipedia.org/wiki/Basic_Linear_Algebra_Subprograms). Matlab does it, as is any other numerical software under the Sun. Nobody should be writing own code for matrix multiplication.

Well I haven't run any kind of battery of tests over this code to be sure (you should certainly do this in case you intend to deploy alike code in any kind of production environment: use either some well-known Gaussian quadrature code, or a tool like Mathematica, to calculate integrals of number...

The procedure above will produce correct answer, but for the fixed integration interval [0,2*delta]. I guess you may wish to change the code, in order to support integration over the arbitrary interval [a,b], to something as follows: class Test { static double F(double x) {...

Matrix inversion is O(n^3), and Gauss-Jordan is the way to go if you want to do it by hand. Performance-wise, the only jumps you should see when increasing n should be when you reach points when your matrix is not able to fit cache (L1, then L2) anymore; otherwise the execution time should...

Probably mistaken with Fortress programming language - both appeared approximately at the same time, and I remember I too was confused about which one is which back at the time. F# seems rather nice language to me too, even surprisingly nice for something coming from Microsoft... And - I'm not...

I don't consider this particularly important - I just don't expect GPU programming to become some kind of skill required from an average programmer, so that there may be need to lower barrier to entry in any way possible, including allowing them to program kernels in language that is presumably...

Suggestions from Joy would probably have much more weight than mine, but I'd like to add my 2 cents regarding good book on some of above topics: As for C++, back then when I was learning the language, the "C++ Primer" book by Stan Lippman served me very well. On the other side, for someone...

If you really want to run executable from your makefile, then you could certainly do, and one possible approach is as follows (beware of indented lines, you should use tabs there, as mentioned in previous reply): .PHONY: all run all: Main run Main: Main.cpp Headers.h g++ -o Main Main.cpp...

Check for this on Glassdoor and alike sites - you'll find plenty of employees reviews, and salary and interview information.

Regarding popularity: among (serious) developers using Unix/Linux as their primary development platform, vi and emacs are definitely most popular editors, used by far more people than any other alternative; on other platforms however, I'd say their usage is minimal. Emacs is probably the most...

I'd second suggestion from alain: when having control over the choice of VCS, these days there is really no reason to choose anything but distributed VCS, like Mercurial (which should probably come more natural to an existing SVN user) or git. Advantages of a distributed VCS are numerous (user...

Statistics like what you mentioned above are always skewed; besides, these in particular seem rather outdated... When I stated that Perl is irrelevant, I was primarily considering how much newbie programmers are learning the language. Throughout past couple years there were number of computer...

They started the language re-design, circa 2000, and still haven't delivered, so the language became mostly irrelevant in the meantime (well, Perl advocates will fiercely claim that the new version, Perl 6, is actually usable while ago, and that old version, Perl 5.x, was actually updated...

Don't know you if you're trying to be ironic here, but actually above is mostly true. For example, if you're up to getting any kind of performance from your x86 node, you simply have to deal with SSE (either through inline assembly, or using compiler intrinsics). And even if you are dealing...

While I'd say good knowledge of hardware is indeed pre-requisite for doing HPC work, I don't think C++ knowledge is needed that much. Fortran and C are languages of HPC, C++ is not.

Sorry, but I'm afraid I don't understand your intention here... The sequence of statements your wrote above is not valid Mathematica code - Mathematica has rather strict syntax, for example all types of parentheses have complete distinct purpose etc. So if you don't know Mathematica...

From what I was able to learn, as an outsider, I'd say many of quant shops are suffering NIH syndrome, and are implementing lots of stuff from scratch. Doing this for low-level stuff, like linear algebra routines, is sometimes needed, in case new approaches like GPU or FPGA appear with...

Now you have not one, but three parentheses missing in your formula for R(T)... In any case - I don't have Matlab copy available, but I do have Mathematica installed on my machine, so I thought running your data through Mathematica regression routine. So if I try with R(T) formula you provided...

You have missing parenthesis in the formula for R(T), so it is not possible to try fitting the data you provided; you also need to write down the value of r0 parameter.

MPI is still big in HPC today; but probably there isn't much talk about it because MPI is the dominant standard for the domain of message passing parallel programming (i.e. programming for clusters of machines connected by high-speed networks), while there is still no dominant standard for the...

Ken Thompson's opinion is actually typical (well, sure it is: Ken is father of all of it, after all) opinion about C++ in the world of Unix programmers community (note that I don't mean on kernel programmers only here, but on all of us old-school guys that learned programming on Unix systems...

I'd like just to jump in with some reading recommendations for brushing up Unix skills. As for work with Unix shell and commands, I'd recommend good old A Practical Guide to the Unix System by Mark Sobell. Note also that there exist more recent, Linux based, versions of this book, but in any...

From an outsider (my background is in Comp. Sci./Eng. and Applied Math) point of view, I was too somewhat surprised with the obsession with C++ among quants (well - among admittedly very limited sampling of quants that I was in touch with, although I seem to get the same impression from lightly...

Never used it in the context of wrapping for Web, but I find SWIG to be by far the most mature solution for wrapping C/C++ code for myriad of scripting languages.

Andy, I was working as an external HPC consultant on this particular CUDA based options pricing project, so at the moment I'm still bound with various NDAs etc. Now, I can certainly provide you with contacts on the PM, if you would like to discuss the details with these guys (probably be...

I too was recently heavily involved in writing pricing code in CUDA, and was also doing lots of HPC work on other platforms in past couple years, so here are some of opinions of mine: I'd say that going with NVIDIA hardware, and CUDA API, is still the safest option for anyone looking for...

You would be definitely better with doing all of the stuff on the GPU - CUDA is all about pushing as much processing as possible to do in parallel into the GPU. I guess you're doing Monte Carlo, so regarding this you may wish to start with RNG code examples from CUDA SDK; the other part...

Sage is nice, and its functionality indeed overlaps with Mathematica in several aspects - it even supports kind of notebook interface, running in browser. There are differences, though. Sage is more like set of Python interfaces to various (Python or non-Python) scientific packages, so it...

I'd just ignore - in my (10+ years) experience, if a company insist on sending CV in Word format only, they're not worth the effort at all. The main point of (La)TeX is exactly insisting on the form. Do you think that you know more about typography and related stuff than one of the most...