>Elite - 22k of memory
Notepad - 1400k of memory
Explain yourself, modern computing.
Lazy fucks that don't know how to code.
programs are written in more abstract languages with more helper libraries and frameworks. We have "plenty of ram" so devs don't try to optimize
Unused ram is wasted ram
>>53821828
/thread
>>53821828
lack of optimization = wasted resources
>>53821828
So a text editor that uses 7GB of RAM, if 7GB is there, is acceptable to you?
>>53821466
Everyone knows software quality is inversely proportional to hardware capability
Reminds me of that Sinclair interview where the interviewer said something like "you must be really impressed with what modern computers can do," and Sir Clive just goes "No. Because when you look at what we did with a few K of memory we really haven't got very far."
This is why I'm actually getting excited for all the low-power boards. Their constraints could actually give way to some really interesting pieces of software, if people were actually interested in such a thing rather than just bigger and "better."
>>53821882
If it improves the performance
>>53822001
Surely something with a lower footprint will perform faster as it's simpler and has to communicate with the hardware and load less.
>>53821828
>java developers
I hear a lot of this shit about how many things we used to do back in the day with the shittiest hardware, but tell me, what would you expect with current hardware if you were all like "let's not waste those 20-30kb of RAM"?
When I take my current PC which is still sort of a piece of shit for modern gaming, I can't think of a lot of things that could be done with better hardware. I can't think of something that could take all of my RAM and be so fascinating.
>>53822377
That's just not true. There's plenty memory performance trade offs in computer science.
>>53822487
Anon, that might have been true in days gone by, but these days memory is by far the slowest operation that a cpu can do, unless its in the cache already, which is only facilitated by having a smaller footprint, and just to be clear I'm not saying have all your program in the cache all the time, as this would be completely infeasible for larger programs, but simply attempt to design the program such that it is constrained to a smaller memory area over areas of tighter repitition, so that as your program runs, it gets to an area and is able to effectively perform the load once and then operate on the cache exclusively for the duration of the task/loop, then write that back to main memory upon completion before advancing to another section and loading in that area, doing the thing it has do, putting it back, etc.
At least, thats my limited understanding of it
>>53822431
>what would you expect with current hardware if you were all like "let's not waste those 20-30kb of RAM"?
On a user end, having much cheaper hardware that uses less electricity. This would be attractive to both casual and business users.
The downside is that it would require better programmers and hardware engineers, and not everyone can cook, as unpopular as such a thing may be.
>>53822431
If an engineer working on an engine can shave even 1mpg off the engine, they would consider that time well spent, because there's a demand for fuel efficiency in vehicles these days.
The same attitude should be taken to computers.
>>53822598
The cache does the same to RAM, as RAM does to the storage. Programs aren't bigger because of bad programming, mostly, but because of people having enough memory allow more advanced features, like smart code completion.
Then again there's the field of big data in which you process terabytes of data, using 64G RAM just to store hashes so computation takes less time. In this field both memory footprint and computation time is a very important/expensive trade off.
Daily reminder that RAM running at 1.6Ghz can't keep up with a processor running at 3.6Ghz.
RAM and GPU(average 1.0Ghz) are the biggest bottlenecks to your computer aside from mass storage devices including SSD's. If your 960 had a clock speed of 3.6Ghz it would outperform the 980 Ti. Imagine your 980 Ti at 3.6Ghz and your RAM equally as fast
notepad is a pretty lightweight program
1.4mb is peanuts when the average system has 8gb
>>53822661
Yes but there's also time/efficiency trade offs. Most companies would rather dedicate that time spent to adding features, or work on something else entirely, rather than eek performance that fewer users would notice
>>53822915
I hate to break this to you but there's no RAM on the market that can be clocked to 3.2GHz. Eat a dick.
>>53822931
When you think of what little Notepad can do, and what VIM can do, then remember that VIM was made to be ran on the Amiga, 1.4meg for Notepad is a lot.
>>53821828
Bloated code is bloated code.
>>53823082
>comparing windows notepad to a editor like vim
it's like comparing napkin sketches and a drafting table
>>53821466
there are lots of reasons that could happen, for one modern computing requires to function within and alongside modern software, such as the window manager, a multiprocessing operating system, rendering fonts, etc.
it is also better to develop more user friendly and safer software even if it means it will use a little more memory.
>>53823005
That's point, buddy. These components are still bottlenecks to the currently most powerful part of a computer, the CPU.
Why do you think race cars don't have exhausts that go to the back of the car, but instead vent out of the hood in a super wide pipe? To reduce back pressure and allow a faster performance. You rebuild engines and add bigger better cams and stronger pistons when you add massive turbos in order to handle the pressure and increase performance.
Yet here we are limiting our CPU's but not advancing RAM and VRAM chips to be as strong as a CPU can possibly handle. We should stop shrinking our dies at the cost of performance and make sure our CPU's can handle faster computing.
