Dawn operating system

[Geri]

Dawn operating system is a revolutionary technology, the first high-level operating system built for URISC computers.



Other operating systems are designed to run on extremely complex hardware environments, but Dawn operating system is designed for the SUBLEQ architecture, which have no traditional instruction set. This allows a Dawn compatible processor and computer to be designed from a millionth costs than a traditional CPU.

The goal of the Dawn operating system is to bring back the hardware and software development to the people.
Currently, the market is dominated by x86/arm, where a CPU that is capable of blicking the cursor on the screen needs 20 billion usd investment to create, and 30 years of work by 30 corporations and 1 million developers worldwide.

Traditional IT corporations created a technologic singularity, which was not able to show up any innovation in the last 20 years.
We arrived to a point, where painting a colored texture on the screen requires 1600 line of initialization and weeks of debugging - and if you are not want to do that, you can only access child toys, design apps, half gbyte sized libraries, and they are still crashing at initialization in most cases.
The internet, global forum of the free speech - actually controlled by the ISP-s and governments. A wifi device is driven by 5 million of source code lines, and nobody actually fully understants, how they work, a TCP stack is 500.000 lines of code. There are no experts at this area any more - even professionals are just typing random things in consoles to get it working if something is broken, hoping that it will randomly cure itself, becouse they cant debug 30 and 40 million code lines that is responsible for sending a bit on the cable.

Dawn operating system is different. Emulating the cpu itself is 6 source code line in C, understanding the hardware set is very simple.
Dawn operating system itself does not supports any technology, that is enemy of the freedom, while it still offers a nice graphics user-friendly graphics interface with the most common elements.
It is easy to create hardware and software for Dawn - the hardware design is well documented, and simple. Dawn have a built-in C compiler that also offers connectivity to the Dawn platform to create textured sprites, texts, play music, get data from the joystics, from webcamera, or manage the files of the computer.


Dawn is a modern graphics operating system, it supports the folowing features:
- Big-endian 64-bit SUBLEQ CPU
- Mouse
- Keyboard
- Virtual keyboard
- Touchscreen
- Multitouch
- 64-bit sound input and output on 8 channels
- 4 force-feedback joysticks
- 16 RGB cameras
- up to 32-bit RGBA graphics output, transparent screens are also supported
- An own p2p wireless networking protocol based on geolocation and requiring no external network provider cormprations
- CPU hotplug
- Disk drive hotplug
- Up to 100 individual disk-drive
- Basic tools like document reader, paint, wave player, process explorer, etc
- Some simple games like chess, amoeba
- C/C99 compiler that is capable of compiling codes written for the C programming language
- SMP support, up to basically unlimited CPU cores
- Very simple IO system that will stay constant for all future versions
- Memory and disk size support up to 2^63 bytes
- Battery, and power-saving features
- UTF8, unicode, ASCII
- Fast boot and shutdown



download: http://users.atw.hu/gerigeri/DawnOS/download.html

[sam1275]

I agree many of your opinion, and this project seems interesting.
Do you have any benchmark results for comparison?
Also, please fix your spelling on your website, example:

[Geri]

hi, thankyou, i will fix the spelling problems later.
currently i dont have any numberic benchmark results, becouse there is no phisical hardware is available yet.

but judging from the emulator, the performance of a native subleq cpu is similar to x86 or arm since it needs approx the same number of cycles to achieve complex things (like rendering an icon)

the cryptonite for subleq binary code is the multiplications, divisions (especially when using float as data type) and pointer store/reads, so when a lot of these is used in the algorythm, the performance lags behind traditional cpus with hardware multiply units.

[ataradov]

That site is blocked by my corporate firewall. There must be a good reason for that. If you want to share source code, consider things like GitHub, not file sharing sites. EDIT: It is apparently not a file sharing site, but still blocked for whatever reason. EDIT2: "This site was categorized in: Pornography " Corporate firewall knows something

I would disagree with performance evaluation of a single instruction computers. They are a nice curiosity, but not much more. Number of cycles means nothing for performance, you need to compare architectures of caches, branch predictors, prefetech buffers, and all that stuff.

and nobody actually fully understants, how they work

This is a bold claim. Because you don't understand, does not mean that nobody does.

[Geri]

yeah, transistor porn 

multiple kinds of cache and branch prediction can be inserted to almost any architectures, so you cant judge an architecture from an actual implementations of these.

about the understanding, well i didnt yet met anybody who understood x86 or networking well enough. for example they maybe knowd how to do assembly code, or what to write into a bash script to do some network redirects, but understanding is a bit more deep than this. if people would understand these technologies, there will be more than 4 x86 manufacturers. for example the specifications of an x86 cpu, from-to every instructions, interrupt handling, dma, and bios calls is approx 5000 page long, and that still lacks the specifications of how an what flags the cpu actually modifies with the specified instruction, or how it actually behaves when its called by prefixes or in various modes. if we would include all of these, the specs of an x86 cpu would be more than 100k page long - just listing and specifying everything. if somebody says he understands that, i would say he maybe understands the letters in the book, or he maybe understands the things that is written down when he reads it, but i would dubt that that person actually understands the architecture.

[ataradov]

There will be more manufacturers if Intel did not patent everything and sue for patent infringement.

In reality, there is no need for a single person to understand absolutely everything. It is impractical and not really necessary in most cases.

And I don't see now your system solves anything here. You understand it, because you are the only creator. But when you start having more software created, you will lose control and knowledge.

Also, I don't see how you will avoid making TCP/IP stack. Does your OS not support networking?  I trust Linux TCP/IP stack way more than any simplified implementation.

[Geri]

patents not exist in all contry, or not enforced in the same ways, and the become erased after time. basically x86 is too complex, thats why the industry is moved on different architectures, for example, the ARM. the alternative x86 designs, such as rdc or dmp designs having great issues even booting an operating system, x86 have ~10 cpu mode states just to control memory access, almost all of them is totally undocumented, and it late even expanded with pae. basically these alternative corporations needed a decade to even boot an os on x86, and they still cant display a youtube video with decent speed.

you can have 100 000 person - like google, microsoft, amd or intel has - so every small piece of knowledge can live in somebody, but whats tomorrow? what if these people die, becouse they become old, who will create cpus then? a terrorist bombs some cpu factories, where and how will cpus being created then if nobody have knowledge, becouse somebody tought it would be a nice trolling to require 20 million source code lines to send a bit on a wire?

x86 is becoming very much irrelevant aniways, as arm kills it (and it is becoming the new x86 since its making the same mistakes).


tcp: no, i will not implement it, i will stay with the current easy packet format i have implemented. if somebody wants it, can do his own, but no way i will do it ever.

,,But when you start having more software created, you will lose control and knowledge. ''
well i dont want to understand the working of a stylesheet software or angry cow girls game, becouse its irrelevant for the operating system. the hardware is specified well enough, so it not requires special knowledge to understand that. as the software part, i will not insert complex technologies. i will not add tcp, or not add a jpg loader, since its complex. loading a bmp file or a wav is easy, so i written the players/loaders for that. but if a software developer drops in into his software a jpg loader, thats his business, i dont have to do anything with it. but the os and the platform itself will be keeped like this, as a very simple and managable platform.

[ataradov]

patents not exist in all contry, or not enforced in the same ways,

USA is the biggest consumer. Who cares if you can make the part some place else, if you can't sell it in the US?

and they still cant display a youtube video with decent speed.

Can your OS do that?

x86 is becoming very much irrelevant aniways, as arm kills it (and it is becoming the new x86 since its making the same mistakes). 

When there is a single ARM-based computer in TOP-500, we can speak about X86 irrelevancy.

I don't like X86 either, but the reality is, it is the best performing architecture on the market today.

[Geri]

USA is the biggest consumer.

usa is only the biggest consumer of usa-made products - its globally irrelevant if your product is being throwed out, they only takes out 5% of the global popularion and 10% of the economic power. its bad if you cant get it, but it will not decide your success or failure.

Can your OS do that?

my os have no internet browser, no tcp, no flash - and if it depends on me, i will never allow these things around my OS, only free technologies.

When there is a single ARM-based computer in TOP-500, we can speak about X86 irrelevancy.

well how many top 500 computers do you own in your house? top500 is just as irrelevant as x86. 95% of the cpus manufactured is arm, x86 is less than 5%, only traditional geeks still hang with x86, regular people use arm based cell phones and tablets since years. (i hate them, what can i do: thats why i made virtual keyboard to the os)

[Monkeh]

Can your OS do that?

my os have no internet browser, no tcp, no flash - and if it depends on me, i will never allow these things around my OS, only free technologies.

Your OS has no utility.

[janoc]

Geri, I don't want to rain on your parade, but why exactly would I want to purchase your system?

- Runs on an esoteric CPU
- It is not meaningfully compatible with any existing software without (potentially complex) porting
- System obviously isn't open, because you have full control over it, including which software it can run (not even Microsoft or Apple dare to grab that kind of control!).
- It is not able to interoperate/network with any existing system.

Sorry, but this sort of thing was maybe interesting back in the 70-80s where a personal computer was a bunch of chips on a PCB or something like Apple II. Today personal computer (which you are obviously targeting, otherwise why multitouch, force feedback support or similar stuff) that doesn't support TCP/IP networking and has no software apart from what you have supplied is a complete nonstarter.

The CPU discussion is a complete red herring - almost nobody cares today about which CPU is in their phone, tablet or PC or about that ideological gobbledygook you have there. What matters is whether the machine can be used for useful work.

What you have is your personal toy. Which is fine - but then why should anyone pay you for it.

[Geri]

- Runs on an esoteric CPU
- It is not meaningfully compatible with any existing software without (potentially complex) porting
- It is not able to interoperate/network with any existing system.

maybe that is your goal. maybe you want something keep hidden from the public. maybe you need to custom-build a 60000 cpu cores to compute a simulation. maybe you want something in your pocket running from a buttonbattery and solar panel with a passive quartz display for a month of tour in the mountains. regular computer platforms cant do that, but this can. this opens new spectrums of computing, and enhances the factor of the users control over his hardware.

almost nobody today cares about which CPU is in their phone

its technically not possible to notably enhance performance of mobiles and personal computers based on x86 and arm, subleq having more potential, thats why i target that system with my OS.

[Geri]

Your OS has no utility.

i think computing is more wide than watching cat videos from youtube.

of course nothing limits people to write a tcp stack and a browser to watch youtube from my os. i just basically dont think that an operating system itself should have these technologies.

like bruce lee said, be like a sculptor, who removes something from the material instead of adding.

[Avacee]

As a developer how can I use Dawn to monetise my time, more than I currently do, to continue paying for a mortgage, new cars, various sports and hobbies, and several foreign holidays a year?

I confess to failing to see beyond any demand for this except as a "flavour of the month" cool item.

[Geri]

Online Avacee: as a software developer, i suggest you to first wait to have compatible hardware.

after that, for example you can make games for it and create a cool game console from it and sell them for kids, you can teach persons how to code in C, you can tie wires to robots and robotic components to controll them, you can build satelites, etc. you can use it on pretty much anything.

[Monkeh]

And you can do all of that with existing platforms which are already in use and have more applications - plus they run on hardware which exists!

[ataradov]

my os have no internet browser, no tcp, no flash - and if it depends on me, i will never allow these things around my OS, only free technologies.

Well, I just made even better OS. It has absolutely nothing and only exist in my imagination. It is totally bug free and awesome.

[hans]

hi, thankyou, i will fix the spelling problems later.
currently i dont have any numberic benchmark results, becouse there is no phisical hardware is available yet.

but judging from the emulator, the performance of a native subleq cpu is similar to x86 or arm since it needs approx the same number of cycles to achieve complex things (like rendering an icon)

the cryptonite for subleq binary code is the multiplications, divisions (especially when using float as data type) and pointer store/reads, so when a lot of these is used in the algorythm, the performance lags behind traditional cpus with hardware multiply units.

First of all, your website states 50GHz CPU's. Please refer one, because I don't think those exist.

Second; subleq is an abstract architecture to demonstrate a Turing machine can work using 1 instruction. It's still a highly active research area. This could give new insights but has no real practical use. Even simple addition instructions require multiple cycles to implement a basic RISC instruction. Source: https://en.wikipedia.org/wiki/One_instruction_set_computer#Subtract_and_branch_if_less_than_or_equal_to_zero

yeah, transistor porn 

multiple kinds of cache and branch prediction can be inserted to almost any architectures, so you cant judge an architecture from an actual implementations of these.

about the understanding, well i didnt yet met anybody who understood x86 or networking well enough. for example they maybe knowd how to do assembly code, or what to write into a bash script to do some network redirects, but understanding is a bit more deep than this. if people would understand these technologies, there will be more than 4 x86 manufacturers. for example the specifications of an x86 cpu, from-to every instructions, interrupt handling, dma, and bios calls is approx 5000 page long, and that still lacks the specifications of how an what flags the cpu actually modifies with the specified instruction, or how it actually behaves when its called by prefixes or in various modes. if we would include all of these, the specs of an x86 cpu would be more than 100k page long - just listing and specifying everything. if somebody says he understands that, i would say he maybe understands the letters in the book, or he maybe understands the things that is written down when he reads it, but i would dubt that that person actually understands the architecture.

CPU design, compiler design, computing architectures are common courses at universities. New computing architectures researched at these places, for example imprecise (probabilistic) computing for processing of noise insensitive data like video or large scale satellite installations.

