Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Search - "parallelism"
-
I just implemented parallelism in one of my major data collection tools for work, and cut the tool's run duration by 90%. That seems like as good a time as any to go to sleep.2
-
I just realized that CPU's are essentially electronic hookers, for processes.
They'll service any process that calls upon them, but keep switching between various processes so that it can spend time fairly with each of its processes. And of course, they'll immediately abandon their current process if one with a higher priority reveals itself and attend to that one instead. Once it is finished, it'll return to its less-important processes.
Kinda sounds like a hooker to me... 🤔4 -
This Book....
Before doing any systems programming you should definitely read this book... most people think they know what they are doing but in fact they are completely clueless and the worst part is you don’t realize how clueless you are... you don’t know what you don’t know nor do you know how much you really don’t know.. a most people are part of this group, including myself lol.
Computers are much more than a bunch of CPUs, buses and peripherals. (Embedded folks realize this). But this goes beyond embedded this is a systems book, on architecture of computers in general.
Learning only java and the java/C# python and the others SDK/Api and spending your life with horse blinders for what’s going on below only sets you up for failure in the future, and when you that point it’s gonna be a shocker. Could be tomorrow could be 20 years from now, but most people with those horseblinders get to that point and have that “experience” no avoiding the inevitable lol.
I really enjoyed this book in their quantitative approach to teaching the subject. Especially understanding parallelism and multi core systems.5 -
!rant
For all of youse that ever wanted to try out Common Lisp and do not know where to start (but are interested in getting some knowledge of Common Lisp) I recommend two things:
As an introductory tutorial:
https://lisperati.com/casting.html/
And as your dev environment:
https://portacle.github.io/
Notice that the dev environment in question is Emacs, regardless of how you might feel about it as a text editor, i can recommend just going through the portacle help that gives you some basic starting points regarding editing. Learn about splitting buffers, evaluating the code you are typing in order for it to appear in the Common Lisp REPL (this one comes with an environment known as SLIME which is very popular in the Lisp world) as well as saving and editing your files.
Portacle is self contained inside of one single directory, so if you by any chance already have an Emacs environment then do not worry, Portacle will not touch any of that. I will admit that as far as I am concerned, Emacs will probably be the biggest hurdle for most people not used to it.
Can I use VS Code? Yes, yes you can, but I am not familiar with setting up a VSCode dev environment for Emacs, or any other environment hat comes close to the live environment that emacs provides for this?
Why the fuck should I try Common Lisp or any Lisp for that matter? You do not have to, I happen to like it a lot and have built applications at work with a different dialect of Lisp known as Clojure which runs in the JVM, do I recommend it? Yeah I do, I love functional programming, Clojure is pretty pure on that (not haskell level imo though, but I am not using Haskell for anything other than academic purposes) and with clojure you get the entire repertoire of Java libraries at your disposal. Moving to Clojure was cake coming from Common Lisp.
Why Common Lisp then if you used Clojure in prod? Mostly historical reasons, I want to just let people know that ANSI Common Lisp has a lot of good things going for it, I selected Clojure since I already knew what I needed from the JVM, and parallelism and concurrency are baked into Clojure, which was a priority. While I could have done the same thing in Common Lisp, I wanted to turn in a deliverable as quickly as possible rather than building the entire thing by myself which would have taken longer (had one week)
Am I getting something out of learning Common Lisp? Depends on you, I am not bringing about the whole "it opens your mind" deal with Lisp dialects as most other people do inside of the community, although I did experience new perspectives as to what programming and a programming language could do, and had fun doing it, maybe you will as well.
Does Lisp stands for Lots of Irritating Superfluous Parentheses or Los in stupid parentheses? Yes, also for Lost of Insidious Silly Parentheses and Lisp is Perfect, use paredit (comes with Portacle) also, Lisp stands for Lisp Is Perfect. None of that List Processing bs, any other definition will do.
Are there any other books? Yes, the famous online text Practical Common Lisp can be easily read online for free, I would recommend the Lisperati tutorial first to get a feel for it since PCL demands more tedious study. There is also Common Lisp a gentle introduction. If you want to go the Clojure route try Clojure for the brave and true.
What about Scheme and the Structure and Interpretation of Computer Programs? Too academic for my taste, and if in Common Lisp you have to do a lot of things on your own, Scheme is a whole other beast. Simple and beautiful really, but I go for practical in terms of Lisp, thus I prefer Common Lisp.
how did you start with Lisp?
I was stupid and thought I should start with it after a failed attempt at learning C++, then Java, and then Javascript when I started programming years ago. I was overwhelmed, but I continued. Then I moved to other things. But always kept Common Lisp close to heart. I am also heavy into A.I, Lisp has a history there and it is used in a lot of new and sort of unknown projects dealing with Knowledge Reasoning and representation. It is also Alien tech that contains many things that just seem super interesting to me such as treating code as data and data as code (back-quoting, macros etc)
I need some inspiration man......show me something? Sure, look for a game called Kandria in youtube, the creator, Shimera (Nicolas Hafner) is an absolute genius in the world of Lisp and a true inspiration. He coded the game in Common Lisp, he is also the person behind portacle. If that were not enough, he might very well also be Shirakumo, another prominent member of the Common Lisp Community.
Ok, you got me, what is the first thing in common lisp that I should try after I install the portacle environment? go to the repl and evaluate this:
(+ 0.1 0.2)
Watch in awe at what you get.
In the truest and original sense of the phrase (MIT based) "happy hacking!"9 -
I started to learn programming to be liked by a girl and since then I periodically ask myself if this is really for me... And periodically, right as I start thinking it may not be for me, I find myself solving issues programming stuff, seeing parallelism between software development and anything and randomly toss out the classical: "you know, with and Arduino you could(...)" To people put of nowhere.
