What does /g/ think of MATLAB? Has anyone ever used it to program

>>52619518
it was pretty gud in the 90's

>>52619518
I had to learn it in uni. Horrible language, but it has huge number of modules/libraries. I have no experience in programming Arduinos but I would start with C or maybe C++. There is no reason to tie yourself to a fucking expensive language.

>>52619518
MATLAB is fine. I personally used it in school, but I wouldn't recommend using it for Arduino because you might as well use C. My guess is that it would be easier (after an initial learning curve), more efficient, and better documented than MATLAB's plug-in.

MATLAB is fucking stupid for anything that isn't 100% matrix multiplication.

A programming language made for people who don't program. Strongly dislike it, but sadly everyone uses it for something. I'm glad I somehow always manage to avoid it.

Don't waste your time on it... I haven't used matlab extensively but every moment I have used it I hated it. Using it for arduino is way overkill and probably way more complicated than just using C. Save yourself the trouble and just use C.

Using MATLAB for Arduino is beyond retarded. Still, MATLAB is a pretty nice software, but it's time for Julia to take its place.

>>52619834
This.

>>52619518
Horrible language, huge waste of money, designed for engineers who really can't be bothered to learn proper programming techniques.
You can do everything Matlab does faster in Python or R.

>>52619696
Octave has excellent compatibility so not exactly expensive

ITT: CS majors that are mad they got a D in MATLAB class and it doesn't work like Java.

EE here, Matlab is an indispensable tool for analysis and design
That being said programming in it is dumb

>>52619518
It's shit, but useful.

>>52620005
Are you talking about matlab or simulink?

>>52619991
>You can do everything Matlab does faster in Python or R.
Bollocks. The only thing faster than MATLAB (on a single computer) is C/C++, but ain't nobody wasting time when you can get a comparable performance with just a fraction of the effort. Python and R are at least twice as slow and offer none of the benefits, except price, which hardly anyone who needs MATLAB cares about. The only other viable alternative is Julia, but it's still missing many feature and libraries.

>>52620044
Both are very useful.

>>52620049
>Bollocks. The only thing faster than MATLAB (on a single computer) is C/C++

I think he intended fast to develop, not to run.
That said MATLAB is very useful.

>>52620099
When comparing development, nothing comes even close to MATLAB. Nothing.

>>52620049
>I'm using precompiled c functions in matlab but not in python
both are shit slow outside of their special functions

OP here. I ask this because I have to use it for class anyways so I figured I could get a leg up and avoid the curve to use it for arduino. Iowa State gives it to engineers for free so

>>52620110
Yes, both can (MATLAB does it by default) use the standard math libraries -- ATLAS, BLAS, etc. The thing is, MATLAB also has some of the best algorithms behind their implementations. Often literally the best discovered ones.

>>52620049
>Python and R are at least twice as slow and offer none of the benefits
Numpy/Scipy use many of the same numerical and scientific programming libraries found in C and Fortran (Lapack/blas for example).

The majority of things you do in MATLAB will not be slower in python.

>>52620049
The problem with Julia is less about missing features and more about backwards compatibility being broken with each update. Makes it too annoying to write any large projects in it right now. Missing libraries aren't really an issue because the FFI is surprisingly not shit.

>>52620049
>The only thing faster than MATLAB (on a single computer) is C/C++

Is this true?

>>52620123
Because it is everywhere, it can't hurt to have some skill. Also, some really useful python libs like matplotlib are heavily matlab-influenced. So go for it. Personally, I could never handle matlabs bullshit
>column and row vectors are completely different things and you'll always have the wrong one for everything you're trying to do herp di derp

>>52620165
They use the same libraries but in a much more inefficient way. I've done comparisons for projects before, you can do them yourself (both Python and MATLAB offer easy to use timing facilities).

>>52620171
Depends. Their optimisation library is really poor. Last I checked, no proper convex solver, everything is slow and formulation is just painful.

>>52620191
Yes.

>>52620123
Nah just use C, but you could interface it to MATLAB with Serial.print
Take some analog data and print it over the UART, then read it in in MATLAB and filter it, fourier transform it or plot it.

