Increase the Efficiency of Plotting G-code

Hi,

I have written a code that reads a .gcode file, finds the x, y and z values and then plots a graph so you can see the part however, its quite slow to run and was hoping on some ideas on how I could speed it up.

I think its the part where I specifiy to only draw a line if the previous g-code line is a G1 (which is an instruction to extrude in g-code), I have highlighted this section in bold.

Any help would be great. I have attached the code below.

%%

% fid = fopen('VK_Holllow_Pyramid.gcode');

clear

tic

fid = textread('VK_Holllow_Pyramid.gcode','%s','delimiter',';');

lineNumLaura = 1;

xVal = 0;

yVal = 0;

fVal = 0;

gVal = 0;

layerNum = -1;

for lineNum = 1:size(fid,1)

v = strsplit(fid{lineNum},' ');

if regexp((v{1}),'LAYER:*') %Find layer number

layerNum = layerNum + 1;

end

% check if first value in 'G' to mark start of data

if regexp((v{1}),'G*')

if strcmp(v{1}(2),'0') || strcmp(v{1}(2),'1')

laura(lineNumLaura,4) = str2double(v{1}(2));

for i = 2:size(v,2)

if ~isempty(regexp((v{i}),'X*')) || ~isempty(regexp((v{i}),'Y*'));

laura(lineNumLaura,1) = layerNum;

if regexp((v{i}),'X*')

xVal = v{i};

laura(lineNumLaura,2) = str2double(xVal(2:end));

end

if regexp((v{i}),'Y*')

yVal = v{i};

laura(lineNumLaura,3) = str2double(yVal(2:end));

lineNumLaura = lineNumLaura + 1;

end

% if regexp((v{i}),'F*')

% fVal = v{i};

% laura(lineNumLaura,4) = str2double(fVal(2:end));

% lineNumLaura = lineNumLaura + 1;

% end

end

end

end

end

end

%%

toc

z = laura(:,1);

x = laura(:,2);

y = laura(:,3);

g = laura(:,4);

fig = figure('visible','off');

for i = 2:numel(z)

if g(i-1) == 1

plot3([x(i-1), x(i)], [y(i-1), y(i)], [z(i-1), z(i)],'LineWidth',0.8, 'color', 'r')

hold on

end

end

set(gca,'visible','off')

% set(gca, 'Projection','perspective');

set(gca, 'CameraPosition', [100 0 0]);

NoLayers = max(z(:,1));

figure(fig)

hold off

[Jan]

Please format your code using the buttons on top of the edit field. Then the code is easier to read.

Why do you assume, that the plotting consumes more time than the reading of the file?

regexp((v{1}),'LAYER:*') is much slower than strncmp(v{1}, 'LAYER:', 6).

You execute (regexp((v{i}),'X*') twice, which is a waste of time.

Answer 1

[Jan]

Why do you assume, that the plotting consumes more time than the reading of the file?

regexp((v{1}),'LAYER:*') is much slower than strncmp(v{1}, 'LAYER:', 6).

You execute (regexp((v{i}),'X*') twice, which is a waste of time.

For plotting - replace

for i = 2:numel(z)
    if g(i-1) == 1
        plot3([x(i-1), x(i)], [y(i-1), y(i)], [z(i-1), z(i)],'LineWidth',0.8, 'color', 'r')
        hold on
    end
end

by

match = (g == 1);
m1    = [match(1:end-1); false];  % [EDITED, typo fixed: , -> ;]
m2    = match;
m2(1) = false;
plot3([x(m1), x(m2)], [y(m1), y(m2)], [z(m1), z(m2)], 'LineWidth', 0.8, 'color', 'r');

Does this work? I cannot try it due to the lack of input data.

[Laura Carnegie-Brown]

It gives the folowing error message:

This code should be usabe with any .gcode file.

% Error Message 

Error using horzcat
Dimensions of arrays being concatenated are not
consistent.

Error in helptest (line 62)
m1    = [match(1:end-1), false];


% Code

% fid = fopen('VK_Holllow_Pyramid.gcode');
clear
tic 
fid = textread('VK_Top_Hat.gcode','%s','delimiter',';');
lineNumLaura = 1;
xVal = 0;
yVal = 0;
fVal = 0;
gVal = 0;
layerNum = -1;

for lineNum = 1:size(fid,1)
    v = strsplit(fid{lineNum},' ');
      
    if regexp((v{1}),'LAYER:*') %Find layer number  
        layerNum = layerNum + 1;
    end
        
    % check if first value in 'G' to mark start of data
    if regexp((v{1}),'G*')
        if strcmp(v{1}(2),'0') || strcmp(v{1}(2),'1')
            laura(lineNumLaura,4) = str2double(v{1}(2));
            for i = 2:size(v,2)
                if ~isempty(regexp((v{i}),'X*')) || ~isempty(regexp((v{i}),'Y*'));
                    laura(lineNumLaura,1) = layerNum;
                    if regexp((v{i}),'X*')
                        xVal = v{i};
                        laura(lineNumLaura,2) = str2double(xVal(2:end));
                    end

                    if regexp((v{i}),'Y*')
                        yVal = v{i};
                        laura(lineNumLaura,3) = str2double(yVal(2:end));
                        lineNumLaura = lineNumLaura + 1;
                    end
                end
            end
        end
    end
end
%% 
toc
 
z = laura(:,1);
x = laura(:,2);
y = laura(:,3);
g = laura(:,4);

% fig = figure('visible','off');

match = (g == 1);
m1    = [match(1:end-1), false];
m2    = match;
m2(1) = false;
plot3([x(m1), x(m2)], [y(m1), y(m2)], [z(m1), z(m2)], 'LineWidth', 0.8, 'color', 'r');


% set(gca,'visible','off')
% set(gca, 'Projection','perspective');
% set(gca, 'CameraPosition', [100 0 0]);
% set(gca, 'CameraPosition', [100 0]);

NoLayers = max(z(:,1));

figure(fig)
hold off

[Laura Carnegie-Brown]

To answer your question on how I know plotting takes more time, I have put a timer on each section to work out where the inefficiencies are

[Jan]

I do not have any .gcode files. Do you want me to search some in the internet?

This line contained a typo:

m1    = [match(1:end-1), false];
%                      ^  must be a ;

I assume, you can fix this by your own.