We need not a desktop as thin as a notebook. Current Xbox sized home desktops are fine. What we need is larger RAM sticks and faster RAM speed. We need computers that turn on in <5 available to everybody. We need cheaper SSD/PCIe SSD prices. We need graphics cards that work at much higher speeds.
We don't need smaller computers for the sake of it being smaller.
>>53823128
seems like a valid comparison to me. A program that can do much more (vim) uses fewer resource than a program that can do little (notepad). Compared to vim, notepad uses a lot of resources.
>>53821882
Leave Emacs alone. LEAVE HIM ALONE!!
>>53823283
Yeah, but Notepad runs in a window and it isn't just text based. You are just taking into account process itself. Now tell me how much RAM does that terminal emulator use.
>>53823550
Vim has several graphical frontends (or used to have at least, it seems that they only work on the gtk one nowadays)
>>53823800
And I assume it uses more than 1MB of RAM in the end... So "muh vim uses less" is again just stupid unix/linux elitist bullshit. Who the fuck cares if it's 1MB or 2MB, if in the end, you just use one instance of it at the time.
>>53824303
Depends, the Amiga version for example doesn't use some heavy toolkit/terminal emulator.
(Again, not sure if the Amiga is still being updated)
>>53823231
But Anon, we cannot afford supercomputers, which are designed for high sustained performance in mind. And high sustained performance requires high sustained RAM bandwidth.
>>53824555
But Anon, that's my point. These technologies and upgrades should be worked on for distribution at a consumer level. Would it not be a just goal to have speedy computation available for everybody? I can't believe hard drives are still a thing. SSD's should have long been in development and production. The speed they offer is outrageous. SD cards are cheap as shit. It's only a matter of time before SSD technology gets as cheap as toilet paper.
>>53823231
No one but nerds use desktop computers these days.
>>53824596
>80 decagon core and 80tb fast ram for shitposting on facebook
We only need so much power for normie tasks.
the best thing that could possibly happen for computing today is if all code, all books on code, all computers, all storage media, all everything, just fucking vanished and we had to start over.
>>53824686
>Wanting to revert society back to the 1800s because of programming elitism
Dynamic high level languages are the future anon. There is no need for other shitty languages any more.
>>53824652
There's nothing wrong with a quad core CPU or higher that can manage to load instantly any application and completely nearly any process instantly.
I'm not asking for server computers in our home. I'm asking for more advancement in currently technology. I'm asking for SSD as big as HDD and faster than current RAM.
>Hectacore Hyperthreaded CPU @ 5.0 GHz
>64Gb Quad Channel RAM @ 5.0 GHz
>2TB SSD as fast as RAM housing instantaneous OS and files
>GPU @ 5.0GHz
I was never arguing already implemented server power for home usage. Just that all parts of the computer be on the same level; enough to do everything instantly with no wait times.
>>53824721
And it's happening. We haven't stagnated, SSD sizes are increasing rapidly. Ram is a little slower, since there's so many standards and comparability in play.
>>53824746
I see far more work being done on SSD size than SSD speed. Same for VRAM. I see way more work being done on the size of VRAM rather than its performance. Because of the 4k meme.
But work is being done. Especially for SSD, Id rather take cheaper $/GB right now than speed. But still. I wish normies wouldn't have gotten into computers. It ruined the entire field. Look at sky meme. Being so small and flimsy that laptop manufacturers are being forced to solder on the CPU's to avoid bracket thickness. Bending and breaking under heatsink pressure. Literally publicly admitting cutting performance for smaller dies.
>>53821466
240p vs 1080p
>>53824789
What are you talking about?
CPU Cache, i.e. fucking fast memory, is increased all the damn time. Along with smarter caching strategies to not get bogged by external memory access
>>53824845
1080p notepad with dynamic HDR and text bump mapping.
>>53821466
>BBC Micro - 32KB of RAM
>Modern computer - 4GB of RAM
>Elite running on BBC Micro - 69% of RAM
>Notepad running on modern computer - 0.0003% of RAM
An equal comparison would be Notepad using 2.76GB of RAM.
>>53824442
Neither does Notepad on Windows 3.1.
>>53825049
I wouldn't be so sure.
Win 3.11 was bloated shit and needed several MB to function.
Workbench ran on machines with less than 512kb
>>53825029
>>Notepad running on modern computer - 0.0003% of 4GB
your general point still mostly stands, but that would only be 12kB you moron.
So why are there no modern editors written entirely in assembly?
Because assembly is ass, it's in the name.
>>53825088
No, Notepad on Windows 3.11 really won't use 1400kB of RAM. It will increase with every Windows version though. Most of it's RAM usage is loaded GUI and nothing else. You are just comparing something totally different to vim. The Windows (or more like DOS) counterpart to vim would be edit.