>>filterbuilder
>>f1=fft(data);
>>f1=abs(f1);
>>plot(f1)

>>52620191
>Is this true?

Yeh, but writing fast C++ code is probably more difficult. Anybody with experience in both MATLAB and C++ can confirm?

>>52620049
Have you ever heard of numpy and scipy?

Ipython+Numpy+Scipy+Matplotlib = Matlab/Octave with a less shitty language.

>>52620191
It is not, and I doubt he has tried them all to verify that statement.
For example, Java with modern just in time compilation is way faster.

File: benchmark.png (41KB, 940x376px) Image search: [Google]

41KB, 940x376px

>>52620191

>>52620279
My colleagues had some data they were trying to process and plot with MATLAB it was too slow, they used open GL to get it to work in real time.

>>52620298
fucking this, matfags

>>52620315
>The city of Octave

>>52620317
>My colleagues had some data they were trying to process and plot with MATLAB it was too slow, they used open GL to get it to work in real time.

Sounds like the problem is in the plot part, not in the number crunching part.

>>52620303
>>52620315
>tfw when you think you're trolling real hard and then someone proves you actually right.
Also, lol GNU Octave

OP just program your arduino in whatever you're comfterable with
Matlab is fine, it has a ton of well organized documentation and someone probably coded and published everything you need already so if you arent looking to become a matlab grandmaster you can save alot of time by reusing code
However if you're looking to pick up some programming skills figure out what language is relevant for aerospace engineering, Im guessing its going to be either python or lisp/scheme (if it isnt Matlab) so try either of those

>>52620350
Possibly, as I said I wasn't involved. Fiber optic shit

How the fuck do you even program an AVR processor with a matrix processing system?

Holy fuck, is C to mainstream for you?

>>52620315
Javascript
mandel
0.66

Javascript faster than Fortran. What a time to be alive.

>>52620450
nodejs haters BTFO

>>52620450
Language benchmarks, especially of vastly different languages are, and always will be, completely useless.

>>52620473
>Language benchmarks, especially of vastly different languages are, and always will be, completely useless.

You must take benchmark with a grain of salt. But they stiil give you a ballpark figure.

>>52620432
clearly you have never used it.

>>52619518
I actually used it for a simple interface for a killswitch with some coworkers compiled (matlab) programs. I have the arduino board with relays on it to kill the power source so the cache/packets would refresh. So less drops happens and data can start back up and continue collecting if no1 is around to help it without the arduino connected.

>>52620512
The author can be a faggot and use fast code in one and slow in other language.
Or he can be just incompetent.
Or he uses different environments.

What is it with /g/'s hardon for C?

>>52620565
No one has come up with an alternative in 30 years.

>>52620565
Probably something about it being one of the fastest, longest used, most useful languages?

>>52620513
I haven't used what?
I've used Matlab on a course I took to get ETCS credits and passed that with an A.
I've programmed an AVR recently in C, compiling with GCC and flashing with avrdude.

But yeah, I've never thought to use an expensive software with half-assed FORTRAN-inspired language designed for processing large data sets to program a fucking MCU.

>>52620565
It's pretty much the fucking standard for programming micro-controllers.
It's also useful for other stuff. The amount of software written in C is humongous.

>>52620565
Geek machismo.
Muh pointers.
Muh bound checking is for fags.

>>52620593
>>52620594
>>52620646
What is it with /g/'s hateboner for C++?

>>52620660
http://harmful.cat-v.org/software/c++/linus

Here's a taste:
>C++ is a horrible language. It's made more horrible by the fact that a lot
of substandard programmers use it, to the point where it's much much
easier to generate total and utter crap with it. Quite frankly, even if
the choice of C were to do *nothing* but keep the C++ programmers out,
that in itself would be a huge reason to use C.

>In other words: the choice of C is the only sane choice. I know Miles
Bader jokingly said "to piss you off", but it's actually true. I've come
to the conclusion that any programmer that would prefer the project to be
in C++ over C is likely a programmer that I really *would* prefer to piss
off, so that he doesn't come and screw up any project I'm involved with.

