did any of those 'lets make python fast' projects pan out?
>>102468731Sounds like what average python developer would do
>>102468731Unless GIL is removed Python will never be fast.
That would require a level of competency that no one in the Python community has and no one outside of it wants wasted.
how come thirdies get attacked by all sorts of scary animals in cities, while firsties are fine>inb4 black people and arabsboo get better material>inb4 australiauh
>>102468731how can a big ass mofo like this just randomly visit and greet you in your kitchen in the first place? or was it her pet?
>>102468731I don't get it.Can you not just bite back the snake back?If it tries to eat you just try to eat it back.
>>102469148firsties don't live in tropical jungle hellholes, at least not by choice.
>>102468731why didnt she just stab it with a knife
>>102468731Nope. The jury is still out on Mojo tho
>>102469148>>inb4 australiaThat's basically a democratic dictatorship by now, so just another turd world country.
Newer versions of python are much faster and I think there's an official GIL-free version still in dev channels. Python will never compete with shit like C++ for perf. but most use cases doesn't matter.
>>102468731‘Fast Python’ is pointless. Python is plenty fast if your code isn’t retarded. If you’re doing something that’s slow in Python even with perfect code, then you shouldn’t be using Python in the first place.Python is not meant for all purposes, mental midget pythonistas need to learn this lesson.
>>102471980even if it was 1/10 C++'s speed it would be incredibly usefulalso you can do python multiprocessing for parallel stuff where each has it's own GIL
>>102469154It's Thailand. Unless you're in downtown Bangkok, one of those is within a few meters of you.
>>102469148there's tons of bear and cougar attacks in the us
>>102469154>he doesnt have to fight off a pack of red faced monkeys every week to make sure the vegetable garden isn't absolutely rektYou firsties have it easy
>>102468731they all went for the JIT compiler crap that was popular because javascript had it (because it needs it) instead of properly compiling shit and moving STD code into C like PHP does it.
>>102468731python is basically a wrapper for numpy/scipy/sympy/tensorflow at this point, for anyone who uses it for anything that needs performance, and those are written in gigaoptimized platform-specific C and GPUese. There simply isn't enough demand to make neo-Perl faster.
>>102469218>snakes won't bite you if you leave them alone
>>102472928Yeah but they eat campers, they don't attack people living in cities.
>>102469148>>inb4 australiaI though you were talking about first world counties, what does this have to do with the topic
>>102469218FUGG:DD:DD:DDDDDDD
>>102469218what kinda ninja snake is that?does it bite the guy at the door or how does it latch onto him?
I hate python so much it's unreal. I have probably written 100k lines of python in the past two years. it's not that the language and ecosystem is terrible. It's that there's too many clueless coders who use it
>>102469218Just stomp on the fucking->no foot wear at all>not even sandalswell shit, they fucked themselves over then.
>>102474395https://www.mirror.co.uk/news/world-news/puma-video-watch-moment-puma-3093075wrong again
>>102473700Snake got pissed when it got the door closed on it.It just wanted some company or eggs.Eggs do taste pretty good.#notallsnakes #snakelivesmatter #gimmethoseeggs
>>102468731The polars library uses Python as an interface for Rust for data processing tasks (much faster than pandas which uses C). You can also use pyspark if you want to use Python as the interface to a JVM coordinating cluster computing in Spark. Some libraries do support multithreading/multiprocessing and therefore can go pretty quick. So, yes, but only because Python is using other technologies rather than being fast itself. Since Python can interface with all these different technologies, I think it's clearly the best option to learn in 2024 and specialize into your given industry using its purpose-built libraries. I also really like Jupyter notebooks because it's a similar story of one interface for many solutions (scripting, visualization, reporting, documentation, etc with few downsides).
>>102468731Python's problem is not really the implementation, it's just a badly designed language. It's basically impossible to make fast without cutting off a lot of features and breaking compatibility
>>102476419Part of that is that there is a thick wad of complexity between the surface that most programmers see and the way things work at the bottom. Which wouldn't matter except many common libraries use that shit to do their clever tricks, so if you try to optimize it out and get some real speed, you end up breaking lots of people's code. Yet if you don't, you can't make a big difference to the speed at all.Also, the base of the implementation has some shit ideas (like how the GIL works, the threading and memory management model in general, that sort of thing) that hurt things a lot, but the real problem is a bit higher up.
>>102468731I always thought you can simply somehow translate all this into JS and have node/v8 fix the performance for you with Google's auto-magic. But I've never seen it done. If you do it that way, you can probably even replicate the native interfaces.
>>102474516>does it bite the guy at the door or how does it latch onto him?It uses one of those claws from Dune. It is the worm that rides now
>>102476396A snake's tail posted this.
>>102476419A good interpreted Python implementation could easily be about 10x faster with the same language features. The issue is C extension compatibility. The C interface exposes the structure of Python's VM, so the Cpython developers are now locked into working around that old slow design for backwards compatibility.
To fire up a python interpreter and print "Hello, World!" translates to 45 million CPU instructions.You can print "Hello, World!" with 40 cpu instructions.Start there.