What is essentially taught is qualitatively reasoning within a context. Nobody fully understands a project if it becomes larger than few ten thousand lines of code. Not to the point where you know exactly what is happening at each function and line of code.

The trick is you can isolate a problem to a particular domain, analyze that context really well and come up with good solutions to solve the problem.

Of the mentioned 100k pages you perhaps only need 20 pages to do your job on a specific day. These OS and CPU projects are not one man bands, but rather a workforce of thousands work on them.. through abstractions and trust.

There are the same modern society principles as that we buy food and clothes in a store.

you can have 100 000 person - like google, microsoft, amd or intel has - so every small piece of knowledge can live in somebody, but whats tomorrow? what if these people die, becouse they become old, who will create cpus then? a terrorist bombs some cpu factories, where and how will cpus being created then if nobody have knowledge, becouse somebody tought it would be a nice trolling to require 20 million source code lines to send a bit on a wire?

Knowledge is a distributed good thankfully. It's not like these handful of companies are the driving factor behind everything. Universities and spinoffs still take their fair share in this, which in case of universities is mostly open format in publications and papers. Companies can also use these inventions as they want.

What some of these companies do well is take a demand and fulfill them with current available technology.

Your OS has no utility.

i think computing is more wide than watching cat videos from youtube.

of course nothing limits people to write a tcp stack and a browser to watch youtube from my os. i just basically dont think that an operating system itself should have these technologies.

like bruce lee said, be like a sculptor, who removes something from the material instead of adding.

An operating system should facilitate programs to run properly.

In the most abstract way operating systems should protect hardware access, but also abstract it through driver models so software can be hardware independent.
This is stretched to middleware stuff like file system and protocols stacks. Adhering to standards make it is interchangeable with other machines, practically achieving the same goals but at a higher abstraction level.

[Zbig]

Can't help but think TempleOS while reading this. If you haven't heard about it and the guy behind it, look it up. Makes for quite fascinating if somewhat depressing read.

[Geri]

First of all, your website states 50GHz CPU's. Please refer one, because I don't think those exist.

online docs talking about 50 ghz ic-s, like in this doc http://www.ursi.org/proceedings/procGA02/papers/p2121.pdf

Of the mentioned 100k pages you perhaps only need 20 pages to do your job on a specific day. These OS and CPU projects are not one man bands, but rather a workforce of thousands work on them.. through abstractions and trust. There are the same modern society principles as that we buy food and clothes in a store.

of course you are right when you say that modern societys can move large groups of people to work on something. the problem is having a lot of people working together WILL NOT make you to get better quality of food, hardware, software, machinery, cloths. it causes the opposite - large groups are loosing productivity and they are forming waterheads above the society.

i am coming from a former communistic contry where everybody was forced to be part of large groups, for example the government forced the land owners to hand over they lands, then they deported them, the government forced peasants to join collectives where they farmed together in large well-organized groups.  and yes, they even manufactured CPU-s.

the result was unimaginable poverty. you were not able to buy fresh flesh on the market, becouse it was stinky, the roll was 3-4 day old and inedible, i remember not having shoes for a while but i had to walk on the street sometimes, and the melted bitumen burned my leg so i had to walk on the grass where the stones cutted my feet.

just look at the computer games nowdays, all of them are shit, the last enjoyable computer games were made approx 20 years before, the more-,,community''-like computer science produced nothing but useless shit, organizing people into eronomous groups does not creating more value, its just creating shit, just as communism did: no prosperity, no more or better goods, just stagnation, and depressed people.

if forming groups and fakely expertising something with serious faces would cause products to born, then african nations would be the number one superpowers on the world.

for example x86 dissipates approx 99,9% of its power consumption as heat, becouse its a dinosaur, a living design flaw, it probably even contributes to the global warmimg.

nowdays people work through a life, and they still cant buy a house thrown together from laces, becouse sitting in large groups just basically dont work, it not creates value.

nobody will come and point out the problems, nobody will solve the problem, the things will not fix theyself. x86 and the current conception of IT dead, nobody will ever made any notable on x86 on arm, the current stagnation with windows, linux, android will stay FOREVER, becouse they canot fixed.

what i offer, is not a miracle, but a realistic alternative, a way out from the technologic singularity. maybe the only way out.

[Geri]

Zbig: yeah, i heard from him. he told that 640x480 and 16 bit color display is a covenant with god, just like circumcision.

(both is obsolote.)

[Monkeh]

just look at the computer games nowdays, all of them are shit, the last enjoyable computer games were made approx 20 years before

Now then, that's a matter of opinion.

for example x86 dissipates approx 99,9% of its power consumption as heat, becouse its a dinosaur, a living design flaw, it probably even contributes to the global warmimg.

Uh, and your CPU which doesn't exist does how well?

And x86 is a generic term applying to a huge range of instruction sets. As for the hardware, that's even more diverse..

what i offer, is not a miracle, but a realistic alternative, a way out from the technologic singularity. maybe the only way out.

What you offer is an interesting toy project.

[Geri]

subleq cpu-s exist, however, only as prototypes, and they not yet compatible with Dawn, becouse they has to be modified a bit to be compatible with dawn io. they are far more effective compared to x86 and arm, from the same transistor count and wattage that can host 4 x86 cpu cores, you can have 2000-20000 subleq cores.

What you offer is an interesting toy project.

yes, innovation always comes from below. dont forget that 8008/8080/8086 cpus were meant to be cheap microcontroller toys (even if they alreday were infected with they problems).

[Monkeh]

dont forget that 8008/8080/8086 cpus were meant to be cheap microcontroller toys (even if they alreday were infected with they problems).

Yes, it was meant to be a microcontroller - before microcontrollers arrived.

Alternate fact!

[hans]

First of all, your website states 50GHz CPU's. Please refer one, because I don't think those exist.

online docs talking about 50 ghz ic-s, like in this doc http://www.ursi.org/proceedings/procGA02/papers/p2121.pdf

That are just single gates, analyzed at the physics level. That is still a long step from any operating digital logic IC.

Clock routing in high speed digital chips is one of the difficult and most power consuming things, and we're more than an order of magnitude out from 50GHz.

just look at the computer games nowdays, all of them are shit, the last enjoyable computer games were made approx 20 years before, the more-,,community''-like computer science produced nothing but useless shit, organizing people into eronomous groups does not creating more value, its just creating shit, just as communism did: no prosperity, no more or better goods, just stagnation, and depressed people.

Most of them are.
Not all.

for example x86 dissipates approx 99,9% of its power consumption as heat, becouse its a dinosaur, a living design flaw, it probably even contributes to the global warmimg.

I would even go as far as saying it dissipates 100%, that's how physics work.
What dominates depends on how the chip is designed, but usually its something to do with the clocks.

nobody will come and point out the problems, nobody will solve the problem, the things will not fix theyself. x86 and the current conception of IT dead, nobody will ever made any notable on x86 on arm, the current stagnation with windows, linux, android will stay FOREVER, becouse they canot fixed.

what i offer, is not a miracle, but a realistic alternative, a way out from the technologic singularity. maybe the only way out.

There have been attempts at asynchronous CPU designs. That solves the clock power problem. Sometimes they have been based on existing ISA's to stay compatible with current ecosystems. Trouble is before these CPUs were completely designed & tested the world has moved onto to faster CPUs.

The commercial drive is in synchronous systems and in particular x86/ARM, so no wonder they are the fastest.

Join an university or a research house. There are people working on solving these problems. They research interesting computing architectures. Some are dead ends, some are purely academic at first but perhaps find applications 20 years down the road.

[T3sl4co1l]

SUBLEQ is an academically interesting architecture, but it has absolutely no practical application.

Just as a Turing machine, or anything else simple and Turing-complete (like other simple TM-like programming languages, like most any of the *Fuck family; alas that's not vulgarity, that's just their name, like BrainFuck), is academically interesting, but not applied practically.

The trend has been to increase the amount of processing done per cycle, without impairing the overall speed or pipelining of the CPU.  The computation-per-cycle accessible with modern semiconductors is vast, many orders of magnitude greater than that possible with an oversimplified architecture like SUBLEQ.  It might be possible to clock a SUBLEQ at perhaps 20GHz, but it would have to run well over 20THz to even begin to compete with any top commercial CPUs.  It's physically impossible to execute extremely rapid calculations, at least on semiconductors, and with present technology.

The reason, as always, such designs exist, is because there is a market for them; if you are upset that they are expensive to design, and closed source, you are welcome to choose one of the many free and open cores available today.  You'll need to synthesize them into an FPGA, since ASIC design is still financially inaccessible to the casual programmer, but this isn't a big impact on most activities you'd be playing with.

Tim

[Hideki]

Much like solar roadways, this project is guaranteed failure, but you are of course free to use the rest of your life believing that it will become successful.

[Geri]

i dont know much about the actual design of ic-s and logical gates, i cant really form opinion in that, i just want to illustrated the wide scalability of mcu-s in correspondence with subleq, but if you know a better and more accurate way to do it, i will nodify it on the website.

There have been attempts at asynchronous CPU designs. That solves the clock power problem.

interesting and unrelated fact: originally amd designed the FX series to run the fpu at 3 ghz and the alu at 5 ghz like in the old times, but they were unsuccesfull

[Geri]

T3sl4co1l, what is your oppinion about a strong superscalar subleq server cpu? how much pipelines are possible to add, how many instructions per clock is possible? i calculated that for my os 3-10 billion subleq instruction per sec on 6-8 core is enough for multimedia usage, or 3 billion instructions on 4 cores for fluid office usage. of course the os runs far below this too, but becomes more sluggish.

[ataradov]

i calculated

Can you show those calculations?

Because you are basically talking about modern CPU frequencies, but it is quite obvious that subleq architecture requires more instruction to achieve the same same result as any RISC or CISC. So it is basically guaranteed to be slower than modern CPU at the same frequency and the same outside conditions (memory, buses, etc).

[Geri]

ataradov, i multiplied the data from the performance of the emulator.
on 300 million instruction per sec * 4 cores (emulated on my pc) it typically does the things on 0.3-2 fps

[ataradov]

So 4 cores at 300 MHz give you 2 FPS graphics? Comparable ARM cores at the same frequency play video without any problems. So I don't see the advantage here.

[Geri]

no, they dont. they have a built in hardware video decoder to play videos.

also you mistaking emulators with real hardware. in qemu, when emulating arm or x86, those system again just moving on 0.3-2 fps, this is a general speed rate with modern operating systems when they are being used in emulators.

[wraper]

x86 is becoming very much irrelevant aniways, as arm kills it

Tell this to data centers where most of the x86 profit comes from.

[ataradov]

no, they dont. they have a built in hardware video decoder to play videos.

Atmel SAM A5D3 plays 720p video at 536 MHz with purely software encoder on a single core. And that's with Linux running as the OS.

[Geri]

Hideki, no problem, i alreday reached 4000 unique visitors and 300 download, so i am alreday better than most of the kiddos with their very own Black Hat Josiph 2.0 linux distributions

maybe i should email tanenmbaum professor to ask him what grade he would give on this beautiful monolitic kernel design melted together to an unseparated gui controller in 2017

[amspire]

Geri,

Congratulations on how far you have come. I did download and run Dawn, and it said I had disk error on Drive D: at 26844074. It was slower then Windows/286 running on a 80286 CPU with 2M RAM back in the mid '80s, so you are starting a long way back.

I do think that ultimately that the current path of software is not very sustainable. As a mind game, imagine a colony ship was being prepared to leave the solar system. You couldn't use most of the current technology - you would have to limit essential technology to things that are manageable and replaceable on the ship. Definitely no closed source software and no closed source chip architectures.

Trouble is even if you can run on a beautifully simple cpu, the software will grow to be as complex as anything today. If it is successful, thousands of programmers will refine and optimise their speciality area. The OS cannot succeed if the developers are told they have to live in the past.

If there is going to be a breakthrough, I think it will be in how to be able to efficiently navigate a complex system without having to fully understand the complexity. If you think about all the towns and cities in the world, you can find them on Google maps or Openstreet maps, and at  basic level find your way around the town. You need more information and you find a local translator or guide. If you need something done, you find a local expert.

I think ultimately, software needs something like that - some kind of overriding structure for every component of the system to make it fully manageable. In this case, the local translators, guides and experts will be somehow implemented into this structure. Do not ask me how, but just as an example, if something needs recompiling, the system knows exactly how to do this. You do not need to know how to find the compiler, what version, what compilation flags, what libraries, how to install the updates, etc.

Of course if you achieve all this, you now have a hacker's delight. A cpu with no hardware security, and code that is totally easy to navigate and change.

Richard

[Geri]

ataradov - depends on the codec and the drivers. back then a 350 mhz k6 also was able to play 640x480 movies, but again, depends on the format, the codec, and the drivers. generally, anything below 2 ghz is unable to play anything, even my 1,6 ghz atom stutters a bit when playing 480p if unaccelerated nowdays.

[ataradov]

Codec is H.264. You can believe whatever you want. But I don't see the point if you are basing your comparisons on things you remember "back then", and not actual performance of the hardware. There are many things contributing to that performance, and I bet that K6 was running Windows and a ton of antivirus software in the background.