>>53825151
there's win32pad
>>53822848
>Programs aren't bigger because of bad programming, mostly, but because of people having enough memory allow more advanced features, like smart code completion.
this is wrong though
people just use 30 layers of indirection and whatnot
Notepad++ and pspad both load the entire file into ram and have issues with text files of a gigabyte or more
meanwhile vim:22.3MB ram usage on a text file with 55M lines
right now there is Steam with 100+50 MB and sharex with 58 MB
neither of these programs do anything particularly complex
>big data
this is about desktop computing, in big data high ressource usage is of course justifiable
>>53821466
>notepad
>1400k of memory
Eeeehhh, no.
>>53823082
the one is using a gui while the other doesn't
>>53825445
VIM's windows port uses a GUI too
>>53825219
win32pad is written in C using win32 api--wtf are you talking about?
>>53825485
and how much memory it's gui port is using?
>>53821466
Elite was programmed directly to the hardware and allows you to do fixed amount of things. Notepad is programmed to the Win32 API that allows it to run on any system running Windows from the past 20 years, and it allows you to input any kind of text and display it in any kind of font that your system has, search in it, edit it, do search & replace, etc.
Hell, a third of that 1400k is probably the font support, the other third is having a big text file open, and parsed in memory in some kind of rasterized format.
>>53825151
there is e3 though I ain't sure if you can call it modern
>>53825503
gVim is 4.5 MB after starting, 22 MB when opening a 1 GB text file
notepad is ~1.9 MB after starting and can't open said text file
>>53825492
My bad, there's was a somewhat popular editor similar to win32pad, which I confused it with, written in assembly but I don't remember the name.
I found HiEditor which is written in assembly and has more features than fucking notepad.
It's not the one I was thinking of though.
>>53825321
>sharex with 58MB
What shit are you smoking?
It only needs 0.5MB of RAM when it's running in the background. 5.6MB if the GUI is open. If you take a screenshot it shoots up to 30MB but keep in mind that a single raw 1080p screenshot RAM needs 8MB alone and I'm using several monitors. Your only valid complaint is that windows is shit at reclaiming memory.
>>53825558
>higher initial memory usage for superior capabilities
modern software in a nutshell
what you said proves the faggotry of this thread
probably the same thing applies to OP example
>>53825631
that's not true though
steam skype are very widespread and use a ton when doing literally nothing
literally just restarted it
it also crashes from time to time
maybe it's just shit on windows
altough i don't get why a c# program would work better on linux
>>53825819
forgot pic
>even ed uses 792k on a modern system
Kill me
>>53824600
nerds and people with jobs
>>53821466
>elite
>wire vector graphics
>notepad
>30 years of microsoft "Well we cant change this without breaking everything so add a hundred more lines for that new function/compatibility update"
>>53825872
that's just boilerplate stuff though
nobody really cares about a megabyte or two
losing 300 MB of ram to 3-4 applications that always run in the background bothers me though
>>53821951
Based clive
>>53825357
>tfw puush has a lower ram usage than editor
>>53821951
We're setting up code camps in ghettos now and paying high school dropouts to take coding classes, even covering their rent in exchange for them learning programming (or at least warming a chair in the classroom). Of course shoving more and more low intelligence individuals into the software development industry is going to result in crappier inefficient code.
>>53827056
i want the "everyone can program" meme to end
>>53827077
It will... when all white and Asian males are wiped out. The it will be "no one can code".
>>53827056
>shoving more and more low intelligence individuals into the software development industry
the real reason they're doing this is to drive down wages. of course the CEOs want everyone to code.
this
>>53821828
without this
>>53821854
You've written an OS with only 2 kilobytes of machine code, right /g/?
>>53829814
>2 kilobytes
Bloat, 64 bit enough.
>>53821466
>Notepad, 1400k of memory
>Elite - 1.1G of memory
>and charging $60 to land on shitty planets
Explain yourself, FDev
>>53825109
>STANDS
>>53830213
1GB for a game like Elite Dangerous is pretty good.
Unlike Elite Dangerous!
We're at a point where more power is seen as a justification to shitty optimisation, intel realised this isn't a valid long-term strategy which is why skylake isn't just more power, but more efficiency. I bet computers will be 10% faster but use 80% less power in 10 years.
>>53829814
Depends what you call an OS, I wrote a REPL / shell / text interpreter whatever you wanna call it, gets keyboard input, performs pattern matching test, if a pattern matches, execute program at indicated location, I wrote two programs for it, fizzbuzz and a guessing game, all of it was in hex, probably less than two KB, for a machine I made in logisim, slow as fuck and seriously ugly, but it worked, albeit with no argument passing or anything
>>53821466
laziness/ ineptness / bad design =
microsoft infastructure/programming
>>53829814
How inefficient. 380 bytes is more like it.
http://www.retroprogramming.com/2011/03/itsy-os-simple-preemptive-switcher.html
