Matlab

Here follows a set of short but hopefully instructive examples illustrating the problems of getting good performance in Matlab. Note that Matlab 6.5 (and later) has a JIT-accelerator, so, it is important that you keep track of what version you are using.

It is important not to cut-and-paste (it is very slow for some reason; my guess is that the JIT-accelerator is not run in this case). So put the code in m-files and execute it by typing the name of the m-file, as usual.

It is also important to read the last part of Benchmarks (the part about SpeedStep, PowerNow!, Cool'n'Quiet).

• Dynamic memory allocation
   
• Using ()
   
• Avoiding details
  
  
• Fetch, read and run strange.m
  

• Matrix multiplication
   
• Inlining

• Using arrays for indexing
   
  
• Sparse matrices  
   

 Include the source code in your report.