[Geri]

amspire, thankyou very much for spending some times to try my system. the block read error is actually a result of the emulator malfunction on windows, and its not related to the dawn os itself.

only the linux version of the emulator is tested properly, windows is a last minute compile (i ported it within a hour in a rush, becouse i planned to release the system by tonight), since i dont have windows at all, and i was not able to test it. i will fix the problem once i get close with a windows machine, so windows users will be also able to test it with proper and normal disk access.

[amspire]

OK. I will try in on Linux a bit later on.

[cyberfish]

First of all, your website states 50GHz CPU's. Please refer one, because I don't think those exist.

online docs talking about 50 ghz ic-s, like in this doc http://www.ursi.org/proceedings/procGA02/papers/p2121.pdf

50 GHz there refers to things like single latches, and those chips are incredibly expensive. You are definitely not getting a 50 GHz chip for $1USD. You won't be able to make a subleq CPU at anywhere near 50 GHz, for any price.

The subleq instruction is roughly of the same complexity as a conditional branch on ARM, which is one of the most complicated instructions on ARM. That means, more realistically, you are looking at maybe 100MHz chip for $5, or a 1GHz chip for $20.

Now the real reason why subleq is terrible in practice is because all common operations (like adding numbers) that can be done on ARM in one instruction will take hundreds of instructions on subleq.

There's more to an instruction set than simplicity - it needs to be useful, and that means executing common operations quickly. You can have the simplest instruction set possible (like you do here), but it's practically useless if it needs a handful of instructions per "useful instruction", and hundreds or thousands of instructions for more complicated operations like multiplications and divisions. An ARM CPU can do a multiplication in a single cycle.

A modern Intel CPU can do 4 simple instructions per clock. That's equivalent to something like a hundred subleq instructions.

[Geri]

UPDATE:
2017 march 15. 

Hotfix - Fixed the disk access bug in win32/w64 version of emulators. Download links and files are replaced.

[Geri]

(amspire, this fixes your bug under win)

[Geri]

cyberfish - a modern x86 cpu can do as many instructions per clock as many pipelines does it have. if it have 2 pipeline, it can do 2, if it have 4, it can do maximum 4 (in most cases, in theory). the same applies to subleq and arm. modern arm cpus have 3 pipelines. having more pipelines than 1 called superscalar design, and you can put ugly amouts of pipelines into a subleq cpu too if you want, more than into a very complex x86. the question is of course, how much you can effectively add to keep the things easy and effective.

having more pipelines will incrase the alu size, and that decrases the clock. so there should be a sweet spot.

if you say 50 ghz is not possible, you probably have right, maybe i will erase these datas from the website, becouse i dont want to advertise with bullshit

[ataradov]

the same applies to subleq

No, it can't. Branches are killers of pipelines, and all you have is branches.

[Geri]

if C points to the next instruction, its not a real branch, the hardware can optimize it out.

[MK14]

From reading this thread and the OP's website...

So we are talking about a 20,000 Core, 50 GHz, cpu, which is as fast (or faster) than the top Intel cpu at the time of availability.
Which only needs ONE competent Engineer for say 6 months (or so), to design it for the latest chip technology.
Unlike Intel who would need 500 or thousands of engineers, over a ten year period, to design their chips.

Crazy Intel, with their latest, 22 (between 1 and 22+ cores) core desktop/server chips, used in hundreds (or more ?) of millions and millions of computers/servers all over the world, built up over the last 46 years. You don't need the X86's thousands of instructions, silly Intel.

Just one single instruction, 50 GHz, 20,000 cores, and just a few months or so, for the lone engineer with a reasonable amount of competency, to do it all. On the latest generation 7 or 10 nm chips.

So I've got some questions for the OP...

At 50 Ghz, how much power will just a single core use ? (Ignoring some minor/tiny little problems, such as the speed of light and the laws of Physics etc) ?

What is that figure multiplied by x20,000 cores ?

How many dedicated local power stations will each cpu need Will it be practical to cool the cpus at a realistic cost ?

How exactly do you envisage your operating system and the software that runs on it, to usefully/efficiently use the x20,000 cores to do stuff, for most typical/average users ?

Why would people want to use your OS, over existing OS's, such as Linux and Windows ?

[ataradov]

You don't need the X86's thousands of instructions, silly Intel.

Nah, Intel just could not convince people to not use TCP/IP and Flash

[hans]

cyberfish - a modern x86 cpu can do as many instructions per clock as many pipelines does it have. if it have 2 pipeline, it can do 2, if it have 4, it can do maximum 4 (in most cases, in theory). the same applies to subleq and arm. modern arm cpus have 3 pipelines. having more pipelines than 1 called superscalar design, and you can put ugly amouts of pipelines into a subleq cpu too if you want, more than into a very complex x86. the question is of course, how much you can effectively add to keep the things easy and effective.

having more pipelines will incrase the alu size, and that decrases the clock. so there should be a sweet spot.

if you say 50 ghz is not possible, you probably have right, maybe i will erase these datas from the website, becouse i dont want to advertise with bullshit

x86 is CISC. This is the reason it has multiple execution units and deep pipelines to accelerate execution, because the instructions are 'complex'. Microcode in the CPU's chops up instructions into smaller instructions potentially, and also will reorder instructions when it can to keep these units and pipelines filled. Caches and branch predictors are in place to keep the machine running for as much time as possible.

Having multiple ALU's and execution units is helpful, but there probably is a saturation point. Code will eventually branch and branch predictors get it wrong.. There is a balance between all these units.
For example, Intel probably has enough transistors left to give each CPU core double the amount of ALU's. After all; GPU's occupy most of the chip area now. But if the controller will never use those extra ALU's in the datapath because of these limitations, there you have hit a saturation point.

if C points to the next instruction, its not a real branch, the hardware can optimize it out.

C is just a way to express your meanings that is translated to assembler.
SUBLEQ is an architecture that completely functions on being able to branch. This makes it horrific to accelerate.
If you do want to accelerate it, then it probably ends up in the same complexity of x86. Can it be done? Perhaps. Will it be done? Most likely not, because we already have x86 and ARM which are much faster.

[Nerull]

ataradov - depends on the codec and the drivers. back then a 350 mhz k6 also was able to play 640x480 movies, but again, depends on the format, the codec, and the drivers. generally, anything below 2 ghz is unable to play anything, even my 1,6 ghz atom stutters a bit when playing 480p if unaccelerated nowdays.

Your response to every question seems to be sticking your fingers in your ears and screaming.
 
It seems abundantly clear that you don't understand anything about the system you are proposing - which you claim is simple to understand. It's easy to rant about how terrible modern CPU design is if you no absolutely nothing about CPU design other than what you read in a reddit thread one time.

[cgroen]

I can't help it, but this thread reminds me of this one: https://www.eevblog.com/forum/eda/isn't-it-strange-that-you-can-make-better-pcbs-at-home-than-factories-make/

This is, at best, an interesting hobby project for the creator (which is also extremely ok, but don't try to turn it into a "save the world" type of technology....)

[newbrain]

Just this.

[janoc]

- Runs on an esoteric CPU
- It is not meaningfully compatible with any existing software without (potentially complex) porting
- It is not able to interoperate/network with any existing system.

maybe that is your goal. maybe you want something keep hidden from the public. maybe you need to custom-build a 60000 cpu cores to compute a simulation. maybe you want something in your pocket running from a buttonbattery and solar panel with a passive quartz display for a month of tour in the mountains. regular computer platforms cant do that, but this can. this opens new spectrums of computing, and enhances the factor of the users control over his hardware.

Sorry, that's just lunatic delusional nonsense. Especially since you are dreaming about a CPU which doesn't exist and is only a theoretical model used for research. Also I wonder why would you want a custom OS/CPU for "hiding anything from the public" - that's like painting a bulls-eye on it - "Here, look at this oddball thing!". Ever heard about "security by obscurity" and how well it does (not ) work?

And good luck with your "button battery" or "solar panel" claims - again for a CPU which does not exist. That reminds me of the Solar Roadways proponents ...

its technically not possible to notably enhance performance of mobiles and personal computers based on x86 and arm, subleq having more potential, thats why i target that system with my OS.

Right. Without going into the technical discussion whether or not your claim even makes sense, I can use those x86 and ARM cpus today. When am I going to be able to have a smartphone running your SUBLEQ one?

Sorry, but this is "reasoning" akin to - "This software X will change the world, cure cancer and fly us to the Moon and beyond! You only have to wait until a non-existing gizmo Y required to run it will appear, then you will see!". You know, I have built a time machine in my garage, now only if I could find one working flux capacitor ...

[timb]

Can't help but think TempleOS while reading this. If you haven't heard about it and the guy behind it, look it up. Makes for quite fascinating if somewhat depressing read.

This sounded familiar so I checked Wikipedia, and wow, what a blast from the past. I remember when the guy who designed it started posting on Something Awful. It became very clear, very quickly that he had some pretty serious mental problems. I believe he ended up being banned for repeatedly spamming various threads, calling everyone "Stupid Ni***rs".

Before his ban I did run TempleOS and it's 640x480 of crazy. Crazy awesome! Honestly, it's still quite an achievement for a single person. (Though he did put 10+ years of his life into it.)

It's sad, he's clearly a very smart and talented person. It just shows how debilitating mental illness can be. Vice did a good article on the guy, it details his break from reality and TempleOS a bit more.

[Geri]

MK14 - power consumption, and the available core number and clock speed depends on the transistor count and chip size. a subleq core is even much smaller than a gpu clock, and you alreday have tousands of gpu cores.

[Geri]

hans, you are confusing cisc definition with superscalar design. actually having a cisc instruction set causes the opposite, its not very effective when you are doing a superscalar design, becouse it needs extremely complex decoder. x86 chips were non-superscalar for much longer time than risc equivalents, becouse it is much hardver to have a superscalar design from x86.

[Geri]

C is just a way to express your meanings that is translated to assembler.

C is the third component in the SUBLEQ instruction that speficied the pointer to modify the eip if the result of B is smaller or equal to 0.  (yeah, its also a programming language, and a letter in ABC, but you should watch what on you reply of)

probably ends up in the same complexity of x86

its 100 million transistor on x86 just do decode an opcode, how could i end up having 100 million transistor to decode a subleq opcode?

[MK14]

MK14 - power consumption, and the available core number and clock speed depends on the transistor count and chip size. a subleq core is even much smaller than a gpu clock, and you alreday have tousands of gpu cores.

GPUs have a poor ability to cope with lots of branches, within the gpu (OpenCL etc) code. Because they are basically trying to run in SIMD mode as much as possible.
I.e. few/no branches, and instructions are only decoded for one or a smallish number of things, which is then repeated for a huge number of threads, in a typical modern gpu.

Whereas you seem to be comparing that to a cpu, with a huge number of threads to continually decode, and which consists of just branch instructions, which probably will use lots and lots of power.

Intel (others and Amd) have got many core server chips down to a fine modern art.
Approximately 8 to 32 cores per server cpu chip, use up to about 165 Watts (probably more than this).

I don't know what it costs, but it must be something like $10,000,000,000 to develop these modern server chips, and take many years and hundreds if not thousands of highly talented engineers in order to do this.

Whereas you are suggesting that your "thought experiment" for a few minutes cpu (and OS), will reach the speeds of Intel and AMDs latest offerings. For just <$100,000 in development costs, by just one, reasonable engineer.
The cpu will use a fraction of the energy, be considerably higher clock frequency, only need one instruction (NOT proven in practice yet, I believe) and while using your new, world busting, great OS.

Call me over skeptical by all means, but I'M not entirely 100% fully convinced yet at your new cpu and/or OS ideas.

[Geri]

well, the os is there, my job is done with it, hardware not depends on me, so we will see what the hardware people can get out of it. i am sure its revolutionary, thats why i wrote it. this software is begin of a new era.

[MK14]

well, the os is there, my job is done with it, hardware not depends on me, so we will see what the hardware people can get out of it. i am sure its revolutionary, thats why i wrote it. this software is begin of a new era.

When I was a young child (if I remember correctly), there were all sorts of things I thought could be done, considerably better, cheaper, faster, safer etc etc.

But now I'm much older, much more experienced and have a much better realization of the reality of the world, and why various things are the way they are.

It is very easy to say that for $99 development cost, I'm going to develop a new cpu and OS, which will run at 500000 GHz, be 10,000,000 faster than Intels best cpus, cost just $0.0001 each and be ready before lunch today.

But that does NOT mean it can really be achieved.

Why not get a book on digital electronics design and prove me wrong. How long can it take, maybe you'll have the 50 GHz, 20,000 cores cpu basically up and running before the end of this week ?
How hard can it be ?
You can make the 7 nm chips in your bathroom, how hard can that be ?

Who needs a multi (?) billion dollar chip plant to do it ?

[amspire]

Geri, what are your plans for this software. Are you releasing it as open source, or are you going to leave it as a closed demo program?

[Geri]

i am a programmer, creating a chip requires more than reading an electric book within a weekend, as programming/cooking/shaping horse also requires previous experience to do it. lets wait for the actual chip design of the people who are oriented into that direction.

[Geri]

amspire: i will probably group up to other os devs to create a bootable OS on various platforms they know, using the emulated Dawn OS as a gui. this is now the next plan.

[MK14]

amspire: i will probably group up to other os devs to create a bootable OS on various platforms they know, using the emulated Dawn OS as a gui. this is now the next plan.

