How to use a multi core computer to implement parallel computing in Matlab

Start the parallel computing function. For example, if you start the parallel operation function and use parfor end, your computing task will be calculated using multiple cores of your computer. If the amount of computation is not large, it is not necessary, and the time saved is not enough to allocate tasks.

What the? The computer doesn't have multiple cores? Those two options don't work.

1. First of all, open MATLAB on the multi-core computer to enter the command window, and click "New Script" in the shortcut toolbar, as shown in the figure below.

2. In MATLAB, the imopen function can be used to open the image, and the function also needs to specify an element structure as the second parameter, plus the ** to read ** and display the before and after comparison**.

3. Click the "Save and Run" button, as shown in the figure below.

4. Select the location where you want to save the M file and set the file name, as shown in the figure below.

5. In this way, the image is opened in MATLAB, and a dialog box will pop up to display the comparison results of the original ** and the open operation image, as shown in the figure below.

MATLAB Parallel Operations.

At present, most of the newly purchased computers are multi-core, so how can you effectively use multi-core when using MATLAB for a large number of calculations? The current version of MATLAB already supports multi-core parallel computing. Yes, the MATLAB version used is R2007B.

The computer is dual-core.

Let's try it out first:

matlabpool local 2

submitted parallel job to the scheduler, waiting for it to start.

connected to a matlabpool session with 2 labs.

shows that a multi-core configuration is in progress, and then prompts to connect to 2 labs. I also understand this: there are 2 workstations that can run MATLAB locally, so that parallel computing can be performed with the distributed computing toolbox (the matlabpool command seems to be in the parallel computing toolbox).

testparallel

elapsed time is seconds.

In order to perform multi-core parallel operations, the original for loop should be replaced by parfor in testparallel.

While running this, observe the Windows Task Manager, and you can see that there are a total of 3 processes. One of them occupies more memory, it should be the main control, he basically doesn't work, only responsible for allocation, his CPU occupancy is only 1 2% when calculating, and the remaining two processes are dedicated to computing, each of which accounts for about 49% of the CPU. It seems that each matlab process is still a single-core operation, but two processes are opened at once, so the CPU can be used up.

After running TestParallel, the CPU of all three processes is immediately reduced to about 1%.