Line intersections

#21 Posted: 2/16/2018 9:01:36 AM

992 likes in 1675 posts.

Group: User

Wrote
Just an exercise on Smath to test functionalities and capabilities.
Try the attached file and play with it.

Thanks
Marco
InterX.sm (85 KiB) downloaded 168 time(s).

Hi Marco. Here some comments about your great file. As you suggest, I play with it.
Best regards.
Alvaro.

InterX.r1.sm (77 KiB) downloaded 83 time(s).

1 users liked this post

Andrey Ivashov 2/16/2018 12:30:00 PM

#22 Posted: 2/16/2018 11:55:14 AM

Jean Giraud

983 likes in 6866 posts.

Group: User

Wrote
Try the attached file and play with it.

1. Marco => still computing after 15 min ? !
2. Alvaro => red from top to bottom

#23 Posted: 2/16/2018 12:24:05 PM

Jean Giraud

983 likes in 6866 posts.

Group: User

Wrote
Try the attached file and play with it.

Attached is about timing in work sheet.
Observe the sequential timing [black].
t0:=time(1) is in collapsed area.

Contour_8 Hypersurface.sm (1 MiB) downloaded 59 time(s).

#24 Posted: 2/16/2018 12:31:14 PM

Andrey Ivashov

2271 likes in 3734 posts.

Group: Super Administrator

Wrote
Hi Marco. Here some comments about your great file. As you suggest, I play with it.
Best regards.
Alvaro.

1 min 53 sec on my windows tablet pc.

#25 Posted: 2/16/2018 1:09:46 PM

Fridel Selitsky

520 likes in 451 posts.

Group: User

Wrote

1 min 53 sec on my windows tablet pc.

SMath Studio 0.99.6611 time(0)-to =83.735 s on my windows10

#26 Posted: 2/16/2018 1:31:34 PM

NDTM Amarasekera

130 likes in 352 posts.

Group: User

Wrote
Originally Posted by: Andrey Ivashov [url=/forum/yaf_postsm47126_Line-intersections.aspx#post47126][/url]

1 min 53 sec on my windows tablet pc.
SMath Studio 0.99.6611time(0)-to =83.735 son my windows10

Time taken in my win 10 laptop with SS 0.99.6619 shown.

Look within!... The secret is inside you. Best Regards Eng. NDTM Amarasekera - Sri Lanka

#27 Posted: 2/16/2018 1:54:42 PM

Davide Carpi

1417 likes in 2873 posts.

Group: Moderator

5 min 25 s (win 10 x64, intel i5 quad core 2.40GHz)

If you like my plugins please consider to support the program buying a license; for personal contributions to me: paypal.me/dcprojects

#28 Posted: 2/16/2018 3:07:54 PM

Davide Carpi

1417 likes in 2873 posts.

Group: Moderator

Wrote
5 min 25 s (win 10 x64, intel i5 quad core 2.40GHz)

reduced to 30-35 seconds using ranged indices to replace stack/augment/submatrix (most heavy operations were in stack/augment under loops)

InterX.r1 - ranged indices.sm (87 KiB) downloaded 76 time(s).

If you like my plugins please consider to support the program buying a license; for personal contributions to me: paypal.me/dcprojects

1 users liked this post

Andrey Ivashov 2/16/2018 3:13:00 PM

#29 Posted: 2/16/2018 3:11:46 PM

Andrey Ivashov

2271 likes in 3734 posts.

Group: Super Administrator

18 seconds on my env.

Great! Thank you Davide!

#30 Posted: 2/16/2018 3:49:26 PM

Andrey Ivashov

2271 likes in 3734 posts.

Group: Super Administrator

BTW, I would suggest to use another version of plotG which is much faster then previous one (and simplier):

plotG(x,y,char,size,color):line(k:range(1,rows(x)),augment(x,y,el(r3,k):char,el(r4,k):size,el(r5,k):color),2,1)

plotG(xy,char,size,color):line(k:range(1,rows(xy)),augment(xy,el(r3,k):char,el(r4,k):size,el(r5,k):color),2,1)

Best regards.

1 users liked this post

Davide Carpi 2/16/2018 8:54:00 PM

#31 Posted: 2/16/2018 4:18:55 PM

Alvaro Diaz Falconi

992 likes in 1675 posts.

Group: User

Hi. Just one more thing, I guess: the only one remaining matlab function, mwfind replaced. (Orignal work have many) If I'm not wrong, it is now 100% SMath native. But double the amazing time obtained by Davide. But it's just a "scalar" implementation, sure there are better ones.

Best regards.

Alvaro.

PD: I don't replace plotG because don't understand very well that syntax.

InterX.r1 - ranged indices.sm (90 KiB) downloaded 60 time(s).

#32 Posted: 2/16/2018 4:45:48 PM

Andrey Ivashov

2271 likes in 3734 posts.

Group: Super Administrator

Wrote
I don't replace plotG because don't understand very well that syntax.

It is simple - ranged indexes used.

Just take a look at the speed:

Best regards.

#33 Posted: 2/16/2018 4:59:17 PM

Alvaro Diaz Falconi

992 likes in 1675 posts.

Group: User

Wrote

Wrote
I don't replace plotG because don't understand very well that syntax.

It is simple - ranged indexes used.

Thanks for the explanation, now I see the speed increment, and I don't think that it could be that big difference before you show it.