The reason I have not said much about your new OS, is because once I saw the bit on your website that said I needed to send you bit coins payment up front, in order to get the proper (non-core limited) version, I gave up and ignored it.

[amspire]

...
It is very easy to say that for $99 development cost, I'm going to develop a new cpu and OS, which will run at 500000 GHz, be 10,000,000 faster than Intels best cpus, cost just $0.0001 each and be ready before lunch today.

But that does NOT mean it can really be achieved.

Why not get a book on digital electronics design and prove me wrong. How long can it take, maybe you'll have the 50 GHz, 20,000 cores cpu basically up and running before the end of this week ?
How hard can it be ?
You can make the 7 nm chips in your bathroom, how hard can that be ?

Who needs a multi (?) billion dollar chip plant to do it ?

This is not needed at all. Initially, all it needs is for a hardware guy to start doing a design and simulate the delays. There will be the tradeoff between clock speed and having to break the instruction up into more then one clock cycle to work through.

That would probably be enough for a researcher to verify if it is a workable solution or not. Once that work is done, it then becomes a technology that is out there for anyone to pick up if they wish.

The idea can be verified without actually making the chips. It might not be a solution optimised anywhere near to Intel, AMD and nVidia, but that is fine.

I certainly do not buy into Geri's notion of no Internet, no browser. If someone does pick it up, they will implement Internet, all the major compiler and interpreter languages, all the video and image codecs. Sorry Geri but real genius went into designing the tcp and ip protocols. You can make small TCP/IP stacks that work. If you want to fully optimise it so, for example, it can send out packets at full speed without waiting for the ACK to come back from the first packet, then the lines of code will get bigger. Everybody wants connectivity -  with everything.

As a project to demonstrate that it is possible to build a full OS with a GUI on a single instruction multi-core CPU, I think Geri has done an amazing job.

Richard.

[MK14]

...
It is very easy to say that for $99 development cost, I'm going to develop a new cpu and OS, which will run at 500000 GHz, be 10,000,000 faster than Intels best cpus, cost just $0.0001 each and be ready before lunch today.

But that does NOT mean it can really be achieved.

Why not get a book on digital electronics design and prove me wrong. How long can it take, maybe you'll have the 50 GHz, 20,000 cores cpu basically up and running before the end of this week ?
How hard can it be ?
You can make the 7 nm chips in your bathroom, how hard can that be ?

Who needs a multi (?) billion dollar chip plant to do it ?

This is not needed at all. Initially, all it needs is for a hardware guy to start doing a design and simulate the delays. There will be the tradeoff between clock speed and having to break the instruction up into more then one clock cycle to work through.

That would probably be enough for a researcher to verify if it is a workable solution or not. Once that work is done, it then becomes a technology that is out there for anyone to pick up if they wish.

The idea can be verified without actually making the chips. It might not be a solution optimised anywhere near to Intel, AMD and nVidia, but that is fine.

I certainly do not buy into Geri's notion of no Internet, no browser. If someone does pick it up, they will implement Internet, all the major compiler and interpreter languages, all the video and image codecs. Sorry Geri but real genius went into designing the tcp and ip protocols. You can make small TCP/IP stacks that work. If you want to fully optimise it so, for example, it can send out packets at full speed without waiting for the ACK to come back from the first packet, then the lines of code will get bigger. Everybody wants connectivity -  with everything.

Not every single idea, needs to be tried in practice to prove that it has little or no merit.

But if you think it has got merit, feel free to try it out or pay someone else to try it for you.

As a project to demonstrate that it is possible to build a full OS with a GUI on a single instruction multi-core CPU, I think Geri has done an amazing job.

Richard.

Between my last post and this post, I got bored, and decided to write a complete windows OS which runs in a web browser.
I would never joke about such serious matters.

Here it is:
http://www.windows93.net/

[amspire]

Not every single idea, needs to be tried in practice to prove that it has little or no merit.

You can prove the hardware in a simulator - there is no problem with that. You do not need to build a chip any more. Then you can prove the logic in something like a FPGA and you would know exactly how much slower it is running compared to a custom IC.

You seem to be after a final solution - this is a research project. Who would pick it up? Google for one, if there was something about this architecture that exactly fitted a concept they need to achieve. Who knows? The thing is the research has to be done first, otherwise there is nothing at all.

[MK14]

Not every single idea, needs to be tried in practice to prove that it has little or no merit.

You can prove the hardware in a simulator - there is no problem with that. You do not need to build a chip any more. Then you can prove the logic in something like a FPGA and you would know exactly how much slower it is running compared to a custom IC.

You seem to be after a final solution - this is a research project. Who would pick it up? Google for one, if there was something about this architecture that exactly fitted a concept they need to achieve. Who knows? The thing is the research has to be done first, otherwise there is nothing at all.

If you or Geri or anyone else, thinks this new cpu idea has merit, feel free to simulate it using free software, and prove its merits. Then publish the results in this thread, so that we can all see the merits of his/her "new" cpu idea.

Until such time, I will remain unconvinced.

E.g. what is 7.9 * 3.31 or 37.9 / 3.31 ?
They could be just single quick (especial add/sub, probably multiply as well, but maybe less quick with divides) instructions in a modern cpu.

How many of these wonderful, single instructions will it take to perform e.g. 7.9 * 3.31, in single or double floating point ?

If the answer is lots and lots, then it's unlikely to be a very fast cpu architecture solution. Even if we don't try and simulate the cpu, using software.

Unless you think my example (7.9 * 3.31) is unreasonable, for a modern cpu ?

[amspire]

You keep insisting it has to be as good from day one. It is a research project into a different approach. It probably needs whole new concepts in parallel programming. Current software still needs fast single core speed, whereas the kind of sharded databases run by Google and Amazon allows truly independent unsynchonised tasks to happen, even if the information is not always accurate at any point in time.

If you do a search for a product from Amazon from two different locations in a country, you can get two different results, because some shards of data available to the first server are not yet available to the second server.
 
Will floating point be fast? No I think it would be slow. The researchers might persist in a pure model with no floating point coprocessor - the realist will probably say "I am just going to whack FPUs into the thing" in spite of the sobbing from the purists.

This is not going to be my desktop in the next year or next ten years. It is an interesting idea that people like Geri want to investigate, and there is probably going to be some tasks that suits a massive number of cores. I personally have no interest in investigating this myself, but I am generally interested in geniune research into new ideas.

As I said on the first pages of posts, there will be in the future times when the whole current computer solutions have to be totally dumped, and you have to start again. I gave the example of a colony ship leaving the solar system - possible for good. They cannot take Intel CPU's and Microsoft Windows - it would be totally impossible.

Not interested  - not going to happen in our lifetimes? OK, I am fine with that.

[MK14]

You keep insisting it has to be as good from day one. It is a research project into a different approach. It probably needs whole new concepts in parallel programming. Current software still needs fast single core speed, whereas the kind of sharded databases run by Google and Amazon allows truly independent unsynchonised tasks to happen, even if the information is not always accurate at any point in time.

If you do a search for a product from Amazon from two different locations in a country, you can get two different results, because some shards of data available to the first server are not yet available to the second server.
 
Will floating point be fast? No I think it would be slow. The researchers might persist in a pure model with no floating point coprocessor - the realist will probably say "I am just going to whack FPUs into the thing" in spite of the sobbing from the purists.

This is not going to be my desktop in the next year or next ten years. It is an interesting idea that people like Geri want to investigate, and there is probably going to be some tasks that suits a massive number of cores. I personally have no interest in investigating this myself, but I am generally interested in geniune research into new ideas.

As I said on the first pages of posts, there will be in the future times when the whole current computer solutions have to be totally dumped, and you have to start again. I gave the example of a colony ship leaving the solar system - possible for good. They cannot take Intel CPU's and Microsoft Windows - it would be totally impossible.

Not interested  - not going to happen in our lifetimes? OK, I am fine with that.

The problem is that the OP is making wild claims about the huge performance (speed and power consumption, and maybe other claims, I've NOT read all available information) benefits, of this cpu.
BEFORE it even exists, or has run real life benchmarks etc.

[amspire]

The problem is that the OP is making wild claims about the huge performance (speed and power consumption, and maybe other claims, I've NOT read all available information) benefits, of this cpu.
BEFORE it even exists, or has run real life benchmarks etc.

That's what you are worried about.    You got to let guys researching new ideas be a little excited about their technology, as long as they are not ripping us off. I accept they have some familiarity with how such a processor would work and how thousands of cores would communicate and I have absolutely no idea.

What I do see is that Geri has gone to the trouble of developing an operating system with a gui running on a single instruction CPU simulator. Whether Geri turns out to be right or wrong, that achievement is very impressive. If someone is smart enough to do that, I will assume there is some genuine intelligence there.

[free_electron]

my os have no internet browser,

 and if it depends on me, i will never allow these things around my OS, only free technologies.

so, mr. dictator doesn't want his users to learn about other things cause is not going to give them internet access.

how do you share programs for this os ? kinda weird , since you need to get it from the intenret but it doesn't let you go on the internet ...

[MK14]

The problem is that the OP is making wild claims about the huge performance (speed and power consumption, and maybe other claims, I've NOT read all available information) benefits, of this cpu.
BEFORE it even exists, or has run real life benchmarks etc.

That's what you are worried about.    You got to let guys researching new ideas be a little excited about their technology, as long as they are not ripping us off. I accept they have some familiarity with how such a processor would work and how thousands of cores would communicate and I have absolutely no idea.

What I do see is that Geri has gone to the trouble of developing an operating system with a gui running on a single instruction CPU simulator. Whether Geri turns out to be right or wrong, that achievement is very impressive. If someone is smart enough to do that, I will assume there is some genuine intelligence there.

The requests for money (bitcoins etc), completely put me off from trying it, or even reading much more about it.
I'm very wary of downloading executables from low post count, new members as well.

http://users.ininet.hu/geri/DawnOS/download.html

Send 50000 Dogecoin to:
DNNJJT7fmTcYHcpT4HAMLo6UyBdiusEUvi
 


 
Send 0.02 Bitcoin to:
15Hq5Ne11bPepeaBc568CvxwsidNiFFLXW
 

 
Send 4 Litecoin to:
Leyt5DPYdcZPJiVWm2WwfqmgZWuXF44TSz

Let's explain this in a different way.

What if a beginner here, posted that they were having terrible technical problems with their circuit, would you help them ?

If the schematic they had drawn up, needed >£20 worth of bitcoins, in order to download the schematic ?

Would you still help them out ?

tl;dr
Someone starting a thread about an (apparently) very early version of software, which does not even have its cpu available yet for it.
Needs >£20 worth of bitcoins in payment, for the full version.

Basically forget it, sorry!
(I know there is a demo/limited version, but I'm way too put off by the concept of charges to bother with stuff like that)

[amspire]

so, mr. dictator doesn't want his users to learn about other things cause is not going to give them internet access.

how do you share programs for this os ? kinda weird , since you need to get it from the intenret but it doesn't let you go on the internet ...

Don't worry. If it happens, it will have Internet. If it cannot have Internet, it will never happen. Anyway, the TCP/IP protocol is free, isn't it? The RFC's documenting the protocols can be downloaded free. Not sure what Geri is on about.

[cyberfish]

cyberfish - a modern x86 cpu can do as many instructions per clock as many pipelines does it have. if it have 2 pipeline, it can do 2, if it have 4, it can do maximum 4 (in most cases, in theory). the same applies to subleq and arm. modern arm cpus have 3 pipelines. having more pipelines than 1 called superscalar design, and you can put ugly amouts of pipelines into a subleq cpu too if you want, more than into a very complex x86. the question is of course, how much you can effectively add to keep the things easy and effective.

having more pipelines will incrase the alu size, and that decrases the clock. so there should be a sweet spot.

if you say 50 ghz is not possible, you probably have right, maybe i will erase these datas from the website, becouse i dont want to advertise with bullshit

Yes, but a subleq CPU would be very difficult to make super-scalar because of both branching and data dependencies (all in memory due to lack of registers, which means you can't do register renaming). And like I mentioned before, how many instructions you can execute per second isn't important. How much useful work you can do per second is.

A subleq CPU would have a much simpler decoder and ALU than a conventional CPU, but those circuits don't take up most of the silicon area on modern processors anyways - most of the area is used for memory, which a subleq CPU would also need. Cost of a CPU mostly depends on silicon area. A subleq CPU also won't run much faster since the subleq instruction is a pretty complicated instruction by RISC standards. How fast you can clock a CPU doesn't depend on how many instructions you have - it depends mostly on your most complicated instruction.

So you have a CPU that's slightly cheaper, clocked at about the same speed, but has an extremely inefficient instruction set that requires multiple instructions to do anything useful.

When you have a "revolutionary idea" for an alternative way of doing anything, it's always a good idea to try to find out WHY people do things the current way. Is it because all the Intel and ARM CPU designers are stupid and don't know you can have a Turing-complete CPU with one instruction? Or do they know, but chose to design their instruction sets with many instructions? What are those reasons? Hint: it's not because they are stupid.

Now you can say you have different priorities and decided to take a different tradeoff - for example, you can say you really like the simplicity of having a CPU with just one instruction, and you are willing to take a 100x performance penalty for it. That would be perfectly valid.

Saying you can get a 100x SPEEDUP is delusional.