So yeah. Guess it's for me. I hate it but... Wait, I could automate my windows depending on the difference between internal and external temperature... Hmm... Later guys, got a project to deal with!14 -
When file managers copy and delete files within the same partition instead of moving or renaming them…
When Google's Storage Access Framework was introduced, it did not feature a move command, so file managers just resorted to copying and deleting files within the same storage. Not only does this cause needless wear and is much slower, but it also destroys the date/time attribute (it gets changed to current).
When moving files through MTP (miserable transfer protocol, used for connecting smartphones to PC), they are also copy-deleted. This makes moving a 20-Gigabyte DCIM folder impractical. Also, if one cancels the operation, it might end up whoopsie-daisy deleting some files from the source before they have been transferred.
MTP is so bogus that it is incapable of a simple operation that would JustWork™ on mass storage devices. Not to mention, MTP lacks parallelism and its directory listing loading it S-L-O-W. Upwards of a minute for just 1000 files. Sometimes, it fails loading at all.
Also, trying to rename a file through MTP using the terminal through GVFS, even if just within the same folder, it copy-deletes it. If I want to rename a 1 GB 2160p 4K video in a highly populated DCIM folder, I can not do so through the terminal. At least, the 4K video has a time stamp in its internal metadata, but it still renames slowly and adds needless wear to the smartphone's flash memory.14 -
how do you go about working with a teammate that the best thing would be to rewrite the code he worked with for the last 5 months!?
I mean, no separation of concerns, layers upon layers of unnecessary abstractions, unneeded parallelism and mutex and whatnot...
Like, 5000+ lines of code that could be done in 400...3 -
Golang tutorials trying to explain the difference between concurrency vs parallelism be like "Oranges are different from oranges. Oranges are juicy while oranges are juicy. They are not to be confused, u newb"1
-
[CONCEITED RANT]
I'm frustrated than I'm better tha 99% programmers I ever worked with.
Yes, it might sound so conceited.
I Work mainly with C#/.NET Ecosystem as fullstack dev (so also sql, backend, frontend etc), but I'm also forced to use that abhorrent horror that is js and angular.
I write readable code, I write easy code that works and rarely, RARELY causes any problem, The only fancy stuff I do is using new language features that come up with new C# versions, that in latest version were mostly syntactic sugar to make code shorter/more readable/easier.
People I have ever worked with (lot of) mostly try to overdo, overengineer, overcomplicate code, subdivide into methods when not needed fragmenting code and putting tons of variables.
People only needed me to explain my code when the codebase was huge (200K+ lines mostly written by me) of big so they don't have to spend hours to understand what's going on, or, if the customer requested a new technology to explain such new technology so they don't have to study it (which is perfectly understandable). (for example it happened that I was forced to use Devexpress package because they wanted to port a huge application from .NET 4.5 to .NET 8 and rewriting the whole devexpress logic had a HUGE impact on costs so I explained thoroughly and supported during developement because they didn't knew devexpress).
I don't write genius code or clevel tricks and patterns. My code works, doesn't create memory leaks or slowness and mostly works when doing unit tests at first run. Of course I also put bugs and everything, but that's part of the process.
THe point is that other people makes unreadable code, and when they pass code around you hear rising chaos, people cursing "WTF this even means, why he put that here, what the heck this is even supposed to do", you got the drill. And this happens when I read everyone code too.
But it doesn't happens the opposite. My code is often readable because I do code triple backflips only on personal projects because I don't have to explain anyone and I can learn new things and new coding styles.
Instead, people want to impress at work, and this results in unintelligible, chaotic code, full of bugs and that people can't read. They want to mix in the coolest technologies because they feel their virtual penis growing to showoff that they are latest bleeding edge technology experts and all.
They want to experiment on business code at the expense of all the other poor devils who will have to manage it.
Heck, I even worked with a few Microsoft MVPs.
Those are deadly. They're superfast code throughput people that combine lot of stuff.
THen they leave at you the problems once they leave.
This MVP guy on a big project for paperworks digital acquisiton for a big company did this huge project I got called to work in, which consited in a backend and a frontend web portal, and pushed at all costs to put in the middle another CDN web project and another Identity Server project to both do Caching with the cdn "to make it faster" and identity server for SSO (Single sign on).
We had to deal with gruesome work to deal with browser poor caching management and when he left, the SSO server started to loop after authentication at random intervals and I had to solve that stuff he put in with days of debugging that nasty stuff he did.
People definitely can't code, except me.
They have this "first of the class syndrome" which goes to the extent that their skill allows them to and try to do code backflips when they can't even do code pushups, to put them in a physical exercise parallelism.
And most people is like this. They will deny and won't admit, they believe they're good at it, but in reality they aren't.
There is some genius out there that does revoluitionary code and maybe needs to do horrible code to do amazing stuff, and that's ok. And there is also few people like me, with which you can work and produce great stuff.
I found one colleague like this and we had a $800.000 (yes, 800k) project in .NET Technology, which consisted in the renewal of 56 webservices and 3 web portals and 2 Winforms applications for our country main railway transport system. We worked in 2 on it, with a PM from the railway company.
It was estimated 14 months of work and we took 11 and all was working wonders. We had ton of fun doing it because also their PM was a cool guy and we did an awesome project and codebase was a jewel. The difficult thing you couldn't grasp if you read the code is if you don't know how railway systems work and that's the only difficult thing.
Sight, there people is macking me sick of this job11 -
A five-minute overview of the differences between concurrency and parallelism.
https://monades.roperzh.com/concurr...