But actually the plotG(4) routine was which confuse me. I prefer

plotG(xy,char,size,color):if(cols(xy)>1,plotG(col(xy,1),col(xy,2),char,size,color),plotG(Re(xy),Im(xy),char,size,color))

because I think that any good plot implementation must to use complexes but not couple of reals. So, if xy argument have only one column, one can assume that data are as Z = X + iY.

Best regards.

Alvaro.

#34 Posted: 2/16/2018 5:08:15 PM

Davide Carpi

1417 likes in 2873 posts.

Group: Moderator

Quote
because I think that any good plot implementation must to use complexes but not couple of reals. So, if xy argument have only one column, one can assume that data are as Z = X + iY.

Andrey misses this detail of your function about the complex values, but you can still do it; keep the if statement and replace the loops (the case with more than a column as shown by Andrey doesn't require the col() functions, this is redundant and time consuming... If you think that more than 2 columns might be provided, add the case and a single sumbatrix instead

)

If you like my plugins please consider to support the program buying a license; for personal contributions to me: paypal.me/dcprojects

1 users liked this post

Alvaro Diaz Falconi 2/16/2018 5:45:00 PM

#35 Posted: 2/16/2018 6:24:21 PM

Arie

93 likes in 285 posts.

Group: User

Wrote
But double the amazing time obtained by Davide. But it's just a "scalar" implementation, sure there are better ones.

Here's my pure SMATH find function. with similar time to MWFIND.

InterX.r1 - ranged indices_alyles.sm (94 KiB) downloaded 79 time(s).

Feel free to join the SMath Studio Users Discord Channel: https://discord.gg/PayZpJW

4 users liked this post

Andrey Ivashov 2/16/2018 6:41:00 PM, sergio 2/17/2018 6:31:00 AM, Davide Carpi 2/16/2018 8:57:00 PM, Alvaro Diaz Falconi 2/16/2018 10:03:00 PM

#36 Posted: 2/17/2018 9:35:24 AM

Martin Kraska

1222 likes in 2150 posts.

Group: Moderator

Wrote
BTW, I would suggest to use another version of plotG which is much faster then previous one (and simplier):
plotG(x,y,char,size,color):line(k:range(1,rows(x)),augment(x,y,el(r3,k):char,el(r4,k):size,el(r5,k):color),2,1)
plotG(xy,char,size,color):line(k:range(1,rows(xy)),augment(xy,el(r3,k):char,el(r4,k):size,el(r5,k):color),2,1)
Best regards.

For such uses it would be good if the augment and stack functions could expand scalar arguments (or more generally any arguments not matching the maximum number of rows or cols) to uniformly populated vectors of the required size).

Then the above example wouldn't even need a helper function as you could write

augment(x,y,"*",10,"red"

Martin Kraska Pre-configured portable distribution of SMath Studio: https://en.smath.info/wiki/SMath%20with%20Plugins.ashx

2 users liked this post

Davide Carpi 2/17/2018 3:23:00 PM, Radovan Omorjan 2/17/2018 12:43:00 PM

#37 Posted: 2/17/2018 9:39:40 AM

Andrey Ivashov

2271 likes in 3734 posts.

Group: Super Administrator

Very good suggestion, Martin! I think I could do this.

Best regards.

4 users liked this post

Davide Carpi 2/17/2018 3:23:00 PM, frapuano 2/17/2018 10:20:00 AM, Radovan Omorjan 2/17/2018 12:43:00 PM, Martin Kraska 2/17/2018 9:41:00 AM

#38 Posted: 2/17/2018 1:20:11 PM

Radovan Omorjan

325 likes in 2052 posts.

Group: Moderator

Wrote
For such uses it would be good if the augment and stack functions could expand scalar arguments (or more generally any arguments not matching the maximum number of rows or cols) to uniformly populated vectors of the required size).

Then the above example wouldn't even need a helper function as you could write
augment(x,y,"*",10,"red"

I also vote for that augment() and stack () functionality.

In that case, we could even have "plot" functions (with the same functionality as augment())


plot(x,y)
plot(x,y,"*"
plot(x,y,"*",10)
plot(x,y,"*",10,"red"

This efforts are, of course, not justified if the micro Mathematics plot style would be introduced.

Regards,
Radovan

When Sisyphus climbed to the top of a hill, they said: "Wrong boulder!"

1 users liked this post

frapuano 2/19/2018 7:14:00 AM

#39 Posted: 2/18/2018 10:58:48 PM

Jean Giraud

983 likes in 6866 posts.

Group: User

Hello Marco,

I'm trying for the intersection of the two Trifolium, .
blue => native ... red => rotated.
They are created in "Page1 Trifolium.sm
Then recalled for the InterX(T1,T2)
Nothing happens >>> 1 hr ? [Win 7 Starter Edition ... 1.66 GHz]
Is it supposed to work ? Thanks in advance.

Page1 Trifolium.sm (32 KiB) downloaded 49 time(s).

Maths InterseX [Marco].sm (51 KiB) downloaded 51 time(s).

#40 Posted: 2/19/2018 12:26:16 AM

Alvaro Diaz Falconi

992 likes in 1675 posts.

Group: User

Hi Jean.

You are using an old version of the program. Try this one.

Notice that the point (0,0) is found several times by the algorithm. It could be interesting to check if it get the correct number of crosses at the origin.

Best regards.

Alvaro.

InterX.r1 - ranged indices_alyles.sm (88 KiB) downloaded 68 time(s).

SMath Forum

Line intersections - An example of translation of a function from Matlab to Smath - Messages