The reason why your code is so slow is not because it's running on an emulated CPU. Emulating a very simple CPU like subleq is very easy, and the slowdown should be no more than 5x at worst. That means on a modern Intel CPU, you should be getting the emulated speed of a 2010-era Intel CPU. It's clearly much much worse than that.

No, the real reason it's so slow is because the instruction set is horribly inefficient at actually getting things done.

[Geri]

 cyberfish: ,,most of the area is used for memory'' - most of the area in your bike is used by the frame, but you dont bring yourself 300 kilograms of potato just becouse ,,most of the area is used by the frame aniway''.
also my platform does not need sound chip (1-2% of regular cpu die), ethernet chip (1-2% of regular cpu die), DSP-s (5%), interrupt controller (5%), dma controller (5%), graphics acceleration (30%).
it needs cache, memory, access for polling based device management. and of course the cpu cores itself, which you can insert a few 1000 from while you still optimized down 50% of the die. or you can do far smaller or larger designs as you wish.

Emulating a very simple CPU like subleq is very easy, and the slowdown should be no more than 5x at worst.

no, at worst its like 40-50x less efficient if you emulate it. however if you use dynarec/jit methods, you can go up to 15-20% efficiency - which i didnt implemented, becouse i dont have experience in that, but if you have, you can write a very efficient emulator, and you can see it yourself. actually this thing is not a theory, you actually have the OS itself to play with, you can see your calculations yourself.

[Geri]

amspire: to access internet, you have to implement multiple layers




every layer is disasterous in some ways, and every one of them is highly undocumented. a tcp stack is 10x larger than my OS itself alone, becouse its designed to be that - a complex, impenetrable technology that limits your freedom, and places your communication in the hands of a few large corporation supported by governmental agencies. if it would be so simple, everybody would build internet devices, meanwhile there is only a few chip design available to be widely used. did you even wondered, if its so cheap, open, simple, and free, why it is costs 20 usd per month to pay for the internet? becouse its not.

[ataradov]

What are you talking about? Everything is pretty well documented in RFCs. The reason that TCP stack is bigger than your OS is that it is actually useful for something.

Your desire to build minimalist systems is great, but you can't just randomly drop features that you don't know or understand how to implement.

[Geri]

MK14: the only limitation is in the free version is that will not be able to use more than 8 cores, i think this is a fair and beneficial limitation for both me and the costumers.

[Geri]

ataradov, if the internet networking is so greatly free, why dont you create a device from sctratch that can connect to the internet ?

i will not copypaste untrustable and unrealible codes to my systems just to support something unnecessary. people if want, they write a browser for example, and add they own tcp stack, its they problem, not mine. an operating system is not a must have to obtain a tcp stack. it will be not added.

[Geri]

free_electron: ,,mr dictator'' refuses to add technologies that allows corporations to get control above you in his operating system, and tells the people to do everything on they own in they softwares, but will not add bloatwares in the OS. if that makes me a dictator, well then i am a very good dictator.

[ataradov]

ataradov, if the internet networking is so greatly free, why dont you create a device from sctratch that can connect to the internet ?

Your logic is broken here. First of all, I don't need a device that connects to the internet. Why would I make things from scratch, when there are other people that make things I can just use? Also, define "from scratch". Should I start from collecting sand to make my silicon?

I'm not saying that making TCP/IP stack is easy, it is not. That's why we use reusable code and standard implementations. I let professionals implement this stuff, and I will think about and work on stuff I care about.

[Geri]

oh, so why people even made computers, when they were the ENIAC alreday available?

[ataradov]

oh, so why people even made computers, when they were the ENIAC alreday available?

Because there was a clearly justified increase in performance. With this thing we don't have any increase in performance. The only benefit to your system is that you can understand how it works. This is of no benefit to anyone but you.

ARM instruction "mov R0, [R1 + #5]" is executed in a single cycle. Show us implementation of this in your system.

My turn to ask questions: why sun is yellow when ocean is deep?

[nugglix]

atradov, of course you've to start making a universe... FROM NOTHING!!!

In general, to tell the 800% truth: I can see into the future.
I can see a closed thread and a high chance for a banned user in a future not to far...       

[Zbig]

also my platform does not need sound chip (1-2% of regular cpu die), ethernet chip (1-2% of regular cpu die)

x86 CPUs contain neither "sound chip" nor "ethernet chip". If you think that by pulling arbitrary numbers out of your ass you hide the fact that you have very little idea what you're talking about, you're doing pretty terrible job at it.

[MK14]

MK14: the only limitation is in the free version is that will not be able to use more than 8 cores, i think this is a fair and beneficial limitation for both me and the costumers.

If that means that, larger numbers of cores, such as 20,000 cores, can't be played around with while emulating it and/or that it is closed source. I.e. we CAN'T see or change the source code. Then I'm not happy with the situation.

Even if ALL versions are made free and open source. I still have grave concerns about some of he claims being made about this product.

It is a bit like someone writes a Chess program, which plays a reasonable game of Chess, compared to a beginner. But who then claims it is the best Chess program in the world and charges >£20's worth of bitcoins for it.

Have you managed to sell many copies of your OS yet ?

[Geri]

ataradov: its like i am making a bicycle but you want a roller, and you are asking why this bicycle is a bicycle and not a roller. there is not much in common, both have wheels, but it will not be a roller, it will be a bicycle.

[Zbig]

Also, there are people who wrote their own TCP stack for fun and giggles on this very forum. You have to understand the simple fact that YOU don't understand something doesn't mean it's some secret knowledge NOBODY has.

[ataradov]

and you are asking why this bicycle is a bicycle and not a roller

No, I (and we all here) ask you to justify any of the numbers you are throwing around here. There are a lot of people here working in the industry that know what they are talking about, and your claims don't make sense. That's why we are asking for some evidence.

We don't need an emulator to judge thing "by eye". Give use examples of the code mapping between traditional instructions, and your instructions, and explain how they are better.

[Geri]

MK14: yes, i sold one copy today from the operating system

(also, bad example: i wrote a stand-alone chess program once, it was very-far from the best chess programs on the world, but i still asked money for it, becouse i worked with it, i put effort to it, i put effort advertising it, and sold copies from it. thats how capitalism works, i am sorry, but there is no great loving soviet states any more, i also have to buy food in the store, etc etc.)

also if somebody will not be able to manufacture a 20 000 core chip just becouse he have to pay ~15 usd equivalent dogecoin for the software, i would automatically assume that he is maybe not a person who could be taken seriously.

[ataradov]

MK14: yes, i sold one copy today from the operating system

Dude, c'mon, stop lying.

[Geri]

Zbig: tcp stack alone is a years of sucking, also by doing it you still just have 10% of the work done. basically if its a lifetime to do a fully working networking, maybe that networking standard sucks and not me. if it is so easy to tcp and network, you can just compile it to your software running on my OS without even bothering the thing. if.

[Monkeh]

Zbig: tcp stack alone is a years of sucking, also by doing it you still just have 10% of the work done. basically if its a lifetime to do a fully working networking, maybe that networking standard sucks and not me. if it is so easy to tcp and network, you can just compile it to your software running on my OS without even bothering the thing. if.

And because it's not easy and you don't think you can do it by yourself, it's not worth doing. That's the attitude which got us space travel! Oh, wait..

And, of course, is just going to be used by the big bads to control you.

[Geri]

MK14: yes, i sold one copy today from the operating system

Dude, c'mon, stop lying.

too bad dogecoin transactions are public

https://dogechain.info/address/DNNJJT7fmTcYHcpT4HAMLo6UyBdiusEUvi

whats the problem, it hurts your soul that i am alreday sold more from this software than you probably sold from anything together?

[ataradov]

too bad dogecoin transactions are public

Yep, and anyone can have more than one account. So fabricating a transaction like this is 2 minute work.

[Zbig]

Zbig: tcp stack alone is a years of sucking,

I prefer to code by typing but well, to each his own, I guess.

also by doing it you still just have 10% of the work done. basically if its a lifetime to do a fully working networking, maybe that networking standard sucks and not me. if it is so easy to tcp and network, you can just compile it to your software running on my OS without even bothering the thing. if.

You have a very annoying habit of throwing around made-up figures and judging everything around you by your knowledge and abilities.

[MK14]

Let me help you, by starting to write a great advert for your new OS.

Roll up, roll up, get your new mobile/cell phone OS here.

Phone conversations: Banned by OS, to protect the user from possible dictators
Any kind of internet access: Banned by OS, to protect the user from possible dictators
Sound generation by the hardware: Banned by OS to improve performance a tiny bit
...
...
...
...: Banned by OS because ...

I thought OS's were suppose to help users do what they want to do in life ?

So if you disapprove, your OS won't let me watch EEVblog's youtube videos I guess.
Since they require internet and sound access.

I think you maybe advertising your new OS in the wrong place.

tl;dr
It is up to the OS users to decide what they do in life, NOT the creator(s) of the OS.
E.g. To use the internet or not.

[Geri]

yeah, and i travelled back in time to send me a payment to justify my os on this forum.

[Monkeh]

yeah, and i travelled back in time to send me a payment to justify my os on this forum.

About as likely as a practical SUBLEQ CPU.

[Geri]

MK14: sound payback is of course supported, with software algorythms, so no dedicated sound hardware required beyond adding a dac

also the os does not *limits* or bans you from anything - it just dont have such things implemented. you still can implement these on software levels, thats none of my business. OS is not equals to the softwares running in it.

[ataradov]

Do we also need to implement process security ourselves? You really don't understand architecture of software. Why would every single application that wants internet access bring its own TCP/IP stack with it? This will be a security disaster. And how would they share share NIC resources with on OS interaction?

Also, for incoming connections, there must be only one stack, you can't have two servers running their own stacks.

[cyberfish]

cyberfish: ,,most of the area is used for memory'' - most of the area in your bike is used by the frame, but you dont bring yourself 300 kilograms of potato just becouse ,,most of the area is used by the frame aniway''.
also my platform does not need sound chip (1-2% of regular cpu die), ethernet chip (1-2% of regular cpu die), DSP-s (5%), interrupt controller (5%), dma controller (5%), graphics acceleration (30%).
it needs cache, memory, access for polling based device management. and of course the cpu cores itself, which you can insert a few 1000 from while you still optimized down 50% of the die. or you can do far smaller or larger designs as you wish.

The equivalent is saying 90% of the weight of the bike is a frame, so I'll optimize the seat, and make the bike 1/5 as heavy. Obviously that's not possible.

A die is part of a chip. Sound logic is not usually on a CPU die. Nor is ethernet. It would have to be an extremely big interrupt controller to take up 5% of the space. DMA controllers are very simple circuits. It would have to be a very small CPU to have 5% of the space taken up by the DMA controller. There are also plenty of CPUs without a GPU.

no, at worst its like 40-50x less efficient if you emulate it. however if you use dynarec/jit methods, you can go up to 15-20% efficiency - which i didnt implemented, becouse i dont have experience in that, but if you have, you can write a very efficient emulator, and you can see it yourself. actually this thing is not a theory, you actually have the OS itself to play with, you can see your calculations yourself.

No really, it's trivial to emulate subleq. You don't need an instruction decoder, and it's just a subtraction and a conditional branch. A naive implementation would be able to do that in 5 cycles on the host CPU.

There is no need/use for dynamic recompilation. You have a single instruction. Dynamic compilation takes out the instruction decoding overhead.

[MK14]

MK14: yes, i sold one copy today from the operating system

(also, bad example: i wrote a stand-alone chess program once, it was very-far from the best chess programs on the world, but i still asked money for it, becouse i worked with it, i put effort to it, i put effort advertising it, and sold copies from it. thats how capitalism works, i am sorry, but there is no great loving soviet states any more, i also have to buy food in the store, etc etc.)

also if somebody will not be able to manufacture a 20 000 core chip just becouse he have to pay ~15 usd equivalent dogecoin for the software, i would automatically assume that he is maybe not a person who could be taken seriously.

Well congratulations on getting that sale!

MK14: sound payback is of course supported, with software algorythms, so no dedicated sound hardware required beyond adding a dac

also the os does not *limits* or bans you from anything - it just dont have such things implemented. you still can implement these on software levels, thats none of my business. OS is not equals to the softwares running in it.

These days users generally come to expect quite a lot of features, to be built in as standard into the OS. I won't try and make a list here.
But things like internet access complete with a web browser, and a whole bunch of other features. Are reasonably expected, from many user orientated OS's.

The thing is, many users rely solely on the built in features of an OS. They maybe don't know how to program, or even how to download and obtain extra programs/functionality. Such users would not get on too well with your OS, by the sound of it.

[Geri]

Do we also need to implement process security ourselves?

what benefits you can list from having ,,process security'' in 2017, in the era of high level language viruses?

You really don't understand architecture of software.

probably thats why you made an os and not me why are you trying to put everything to be so personal, even if its backfiring? too bad to ruin your impartial arguments with such things.

and well if you want to share your tcp networking, you can wrap it through the OS data packet networking system, his thing can be standardized even without me - and if somebody is interested, must do it without me, since i will not implement it for the alreday introduced reasons.

[cyberfish]

free_electron: ,,mr dictator'' refuses to add technologies that allows corporations to get control above you in his operating system, and tells the people to do everything on they own in they softwares, but will not add bloatwares in the OS. if that makes me a dictator, well then i am a very good dictator.