>>52620660
Nah I use C++ when I'm not doing embedded.

>>52620660
http://yosefk.com/c++fqa/

Hey related question what is a good language for writing out physics and math formulae? I hate using a pen and paper

At one point I heard matlab lets you do that

File: 1428187374014.jpg (59KB, 453x439px) Image search: [Google]

59KB, 453x439px

>>52620746
>I hate using a pen and paper

>>52620746
Latex

R is better nigga.
It is free as in price and as in freedom. Also a lot of job opportunities since enterprises don't like to pay for things.

>>52620660
Prefer C Plus equality.

>>52619518
Proprietary shit that does nothing that couldn't be done as a library for another language.

>>52620746
Latex. Made by a mathematician for mathematicians.

>>52620924
>Proprietary shit
Opinion dropped

>>52620946
Enjoy paying 2K for the "privilege" to use a shitty language to do the same shit you can do faster and for free with python or R.

Bloated, expensive, piece of Java-based shit. It's mildly convenient for any application which is pretty much 100% linear algebra, but the price, hassle of getting it set up, and overall shittiness makes it not worth it.

>>52620216

You must be a complete fucking retard if you do not understand the row and column vector notations. They follow standard mathematical practice for dealing with matrices.

>>52621051
>Enjoy paying 2K for the "privilege" to use a shitty language to do the same shit you can do faster and for free with python or R.

Octave is free. Besides, you don't know how the world works outside of your basement.

>>52620298

They are still not as fast as Matlab.

>>52621130
I'm pretty sure Mathworks couldn't invent something faster than Nvidia and Intel did.

While I don't use it I saw it being used for various projects, it seems to scale quite well over clusters.

If price is the issue why not Octave?

>>52621163
Price is not the issue, the shitty language is.

>>52621115
Please tell me where I can get an interface to Tensorflow in Matlab.

If you need something easy math scripting, R is better.
If you use the graphical toolboxes, matlab is pretty much the only option.
If you design a controller, or something like that it is very beneficial to just write the functions and draw the design. rather than write it as a function (unless you want to do that)

File: rootlocus.jpg (28KB, 794x428px) Image search: [Google]

28KB, 794x428px

>>52620924
>>52621051
>>52621057
>>52621195
Find me something else that helps you design via root locus

>>52620315
why is pi_sum = 1 so often?

>>52621195
You write one freetard. Isn't this what your idol preaches?

>>52621253
>find me someone capable of drawing simple plots
Oh man. That's a toughie.

>>52621253
Is it profitable to write toolboxes for matlab for retarded engineers, who can't program?

>>52621297
I'm a retarded engineer who can't program, but I try.

>>52621280
>>52621297
>why buy something to do a job when you can waste time writing a theoretical piece of software for free
programmer logic.
Tell me, do you know Laplace transforms? Feedback control?

>>52621340
programming is an end not a means, don't you understand?

>>52621340
Rings some bell. Everything's easy if you have some time, some books and can read.
Don't get so defensive. I'll rephrase:
Is it profitable to make matlab toolboxes for people, who for some reason (time constraints, knowledge, skills) can't create one?

>>52620746
Latex.
It's a typesetting language though, not something to use when fiddling with equations.

Good for engineering I hear. Not really for everyday applications or anything.

>>52621390
What? Programming is the means and engineering is the end.
>>52621397
>Everything's easy if you have some time
>time constraints
you just answered your own question, welcome to industry

>>52621455
>What? Programming is the means and engineering is the end.

I guess irony doesn't come across very well in ASCII.

MATLAB is a fine choice for programming language for the sorts of projects an engineering undergrad would do, like interfacing some sensors and a step motor thru arduino with a PC or something, it is easy to learn and low-effort for the right tasks.

If you have some proficiency with C++ that's generally preferable, so if you have time to learn that first (you'll have to as an engineering student) it might still be your best option overall.

>>52620279
just use blaze (C++ library) for linear algebra. It blows MATLAB out of water and not hard to use at all.