I felt like playing with Julia and MATLAB this Sunday morning.

I found some code that prices European Options in MATLAB using Monte Carlo simulations over at computeraidedfinance. I ran this a couple of times to see what results I should be getting and how long it would take for 1 million paths:. The result varies because this is a stochastic process but we can see that it should be around The following seems to work.

The Julia code appears to be valid, it gives the correct result of Update 9pm 7th October The code is still slower than the MATLAB version but better than the older Windows build. Over on the Julia mailing list , someone posted a faster version of this simulation in Julia.

MATLAB Tutorial - Monte-Carlo Asset Paths

On the Linux partition of my test machine, this got through paths in 8. Members of the Julia team have been improving the performance of the randn function used in the above code see here and here for details.

Using the de-vectorised code above, execution time for 1 million paths in Julia is now down to 7.

[R-SIG-Finance] Monte Carlo Option Pricing formula R code vs Matlab

It might depend on the number of kernels that are actually used. The time consuming task in the look is the generation of random numbers and the expression.

Walking Randomly » European Option Pricing in Julia and MATLAB

This benefits from threads. As I said here: I do not believe that during this year will be. Just try to reproduce micro-benchmark results!? I should try and run it myself, but did you try starting julia with more than one thread? This blog post led to a discussion on Google Groups [1], which led to a bug report [2], which is now marked as closed. Have you tried to run the original program not the devectorized one again?

GitHub Bug Sorry for spam! Kumar On the most recent Julia build, the original version of the program is a little faster but the de-vectorised version is still streets ahead. The devectorised Julia program now completes the calculation in 7. This monte-carlo pricing algorithm is embarrassingly parallel and so I could explicitly code it for multiple threads in both MATLAB and Julia.

BTW, on my macbook pro 2. We expect that the vectorized julia performance will also improve and at least match Matlab once some optimizations are in place.

This benchmark was crucial towards driving the performance of randn for us, and also led to the systematic testing of the quality and correctness of randn. Thanks and keep pushing! Notify me of follow-up comments by email. Notify me of new posts by email.

European Option Pricing in Julia and MATLAB. October 7th, Categories: Financial Math , Julia , just for fun , matlab Tags: The code is still slower than the MATLAB version but better than the older Windows build Update: Dell XPS LX CPU: Intel Core iQM 2Ghz software overclockable to 2.

Windows 7 Home Premium 64 bit and Ubuntu Original windows version was Version 0.

Several versions used on Linux since, see text for details. Leave a comment Trackback. Euler uses only one. It takes around 70 seconds on my 1.

Starting julia with muliple threads appears to make no difference for the code as written. Thanks to you too Viral…your side of the work was harder than mine: E-Mail will not be published. Subscribe to comments feed.