Hello,
I have been given a humongous code which has been implemented at 2013 and the problem is, the code is very slow for the kind of simulation I am running.
Initially, I wanted to reimplement everythings in another faster computer language like Fortran, but it is almost impossible since the code is huge... Hence, I did Matlab profiler to find the bottleneck, I vectorised the code and tried to remove any unnecessary functions..etc. I could save a bit of time, but still no satisfaction.
Could anyone help me to improve the code and speed up? I would really appreciate so.
Numerical method, FVM. application in fast solid dynamics, highly nonlinear problems
As well as bottlenecks, you could find the types of functional evaluations that are computationally expensive and see if these can be approximated (at least initially). I think that MATLAB will allow you to clock how much time the process is spending on various steps.
Also, it may be time to look at the actual math of what the program is trying to do. If for example there are lots of square roots, exponential, or trig functions, you may be able to do better with code that has some pre-computed elements?
If by "slow" you mean that the final few percent of improvement takes disproportionate time, you might be able to save some approximation, and do some restarts with better / different finite steps? You may also be satisfied for initial exploratory experience with an approximation? If the program is really expensive and there is not a way to improve, you should invest in some way to do warm re-start from previous runs.
1) Use parfor loop.
2) Instead of for loop, try to implement vectorization.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running