Wait, so you are saying if I care about freedom, I should use a closed source OS written by someone I don't know, that is missing a bunch of useful features, has not been independently-audited, that I have to pay for, and runs on a non-existent CPU with a horribly inefficient instruction set...

Instead of Linux, a fully open source, full featured, high performance, well tested, and well studied, free OS that runs on about 100 different instruction set architectures?

I find it ironic that you are using "freedom" and "lack of control" to promote a closed-source OS. If I actually care about those things, there are much better options like Linux or FreeBSD.

what benefits you can list from having ,,process security'' in 2017, in the era of high level language viruses?

The benefit is that implementing virtual memory in processor allows OSes to implement high level security features (like process isolation) efficiently. If you don't care about speed, then no, you don't need that. It's not going to be magically faster than the latest Intel processor, though.

[Geri]

These days users generally come to expect quite a lot of features, to be built in as standard into the OS. I won't try and make a list here.
But things like internet access complete with a web browser, and a whole bunch of other features. Are reasonably expected, from many user orientated OS's.

The thing is, many users rely solely on the built in features of an OS. They maybe don't know how to program, or even how to download and obtain extra programs/functionality. Such users would not get on too well with your OS, by the sound of it.

you have totally right, generic online users needing a generic content-consumer operating system like android or IOS. i dont want to competite with apple or google on they own territorry - if i would like to do that, then probably i would just made a linux distribution with a nice logo (and that would probably been ended in 5 downloads and 2 replys totally). but instead, i made this, which is more like an operating system that actually creates something new to the world, opens new opportunities, new chances, opens the doors for new technologies. and it seems it actually was able to mobilize people, it creates feelings, and loud arguments. i never dreemed of that if i actually create an operating system, people will go mad from it. i feel now i actually created the most serious software in my life.

[Geri]

cyberfish: linux is approx 100 000 file and 10000 package to paint a mouse cursor and a taskbar with icons, totally from ~400 million lines of source code. this will not help you to be more free. you also can run only binary files, so the sources must be compiled. what prooves  that a virus is not placed to a random package from the 40k packages in the repo in binary form? what prooves that a virus is not written into one of the source code files hiddenly, disguised as a bugfix? nothing, just trust, and hope that the developers are correct with you.

[ataradov]

what prooves that a virus is not written into one of the source code files hiddenly, disguised as a bugfix? nothing, just trust, and hope that the developers are correct with you.

What proves any of this for your OS?

[MK14]

These days users generally come to expect quite a lot of features, to be built in as standard into the OS. I won't try and make a list here.
But things like internet access complete with a web browser, and a whole bunch of other features. Are reasonably expected, from many user orientated OS's.

The thing is, many users rely solely on the built in features of an OS. They maybe don't know how to program, or even how to download and obtain extra programs/functionality. Such users would not get on too well with your OS, by the sound of it.

you have totally right, generic online users needing a generic content-consumer operating system like android or IOS. i dont want to competite with apple or google on they own territorry - if i would like to do that, then probably i would just made a linux distribution with a nice logo (and that would probably been ended in 5 downloads and 2 replys totally). but instead, i made this, which is more like an operating system that actually creates something new to the world, opens new opportunities, new chances, opens the doors for new technologies. and it seems it actually was able to mobilize people, it creates feelings, and loud arguments. i never dreemed of that if i actually create an operating system, people will go mad from it. i feel now i actually created the most serious software in my life.

That is one way of interpreting it, but...

Analogy (please don't try this for real):
If very late tonight you were to pick up many large stones, and throw them through lots of building windows (pun accidental), and a huge number of peoples houses. Let's also assume some of them saw it was you, or you own up about it later.

Then tomorrow, perhaps there will be a huge angry mob outside your home, and maybe even some policemen. Many shouting, very angry and highly excited. Lots of noise and atmosphere.

This would be a bad, bad thing you have done. Not any kind of good.

So yes there has been much forum excitement in this thread. But that does NOT necessarily mean that you have something good. Sorry!

[Zbig]

what prooves that a virus is not written into one of the source code files hiddenly, disguised as a bugfix? nothing, just trust, and hope that the developers are correct with you.

What proves any of this for your OS?

Because (s)he knows that and the whole universe rotates around him (her?), apparently.

[janoc]

amspire: to access internet, you have to implement multiple layers




every layer is disasterous in some ways, and every one of them is highly undocumented. a tcp stack is 10x larger than my OS itself alone, becouse its designed to be that - a complex, impenetrable technology that limits your freedom, and places your communication in the hands of a few large corporation supported by governmental agencies. if it would be so simple, everybody would build internet devices, meanwhile there is only a few chip design available to be widely used. did you even wondered, if its so cheap, open, simple, and free, why it is costs 20 usd per month to pay for the internet? becouse its not.

Geri, it only shows you don't have a clue what you are talking about. That picture is not a TCP/IP stack but based on the OSI model. TCP/IP doesn't go above level 4 in that.

And I am not quite sure what is "highly undocumented":
E.g. TCP protocol (layer 4 in OSI): https://tools.ietf.org/html/rfc793

All the rest is available for decades at ietf.org as well.

a complex, impenetrable technology that limits your freedom, and places your communication in the hands of a few large corporation supported by governmental agencies

Conspirationalist nonsense.

if its so cheap, open, simple, and free, why it is costs 20 usd per month to pay for the internet? becouse its not.

What does have your ISP fee (someone has to run your DSL/cable/what-have-you, the electricity for the routers etc is not free neither) have to do with the complexity of implementation of a TCP/IP stack?

Furthermore, you do not have to develop your own - there are quite a few stacks TCP/IP available, ready for use.

You may be a competent programmer but I think you are trying to justify your lack of knowledge and tunnel vision by arguments worth of Donald Trump, trying to turn a serious problem into a "feature". It just doesn't work like that, especially not on an engineering forum.

[timb]

cyberfish: linux is approx 100 000 file and 10000 package to paint a mouse cursor and a taskbar with icons, totally from ~400 million lines of source code. this will not help you to be more free. you also can run only binary files, so the sources must be compiled. what prooves  that a virus is not placed to a random package from the 40k packages in the repo in binary form? what prooves that a virus is not written into one of the source code files hiddenly, disguised as a bugfix? nothing, just trust, and hope that the developers are correct with you.

What proves it is that you can see the source code of said bug fix for yourself! Hundreds and thousands of eyes auditing the code. Contrast  that with your closed source OS and I'll take Linux, thank you very much...

Also, if TCP/IP stacks are so large and complicated, how are they available for MCUs? For example, TI did a full implementation that fit in a few KB of flash and ran in an 8051 (the CC3000)...

I've personally used the very small TCP/IP stack that's part of Contiki on small MCUs to implement IPv6 networking (using 6LoWPAN).

It's nowhere near as complicated as you make it out to be.

(You could also use the tried and true BSD networking stack, which is what Microsoft did starting with Windows XP; it's also used in OS X and some Linux distributions. It's well tested and reasonably bug free after so many years in service.)

[Geri]

janoc, do you think the whole internet (see what you quoted) is only the (crappy) tcp stack? you dont need wifi hardwares, ethernet chips, ethernet packets, just a tcp stack, and thats all, you have internet? well then i am the one with a clue. and at least i know what i dont know, and i will not write that i dont know.

donald trump

we make computers great again

[ataradov]

you dont need wifi hardwares, ethernet chips, ethernet packets, just a tcp stack, and thats all, you have internet?

You can buy all of this stuff, you know?

[cyberfish]

cyberfish: linux is approx 100 000 file and 10000 package to paint a mouse cursor and a taskbar with icons, totally from ~400 million lines of source code. this will not help you to be more free. you also can run only binary files, so the sources must be compiled. what prooves  that a virus is not placed to a random package from the 40k packages in the repo in binary form? what prooves that a virus is not written into one of the source code files hiddenly, disguised as a bugfix? nothing, just trust, and hope that the developers are correct with you.

You can strip out 99% of Linux (eg. the whole networking stack, all the drivers, all the support for other architectures, etc), and get something as small and as limited as your OS. It will still be faster, more well tested, and best of all - open source, so I can actually see what I am running.

Making a simple OS is very easy. Making a simple OS that is actually useful isn't.

[Geri]

timb: being able to deliver a small form not necessary means its simply and not-complex. for example you can get a jpg loader from a 3000-4000 code lines, but it still stays a very complex code even if you pumped it to small - the same applies to tcp and internet related codes in my oppinion

[Geri]

You can buy all of this stuff, you know?

you can buy all kinds of stuff, why do you work, why do you create stuff, why do you actually doing things?

[Geri]

cyberfish, if you strip 99% of linux, the maximum you can do is watching a blinking cursor displayed by the vga card, and it will be still slower than my system

[ataradov]

you can buy all kinds of stuff, why do you work, why do you create stuff, why do you actually doing things?

You can't create everything. If you are working on a CPU and an OS for that CPU, then why do you feel like you need to work on peripheral hardware as well? You don't. Let wireless people work on wireless parts, let graphics people work on graphics cards.

[cyberfish]

cyberfish, if you strip 99% of linux, the maximum you can do is watching a blinking cursor displayed by the vga card, and it will be still slower than my system

It has already been explained multiple times by multiple people (with a lot more knowledge in CPU and IC design) why your system would be extremely slow even if someone actually made subleq in hardware... and you haven't answered them, and are still insisting that your system is magically faster than everything else.

Right, we are back to making baseless assertions.

[beenosam]

cyberfish, if you strip 99% of linux, the maximum you can do is watching a blinking cursor displayed by the vga card, and it will be still slower than my system

Prove it.

[Geri]

load it with qemu and proove it for yourself

[Geri]

cyberfish: i think the only logical arguments were about the power and transistor needs of the cache and ram, which of course subleq cant magically fix. of course one inorder subleq cpu core is slower than one inorder cisc core, i never disputed that - i, however, obviously emphasized the advantages and benefits of subleq. in my oppinion, its a better construction than x86 or arm, otherwise, if my oppinion would be different, i would relase my os on x86 on arm, and not on subleq.

[sam1275]

Hello Geri.
I didn't expect you under a heavy attack However I'm not going to join any side of the war
Could you tell me whether your OS is open source or close source?
Thanks.

[Geri]

ataradov: yes. i do the os, i do it as it should. somebody does the cpu, does as he things it should. somebody does the networking hardware, as it should. there are plenty of people doing what they think its good, lot of os-es, cpu-s, network hardware, and in the end, in some form of combinations, the results will show what is a step forward.

[Geri]

hi sam1275, i dont interpret it as an attack, or as a war. bold innovations always collide to acute resistance, especially if it introduces large changes. people are usually fear of the unknown.

the os is closed, it needs a license key to have more than 8 cores opreational

(years before i was a bloody-mouthed x86+opengl fanboy, a lot of my work depended on them, but now, as i know more, i would put both in the garbage bin and strew it with salt)

[beenosam]

load it with qemu and proove it for yourself

When you make extraordinary claims, the burden of proof is on the person making the claims.

[MK14]

cyberfish: i think the only logical arguments were about the power and transistor needs of the cache and ram, which of course subleq cant magically fix. of course one inorder subleq cpu core is slower than one inorder cisc core, i never disputed that - i, however, obviously emphasized the advantages and benefits of subleq. in my oppinion, its a better construction than x86 or arm, otherwise, if my oppinion would be different, i would relase my os on x86 on arm, and not on subleq.

So you seem to be saying you have chosen a hypothetical, non-existent and whom many seem to think, greatly exaggerated and unproven claims cpu over the billions (or at least a huge number), of existing cpus and their computer systems (including handheld and very small devices).

That decision would seem to be at least a little bit, unwise. It seems to go against common sense logic.

If I was designing something, should it use readily available stuff, which already exists, is well proven, low cost and a billion people have them already.

Or should I go for (for no really apparent or compelling good reason) something which does not currently exist, might never exist and even the claims for its existence may actually defeat a number of laws of Physics, such as the speed of light (as regards it's predicted performance) ?

[Geri]

beenosam: how and what will it proof to you if i load a linux kernel in qemu and watch the blinking cursor? becouse that will happen, if somebody just cant proof his calims even to himself, thats not my problem.

[Geri]

MK14, its not illogical, its risky. illogical is if you write another generic os for x86, which alreday have like 600 developed. 

[MK14]

MK14, its not illogical, its risky. illogical is if you write another generic os for x86, which alreday have like 600 developed.

But it is apparently an unnecessary risk.

[Kalvin]

What is so unique about this? A one instruction CPU and the simple hardware could be implemented in a FPGA. This is a good thing, but you should not fool yourslef into thinking that this is unique or superior The simple hardware was available already in 80's, although the clock speeds were lower compared to today's technology. Adding more cores will not increase performance as programming paradigm we have today will not scale very well for multiple cores. Of course the graphics and audio can benefit from multiple cores and SIMD instruction set. How about power efficiency? How do you handle memory bus bottleneck and how you are going to implement caches? Well, your simple CPU will not be that simple any more.

[mib]

Something that doesn't seem to have been mentioned so far is that the subleq architecture is nothing new. It was originally described in a research paper in the mid 80s. It is just one of many single instruction architectures, all of which are academically interesting but useless in the real word. If it was going to change the world it would have happened by now.

Looking at the website, other than a whole load of paranoid crazy, there's very little actual detail. Especially anything about how you've tried to solve any of the well known issues in the architecture.

How have you implemented concurrency ? As everything is using absolute addresses, threading or multiple processors just doesn't? work. And obviously you have no synchronization primitives.

