Check out my new community: !tech_memes@lemmy.world
Why would an RTX 4090 make Python faster?
If you create a meme like this at least inform yourself on how computers and software work.
If 4090 is not making stuff faster then why my games run faster with it? /s
What makes you think python is in optimized and bloated?
Did you know vast majority of AI development happening right now is on python? The thing that literally consumes billions of dollars of even-beefier-than-4090 GPUs like A100. Don’t you think if they could do this more efficiently and better on C or assembly, they would do it? They would save billions.
Reality is that it makes no benefit to move away from python to lower level languages. There is no poor optimization you seek. In fact if they were to try this in lower level languages, they’ll take even longer to optimize and yield worse results.
TBF, using AI as an example isn’t the best choice when it consumes an ungodly amount of power.
They do do it in C. The packages are written in C, python is just used as the wrapper to allow less coding skilled data scientists to easily use it.
That’s like the entire data science joke. It’s C in a python trench coat.
Every single high performance python library is written in C, you just don’t see it
Don’t worry this post was written by a first year computer science student who just learned about C. No need to look too closely at it.
I’m happy if it’s actually running in python and not a javascript app with electron.
Idk, it’s rare for an electron app to literally not even run. Meanwhile I’m yet to encounter a python app that doesn’t require me to Google what specific environment the developer had and recreate it.
Phyton
It used to be pretty terrible, but the frameworks are getting there, starting with the languages they are based on.
Believe it or not, Java has been optimized a ton and can be written to be very efficient these days. Another great example of a high-level, high-efficiency language is Julia. And then there is Rust of course, which basically only sacrifices memory-efficiency for C-speeds with Python-esque comfort. It’s getting better.
Ah yes, those precious precious CPU cycles. Why spend one hour writing a python program that runs for five minutes, if you could spend three days writing it in C++ but it would finish in five seconds. Way more efficient!
exactly! i prefer python or ruby or even java MUCH more than assembly and maybe C
I mean, I’d say it depends on what you do. When I see grad students writing numeric simulations in python I do think that it would be more efficient to learn a language that is better suited for that. And I know I’ll be triggering many people now, but there is a reason why C and Fortran are still here.
But if it is for something small, yeah of course, use whatever you like. I do most of my stuff in R and R is a lot of things, but not fast.
Because when it is to actually get paid work done, all the bloat adds up and that 3 days upfront could shave weeks/months of your yearly tasks. XKCD has a topic abut how much time you can spend on a problem before effort outweighs productivity gains. If the tasks are daily or hourly you can actually spend a lot of time automating for payback
And note this is one instance of task, imagine a team of people all using your code to do the task, and you get a quicker ROI or you can multiply dev time by people
That also goes to show why to not waste 3 days to shave 2 seconds off a program that gets run once a week.
You can write perfectly well structured and maintainable code in Python and still be more productive than in other languages.
This site has good benchmarking of unoptimized and optimized code for several languages. C+ blows Python away. https://benchmarksgame-team.pages.debian.net/benchmarksgame/index.html
SDLC can be made to be inefficient to maximize billable hours, but that doesn’t mean the software is inherently badly architected. It could just have a lot of unnecessary boilerplate that you could optimize out, but it’s soooooo hard to get tech debt prioritized on the road map.
Killing you own velocity can be done intelligently, it’s just that most teams aren’t killing their own velocity because they’re competent, they’re doing it because they’re incompetent.
And note this is one instance of task, imagine a team of people all using your code to do the task, and you get a quicker ROI or you can multiply dev time by people
In practice, is only quicker ROI if your maintenance plan is nonexistent.
