วีดีโอ

Sure, you create the prototype in Python, then rewrite it in C if needed. Porting Python code is rather easy once the initial design is done. CPython extensions can become messy though. Even if you get the code to work, it's hard to eliminate all memory leaks during the interaction with Python because you're relying on Python's garbage collector to free the memory. Not even popular libraries like TensorFlow seem to get that right.

If you concern about Python speed just simply use Numpy , Pandas like libraries. 😂😂😂

Awesome!!!!

what language would you say comes closest to c and c++ compile/run speeds, while still being secure, memory-wise? ive heard that rust it the togo for this, but idk much about rust. wanted to go with c++ and just get good with memory, but at the same time meh cuz White House rly wants us to go with safer languages

Production speed is where python is much faster than C++ and many other languages. This extra speed you're talking about is largely not needed MOST of the time. Your bottlenecks are almost never CPU these days. Being able to create a functional MVP in 1/10th the time is a huge deal. It's something college freshmen don't understand because they have no real world experience.

So basically 11 times worse code, but can be written by lazier or stupider developer.

Python is good for prototyping. It should stay as prototyping. Note: I like Python. I still feel it's misused.

Simple syntax, no memory management, you don’t have to do anything

If you had more practice and developed in C++ on a regular basis you could've written it significantly faster. In fact, I'd argue that C++ is one of the fastest to write languages for common things. You can simply give everything an auto type and it's just as loosey goosey as Python.

Development speed? If you want development speed use to-basic. I can get a simpler program out in a few minutes. Once you learn it you will have virtually no bugs

answer: yes

I code in C++ and never used cmake or json files.

As a C student in Uni I now learn Java and it hurts so Python is not for me

Subset of Python, not normal Python…

The time you’re measuring around the 3:10 mark is the time it takes to declare the function, not the first run time or the jit.

This is just Julia now.

i started playing with code using c++, i never had any formal education, just articles, forums and youtube. then one day i decided to make a game in python. i never once felt like python was any easier or faster to write, maybe it was because i was learning along the way but i feel like it would have taken me just as long to learn how to make it in C++

Classes are not slower than straight C code. Virtual method calls are slow, but if you don't do that, there shouldn't be a difference. Internally, classes are structs and methods are functions taking an implicit "this" pointer as a parameter. A non virtual method call is the same as calling a function in C and it is neither slower nor faster. A virtual call is like calling a function through a function pointer, which adds a level of indirection and the slowdown can be significant in inner loops.

i prefer lua :) i know nothing about coding, im just a basic linux user and out of the box i can use lua, other languages i can see how its written but it still somehow always gives me errors, so thats where i know whats good for me.

With Python you can get more people, fire them and get new ones - whenever you need, because of low entry level and high code production speed. With C you have to valuate your employees - which is always hurt the "business".

As a python developer, i would add that anyone who develop in python for exec speed is a fool. But when developing a product, development speed is king : developing a product in 2 weeks -> ship it -> get customer complain that it's too slow -> rebuild it is far superior than : developing a product in 8 weeks -> ship it -> get 0 customer complains (but lost 6 weeks of revenue)

Now try GO

12 hours to code a tictactoe script is crazy. Calls himself "the code guy"

I think the biggest difference is Python is so much easier to learn than c++. Companies would rather hire someone with 1 year of experience in Python over someone with 1 year of c++ experience.

If you know you can use JIT and packages that are in C anyways in the background. You can completely bypass the performance penalty. I'm surprised that you didn't do a performance run with pipy and cython since you specifically mentioned them. Instead of just keeping the 10x slower comparison from the start... You wouldn't even need to edit your code with them.

Don't forget the conda extension "numba" -- a just in time compiler that needs no changes or any other modifications to your python code. It is very fast. Also remember that the numpy library is very fast.

I've heard that if you use Python on large data sets, using a decent JIT compiler, it can actually perform pretty well. Maybe not as fast as equivalent C++ code, but it can close the gap a bit. I think the reason is it can keep optimising the machine code because the same code is running so many times. I'm primarily a C++ developer, so I don't really know how true that is though.

Crash? Why would I use a variable to represent a number, then change it to represent a string? Who does this?

With that comparison, now remove console output each game

Can you please try the various ways to compile Python and show how fast is each? Would any of them be similar to C/C++ performance?

You don't "iterate over array" with indexes in C++. That's C.

Please also do with c

I program just for fun and no one is further away from a professional than me. Nevertheless, a programming language that doesn't need a variable declaration scares me and is a little weird

Huh? Wouldn't the c++ be less than 200 lines? Most of them being { and }

one thing is the printing which obviously is a bottleneck, but did you compile c++ with optimized build O2 or O3 flag? I expect 10000 games of tic tac toe to finish way faster than in your example

If whatever you're producing will do the job in 100 ms when its coded in python instead of 10 ms if it were to be coded in C++ it doesn't matter at all to the user who doesn't know this fact. Btw the reaction time of the best people around is around 100 ms and realistically it can even take multiple seconds before anyone who isn't an annoying nerd like we are will complain

Write your code in python then ask chatgpt to convert it to c++ 😅

Rizz

The great thing about python is that you can use it for the high-level logic, and then move the performance bottlenecks to a lower-level language.

the number ive always heard was 140 times slower

People complaining about printing to terminal slowing down the program have probably never build any real applications. Yes, it's a bottleneck. Real applications tend to have those. IO is a very common bottleneck - not just printing to the terminal, but also things like calls to databases or APIs can be very significant bottlenecks that have nothing to do with the language you are using. If you factor those in, you often find that switching to a language that is a bazillion times faster doesn't actually make your application faster - you are just waiting faster. That's why Python is so widely used, despite being relatively slow - because its lack of performance doesn't actually matter so much for many applications.

Hmm.

i think python was slower mainly because it somehow took you over 200 lines of code to make tic tac toe

Can you redo the measurements this time without printing to the console because that is slowing code so much. In my case printing in C can slow down code up to 200 times in some scenarios.

bruh, don't compare a compiled executable to a interpreted script... atleast compile the python code to an .exe before comparing, cause that'd be unfair towards python

About the number of lines: Well, in the c++ code many lines contains only a single "{" and the closing "}" can be placed by many editors automatically. So even if the number of lines is doubled in c++, it doesn't mean that developing is also 2 times slower.

Good video. When i saw title, i was thinking like. Oh this is another video of why you should use C++, because it's much faster than python. But hopefully the vid is not about that. For me I use learn and use python because It has many helpful libraries with ready-made solutions, and also bcz it's syntax is very easy, and you can make programs way more faster. In fact many people who speaks about comparing python and C/C++ speed, everyone for some reason forgetting about libraries like ctypes, numpy, etc. This libraries can make python runs at speed that is comparable with C++. So why we should use C/C++ just bcz the speed. It should be used when you need direct access to system components, not only bcz of the speed. Sorry if I offended you.

Could you please make a comparison between javascript and c++ ?

Python is a disgrace. It should not exist. If the community had focused on C++ only instead of creating a bunch of other programming languages, we wouldn't be arguing about speed or ease of development.