Instruction density is crazily bad. On a 64bit implementation, every instruction is 24 bytes long (three pointers). Add that to the fact that it takes multiple instructions to do simple operations and you end up with massive binaries to do anything useful.

I severely doubt there's anything new or original here. You're using one of the various compilers that are available that target subleq. The GUI looks familiar - I've seen that somewhere before. I'd guess the whole os is just one of the freely available embedded os's that you've just fiddled with a little.

If you want to change the world of computing, try actually adding to the existing research. You've obviously spent a lot of time looking into subleq, so come up with some ideas to improve it. You may come up with something genuinely original that's useful on real CPU architectures.

[Geri]

mib, i manually relocate processes and threads when they are being created. api calls are served by the 0th cpu. i choosed 64 bit - knowing that it actually a huge bottleneck - becouse its no point in making 32 bit systems, didint want work for the past. you didnt seen the gui or compiler, becouse they are designed just for this os. subleq is not new, thats right, but putting a high-level graphics operating system around it is, you cant dispute that. now with this OS subleq actually become a serious thing, not just a toy any more

[cyberfish]

cyberfish: i think the only logical arguments were about the power and transistor needs of the cache and ram, which of course subleq cant magically fix. of course one inorder subleq cpu core is slower than one inorder cisc core, i never disputed that - i, however, obviously emphasized the advantages and benefits of subleq. in my oppinion, its a better construction than x86 or arm, otherwise, if my oppinion would be different, i would relase my os on x86 on arm, and not on subleq.

And the biggest one - that to use subleq to do anything useful, you need many times the number of instructions.

And the fact that a subleq CPU cannot be clocked faster than existing ARM chips, because subleq is actually a pretty complicated instruction. The fact that ARM supports other instructions doesn't matter. It's the most complicated instruction that determines how fast you can clock the chip.

So you have a CPU that can run say 1 billion subleq instructions per second, while a comparably priced ARM chip can run 1 billion ARM instructions per second. But on ARM many useful things only take one instruction, while subleq requires many (sometimes hundreds) times the number of instructions to do the same operations.

So claiming that it will be faster is unsubstantiated. It will be much slower in reality.

How, in your opinion, is subleq better? Are the advantages worth the huge speed penalty?

[Geri]

Kalvin, dont worry, tree branches were not unique, but using them as wheels was unique.

[Kalvin]

You would definitely want to add a simple MMU for improved system stability as it would protect the system when one process misbehaves and tries to write or access data beyond its intended address range.

[Geri]

cyberfish: on X area you can have for example 4 arm cores, but on the same area, you can have a few tousands of subleq cores, so its density is far bigger than arm. of course if your software is limited to 1 thread, it will be slower, since as you mentioned, subleq needs a lot of instructions to do various things. but if the algorythms are optimized for subleq, then you have something that performs almost like an asic while it just still stayed a regular SMP/neumann computer

[Geri]

Kalvin: originally i started to design it with MMU, but later i wiped it out after i realized that it would overcomplicate the hardware.

[MK14]

Kalvin: originally i started to design it with MMU, but later i wiped it out after i realized that it would overcomplicate the hardware.

What happens if the various hardware blocks, typically preset in a modern day cpu, which you've decided to drop, because they overcomplicate, or allow dictators to thrive etc etc. Are actually important for the well being of a modern day cpu and its OS/software ?

tl;dr
Are you sure this cpu would even work in practice ?

[Geri]

the os runs on the 64 bit big endian generic raw subleq instruction set without requiring any special cpu behavior.

and no, paging for example is not important, but x86 prety much enforces it if you want to run in 64 bit mode for example.

[Nerull]

cyberfish: linux is approx 100 000 file and 10000 package to paint a mouse cursor and a taskbar with icons, totally from ~400 million lines of source code. this will not help you to be more free. you also can run only binary files, so the sources must be compiled. what prooves  that a virus is not placed to a random package from the 40k packages in the repo in binary form? what prooves that a virus is not written into one of the source code files hiddenly, disguised as a bugfix? nothing, just trust, and hope that the developers are correct with you.

The more I read the more I become convinced you've just copied someone elses work and stuck your name on it. You really don't understand software or hardware, do you?

You keep talking up complexity, I suspect you have no idea how complex your software needs to be to run on your chosen CPU. You aren't working at a low level, so it's all hidden from you - and if you don't know about it, it must not exist!

[Geri]

yes, its the last unrelased work of steve jobs, also the light of the deep hardware understanding just shines out from your anus

[Kalvin]

cyberfish: on X area you can have for example 4 arm cores, but on the same area, you can have a few tousands of subleq cores, so its density is far bigger than arm. of course if your software is limited to 1 thread, it will be slower, since as you mentioned, subleq needs a lot of instructions to do various things. but if the algorythms are optimized for subleq, then you have something that performs almost like an asic while it just still stayed a regular SMP/neumann computer

How you are going to feed those multiple cores with instructions and data? How do the cores communicate with each other and transfer the data between each other? How do the cores access the external memory at the same time? The external memory access will take time and slow the system down unless you can cache the data. Let's assume that you decide that each core will have a local RAM as a cache for the instructions and the data. How much would be sufficient? What happens to the chip area as those caches will increase the die size? Also those wide buses require large die area. How to solve the heat problems as those cores will generate some heat and the heat needs to be taken out from the chip?

[Geri]

the cores can wait for flushing the cache together, they only need to obtain cache coherency when i turn them down.  when a core needs runtime call, the kernel disables it, so it can read all the datas out. cores are being disabled and reenabled several times.

[MK14]

yes, its the last unrelased work of steve jobs, also the light of the deep hardware understanding just shines out from your anus

So you're now needing to call people names, in order to answer politely put technical questions/arguments.

[Geri]

i dont really understand since when is ,,you've just copied someone elses work'' is actually a technical argument?

[Kalvin]

the cores can wait for flushing the cache together, they only need to obtain cache coherency when i turn them down.  when a core needs runtime call, the kernel disables it, so it can read all the datas out. cores are being disabled and reenabled several times.

Sounds very slow to me.

[Monkeh]

the cores can wait for flushing the cache together, they only need to obtain cache coherency when i turn them down.  when a core needs runtime call, the kernel disables it, so it can read all the datas out. cores are being disabled and reenabled several times.

Sounds very slow to me.

But it'll run at 50GHz!!!!

[Nerull]

cyberfish, if you strip 99% of linux, the maximum you can do is watching a blinking cursor displayed by the vga card, and it will be still slower than my system

What are the current capabilities of your OS running on hardware without an x86 OS and emulation stack for help?

[Geri]

Kalvin: faster than x86 where this is a very magical grey area

[Monkeh]

Kalvin: faster than x86 where this is a very magical grey area

Which is the bigger grey area, the one we can see and measure, or the one which doesn't even exist yet?

[Geri]

Monkeh: you planned to be ironic, but its actually realistic - you must sacrafice speed in some areas to get performance and speed from other areas. there were basically no way to make this faster, this just causes delays in the sync but not actually slows down the whole thing as for exampele in x86.

[Geri]

Nerull: it runs on real subleq hardware just as it runs in the emulator

[Geri]

Monkeh: undeterministic (oh my god he told the holy x86 shits in something, blasphemy!!!)

[MK14]

i dont really understand since when is ,,you've just copied someone elses work'' is actually a technical argument?

There is a huge amount of open source stuff, freely available on the internet.
So if someone asks you or claims you might have used at least some, or even lots of it, that is their "technical argument" or question.

But I accept I may be stretching the point by calling it a "technical argument".

[Monkeh]

Monkeh: undeterministic (oh my god he told the holy x86 shits in something, blasphemy!!!)

Your assumption that I bow down before the great god, Intel, is hilarious. And says a LOT about you.

Let me re-iterate: You have an interesting project. What you do not have, despite your belief, is a weapon to destroy all the hardware and software evils in the world and free humanity from the slavery of corporations and governments. And you never will have one.

Next up: Flat earth, and the Sun is Allah's bunghole shining light upon us.

[Geri]

back then i planned to use existing opensource c compiler and memory manager, but it turned out using and modifying them would be a lot more work than creating one from 0. for example there is an existing c compiler for subleq, but it does not supports data types, no float, no relocational tables, it uses special negative addresses sometimes, and hacking it seemed impossible. i had no other chances than doing my own.

[MK14]

Nerull: it runs on real subleq hardware just as it runs in the emulator

You probably need to design the cpu at a very low level (e.g. gate), to properly emulate/simulate and get an idea if it is going to really work well in practice. Your emulator is probably working at too high a level, to really show up any possible flaws in your cpu.
Such as missing (maybe major) hardware blocks and things.

[ataradov]

Can you provide some code samples as generated bu your compiler? Simple things will do. Like adding two floating point numbers, multiplying two integers.

[Geri]

Your assumption that I bow down before the great god, Intel, is hilarious. And says a LOT about you.

https://en.wikipedia.org/wiki/Sarcasm

Let me re-iterate: You have an interesting project. What you do not have, despite your belief, is a weapon to destroy all the hardware and software evils in the world and free humanity from the slavery of corporations and governments. And you never will have one.

Next up: Flat earth, and the Sun is Allah's bunghole shining light upon us.

well i actually see the x86 fans as flat earth fans

[Geri]

ataradov: menu / tools / coder / int main(){float a=1; float b=54;a+=b;;}; / click on .bin icon

from this point i must assume that you are not even seen the operating system

so you basically told its useless without even trying it?

[cyberfish]

cyberfish: on X area you can have for example 4 arm cores, but on the same area, you can have a few tousands of subleq cores, so its density is far bigger than arm. of course if your software is limited to 1 thread, it will be slower, since as you mentioned, subleq needs a lot of instructions to do various things. but if the algorythms are optimized for subleq, then you have something that performs almost like an asic while it just still stayed a regular SMP/neumann computer

No, you can't really. Using many small and simple cores is not a new idea. The NVIDIA 1080 Ti has 3584 cores. To cram that many cores onto a chip, they are already quite limited in memory access patterns and flexibility. Though obviously they still have useful (but very limited) instruction sets. On the Intel side there's also Intel Xeon Phi, with up to 72 cores.

Where did the "few thousands" number come from? Why not few? Or few tens of cores? Or few hundreds of cores? How did you make that estimation?

the cores can wait for flushing the cache together, they only need to obtain cache coherency when i turn them down.  when a core needs runtime call, the kernel disables it, so it can read all the datas out. cores are being disabled and reenabled several times.

This really shows how you have absolutely no idea at all how concurrent hardware works.

I am now also seriously doubting the originality of your program. I wasn't sure but I gave you the benefit of the doubt, but now that you have written a lot more and clearly showed that you have very little understanding of hardware, I do not believe you would be capable of writing something like that.

[Kalvin]

Kalvin: faster than x86 where this is a very magical grey area

x86 is a poor architecture and Intel has used billions to keep it going and beating the competition, but Intel has earned more billions than it has invested in to the process so beating the old horse has been a good and profitable strategy. Without spending the billions, it would have lost more billions not been able to earn even more billions.

[MK14]

ataradov: menu / tools / coder / int main(){float a=1; float b=54;a+=b;;}; / click on .bin icon

from this point i must assume that you are not even seen the operating system

so you basically told its useless without even trying it?

Has anyone in the Universe physically tried your new cpu out yet ?
(I'm not counting high level emulation).

[Geri]

I am now also seriously doubting the originality of your program. I wasn't sure but I gave you the benefit of the doubt, but now that you have written a lot more and clearly showed that you have very little understanding of hardware, I do not believe you would be capable of writing something like that.

this is the most precious praise i even got in my entry life.
but i would rething my knowledge if i would be you, becouse there is no places to get any of this code, i am sorry.

i seriously think that some people from here are actually jealos for others knowledge, and instead of trying to enhance they own knowledge, they are just getting depressed for being realize that at the end they maybe have to go work.

[Geri]

MK14: subleq is not my invention. its not MY cpu. look after it if you are more curious.

[timb]

timb: being able to deliver a small form not necessary means its simply and not-complex. for example you can get a jpg loader from a 3000-4000 code lines, but it still stays a very complex code even if you pumped it to small - the same applies to tcp and internet related codes in my oppinion

Well, your opinion is wrong. A basic, working TCP stack can be written in a hundred or so lines of code.

The fact you confuse paying for internet access and the imagined complexity of TCP/IP shows you have no idea what you're talking about, yet you act like you do. Ergo, anything you say is suspect. QED

[Geri]

timb, if tcp stack is so easy, why dont you just write a small browser with a tcp stack for my operating system?

[ataradov]

timb, if tcp stack is so easy, why dont you just write a small browser with a tcp stack for my operating system?

What TCP/IP has to do with a browser? You really have no idea how any of this technology works.

[MK14]

MK14: subleq is not my invention. its not MY cpu. look after it if you are more curious.

It was more of a rhetorical question. Attempting to highlight the fact you were criticizing someone for not trying out your OS. When the cpu, which is also a big part of this thread, has been tried by no one yet, because it simply does not exist and is apparently not emulated at a low (gate logic) level.

Strictly speaking, in order to properly try/use your OS, designed specifically for the new cpu. Needs to actually have one physically exist, in order to run. So strictly speaking, no one (including yourself) has really used your OS (properly) yet.

[Geri]

i am sorry but html requests are going through tcp

[Monkeh]

i am sorry but html requests are going through tcp

 

[thn788]

SUBLEQ is certainly an interesting concept and a single SUBLEQ-core will without any doubts be several orders of magnitude smaller than any X86 or ARM core, so that you could put thousands of them on a single die, but ever thought about how all of these cores will get their instructions and data?

Just for rough ballpark calculation let's ignore such ugly details like latencies, dependent memory fetches, etc. and only care about memory throughput.

Let's see...

For a 64-bit SUBLEQ instruction we have 0 bytes for opcode and 8 bytes for each of the 3 operands, each specifying an address. Executing such an instruction requires reading (64 bits) data from each of the addresses specified by the A and B operands, writing the calculated 64 bit result back into the address of the B operand and, depending on the outcome of the calculation, either continuing with the next sequential address or the address given by operand C. So we have 24 bytes for instruction fetch, 16 bytes data fetch and 8 byte data write-back - in total 48 bytes - for each instruction executed.

What did we assume for clock-rate? 50 GHz? So a single one of these SUBLEQ-cores requires a (L1-)memory interface capable of sustaining little more than 2.3 TBytes/s. Putting 20.000 such cores on a single die requires a memory bandwidth of... :-D

Impractical you say?  Anybody here likes to calculate the throughput provided by a 40t truck fully loaded with HDDs? All we need to do is shrink it down a little bit... :-D

[ataradov]

i am sorry but html requests are going through tcp

Yes, but nobody claimed that it is easy to make a browser, or even HTML parser. We only say that basic TCP/IP is easy.

[timb]

timb, if tcp stack is so easy, why dont you just write a small browser with a tcp stack for my operating system?

Because, I have no interest in:

A) Reinventing the wheel.

B) Terrible closed source operating systems that nobody will ever use, on a platform that will never exist.

C) Paying you for the privilege to do so.

[Geri]

MK14: well i am not an electric engineer, i cant make a cpu.

---------------------


so lets make things sort, becouse i now have to leave:

0. THANKYOU FOR THE BRUTAL NUMBER OF COMMENTS, I REALLY LIKED THE TALK WITH YOU GUYS (may it seemed the otherwise sometimes, but i really do liked and enjoyed the interest). I see that you dudes have large interest in my operating system, and this architecture. dont give up your curiousity, CONTINUE TALKING from Dawn if you wish to. Debate it, question it, accuse it, do not belive it!

1. if the emulator i have is not good enough, do your own emulators. the full hardware access is documented in menu / help / hardware.txt . without even implementing hardwares, the os will still boot (if you dont add a keyboard or mouse, and you dont paint the frame buffer, you will not be able to do much lol)

2. subleq is not a teocretical hardware, various models built mostly for studying

3.. tcp, wifi, ethernet, udp, ip, ,,internet compatibility'', etc will not be added EVER, becouse its too complex

4. no, i will not write a tcp stack, becouse its too complex

5. no, i will not write a tcp stack, becouse i dont know how to write one, becouse (see the 4. th and 3 th. points)

6. no, i will not use someone elses tcp stack, becouse I NOT ALLOW OTHER PEOPLES CODE IN MY OPERATING SYSTEM AT ALL

7. repeat the previous 4 points until your soul gets used to it, becouse i will not do a tcp stack.

8 . yes, you can compile your own programs with your own tcp stack, i have nothing to do what you do what do you do, its your work

9. subleq cpu have more advantages than disadvantages. adventage is the density, disadventage is it needs a lot of opcodes sometimes

10. i was aware of EVERY case of the problems (for example, the efficiency, the opcode size, design problems related to superscalar, the problems with code separation, the problems with cache synchronisation) discussed in the comments above. seriously. otherwise the os would not be able to run at all. and i either fixed them, i found a way for all of them to still have it working, or i just made a compromise to still have it working relatively well.

11. it will stay paid, it will not be opensource, it will stay free below 8 cores.

12. the architecture targets subleq. not arm, not x86, not mips, and not addleq or bitbitjump, or any other similar urisc architecture. The os will probably come, however, on various other platforms, but only in emulated forms.

13. The emulator i giving is just an interpeter, later i write some more advanced emulator with dynarec/jit. (an interpereted arm linux is 5x slower if you want to know)

[timb]

6. no, i will not use someone elses tcp stack, becouse I NOT ALLOW OTHER PEOPLES CODE IN MY OPERATING SYSTEM AT ALL

Then why did you just ask me to write one for you?

[MK14]

MK14: well i am not an electric engineer, i cant make a cpu.

Then maybe that means that you don't yet have the skill, experience and know how, to come up with sensible/practicable/realistic specifications and other design information for this new cpu.
E.g. 50 GHz may be a tiny bit unobtainable.

If you want lots of people, to go out and use your OS. Then maybe you have to be more flexible, with the rules and thinking, you just posted.

An OS is the peoples main medium through which the users, use the device/computer. So if you have various must haves and various no no's. You may well exclude so many activities and possible usage scenarios, that you will make the OS of significantly poor quality and maybe virtually unusable.

E.g. Banning internet access completely, would severely limit desktop computers, many possible activities.
Such as this post I'm making would be extremely hard to make without any form of internet access.

I expect many possible users would be put off, if there was a huge list of things that are banned in your OS, such as internet usage, other peoples software, other hardware drivers, anything else you don't understand and/or can't implement yourself.

tl;dr
DawnOS the OS that likes to say "NO", whenever possible!
Making the possible, impossible 

[Hideki]

Geri: What has kept this thread going is not what you have created but all your outrageous claims of the superiority of your creation.

It doesn't really matter if it's emulated or running on an actual (future) subleq CPU, Dawn is mostly irrelevant and has little practical value to anyone but yourself, and it seems you will do your best to keep it that way

[MK14]

Our forum has been very slightly discussed here:

http://forum.osdev.org/viewtopic.php?f=2&p=272439

Apparently we confirmed 50 GHz was not fully 100% realistic!

i turns i am bullshiting becouse electric engineers sayd that complex chips could not done on such big clock, only very simple signal processors

[senso]

A serious question.
Why is your OS/CPU better than anything else?

Tell me why it is so much better?

Where can I get a SUBLEQ cpu for research?
Ebay?

If you dont and wont and dont care about TCP/IP, why do you assume its millions of lines of code?
I have TCP/IP running in a micro-controller, and I garantee you that its not millions of lines, in fact, the entire system only as around 40K lines..

And no, you wont get a CPU running at 50Ghz, that paper that you linked is for interconnects as already said, AND its bleeding edge state of the art, you cant go to GoFlo and get such process's at a whim..

[Geri]

HOTFIX:
-Fixed misleading informations on the webpage
-Fixed a bug in the Dawn operating system with the wallpaper when modifying resolution
-Links and files are updated

A video that compares booting Dawn OS on emulated 300 mhz SUBLEQ cpu, and Linux emulated on 300 mhz ARM CPU:

[cyberfish]

I have designed many embedded systems that boot up in less than 0.1 seconds, on a 16 MHz CPU.

Now of course the question is, are they anywhere near as capable as Linux? And the answer is no.

[ataradov]

Linux kernel was done loading at 0:19. The rest of it software that you claim is not going to be implemented in your OS (networking and all that other "hard" stuff), so remove it from the boot.

Although I feel like this feed the troll session ran its course and it is not funny anymore.

[alex89]

Just a reminder ..

[Wilksey]

What in the buggering jesus is this drivel?

How can you compare Linux to that 1980's looking monstrosity!

The latest craze is IoT, something which you would think would require a TCP stack at minimum!

What kind of "games" console do you think you could make with that garbage? Even the Sega Master System would work better than that thing!  Does it support OpenGL? Probably not as it's "too complicated".

And as for selling it, well, i'd like to see your sales figures for the first year, I hope you have a full time income elsewhere as you'd struggle to feed a sock puppet with what you would make from this!

Making an OS for a CPU that doesn't exist and just "emulating" it, means it probably wouldn't work on real hardware and even if it did by some miracle it'd probably mess up the timings.

What a load of sausage gravy!

[Geri]

Wilksey: so basically thats what you afraid of - if systems like this are succeeding then you will have to get a real job

[MK14]

Wilksey: so basically thats what you afraid of - if systems like this are succeeding then you will have to get a real job

Please explain/justify the succeeding

It will also help put my mind at rest, that you are not either Trolling or badly Mistaken.

[Geri]

ataradov: you compare a kernel with an os? very great mistake, linux can even only do text mode on it own.


/ this was me for today /

[MK14]

Just a reminder ..

Thanks!

[senso]

A serious question.
Why is your OS/CPU better than anything else?

Tell me why it is so much better?

Where can I get a SUBLEQ cpu for research?
Ebay?

If you dont and wont and dont care about TCP/IP, why do you assume its millions of lines of code?
I have TCP/IP running in a micro-controller, and I garantee you that its not millions of lines, in fact, the entire system only as around 40K lines..

And no, you wont get a CPU running at 50Ghz, that paper that you linked is for interconnects as already said, AND its bleeding edge state of the art, you cant go to GoFlo and get such process's at a whim..

Care to reply to this questions?

[ataradov]

ataradov: you compare a kernel with an os? very great mistake, linux can even only do text mode on it own.

No, I am comparing a useful OS with networking and other services to something that can't really be called an OS.

Cripple the Linux to your level and then compare how fast it boots.

[beenosam]

ataradov: you compare a kernel with an os? very great mistake, linux can even only do text mode on it own.

No, I am comparing a useful OS with networking and other services to something that can't really be called an OS.

Cripple the Linux to your level and then compare how fast it boots.

It is honestly painful to watch someone so clueless act so stupid. 

[DBecker]

I can add a bit of perspective.

In 2001 we had a build of the Linux kernel that could boot in under a second, and subsequently run regular Linux applications.  (There was a bit of a trick -- it was a 'slave node' in a Beowulf cluster, and initially would only run applications migrated from a full-featured master node.  The time did not count loading the kernel image into memory, and the kernel image did not include low level disk drivers which enforce significant delays waiting for likely-not-installed disks to respond.  The network drivers did not not trigger re-autonegotiation, and the video mode was left as a text console.)

Writing a TCP/IP stack is easy.  Refining that to one that interoperates with everything out there takes several years.  The situation is much improved for IPv4, since most of the flawed stacks we dealt with in the early 1990s are now gone.  Countering that is the complexity of IPv6, higher bandwidth-latency product, and the many higher level protocols that expected to exist.

Writing a quality compiler is now much more difficult.  I did a port of GCC to a then-new processor in just a few months, as a casual side project, when GCC was new.  I couldn't even learn my options in that time frame now.  (Wow, it's scary to think that GCC is now over three decades old.  It's been around for more than half the time that compilers have existed.)

[helius]

How did I know this thread would turn into a soap opera?

[Geri]

senso/others: yes, i will reply. sorry for making you to have wait, but i dont have now much time to reply or read, as i previously written. i will try to reply within a few days. i have found a bugs in the os that also must be fixed, and i have other things to do. i received totally more than 500 comments and questions from the OS. please be patient, and sorry for the wait.

[free_electron]

dawn OS . where os means obsolete system ...

[Wilksey]

I have a real job, working with real hardware (that exists and I can buy) with a real operating system with real people.
What i'm afraid of is someone will believe you and your "next best thing" ramblings and actually part with their hard earned real money.

For your next greatest achievement, build an operating system using only 1 instruction...nop

Watch out for the matrix.

[cgroen]

senso/others: yes, i will reply. sorry for making you to have wait, but i dont have now much time to reply or read, as i previously written. i will try to reply within a few days. i have found a bugs in the os that also must be fixed, and i have other things to do. i received totally more than 500 comments and questions from the OS. please be patient, and sorry for the wait.

Geee, there must be a lot of new business coming for your new nonsense os...or nor ?
You sound like a guy with some skills, my advice, use these for something else than this, or maybe talk with some professionals ?

[free_electron]

You could ask these guys for a quantum machine ...

http://www.atomchip.com/id67.html

[Geri]

2017, march 16. - Compiler ang GUI bugs
Several bugs fixed in the compiler related to inline function calls and runtime library.
A bug sometimes caused missing re-rendering of texts.


now i have a bit time, and i will make now very short answers to the newer questions.

What i'm afraid of is someone will believe you and your "next best thing" ramblings and actually part with their hard earned real money.

what im afraid is that masses buying 400 usd x86 cpus with 200 usd motherboards and 600 usd graphics cards becouse millions of people fake-expertising them, and posing for them with little knowledge with they little bash scripts, routers, shitty games in forums, governments, corporations and chaining humanity to these ineffective, slow and dangerous technologies, becouse they refuse to learn, innovate, and make any notable technologic expansion in the last 25 years.

cgroen: i guess the great way to profit from my *some* skills would be to do works directly for you, while you watching tv from the money i earn you. these times ending in the IT market nowdays.

free_electron: from what viewpoint dawn is obsolote? its more modern in every aspect than android, windows, or linux.

helius:


DBecker: thats what is usually happening with the operating systems, and most of the software. there is a good and effective implementation, which boots instantly, works very fast with very small resources, and as more and more people starts to work on it, from time to time it starts to gain more ,,weight'', its becoming slower, people start to add random unnecessary things that making it even more slower, and in the end you need gbytes of disk space to gave. thats what happened to windows from 95 to 98, from 98 to xp, and from xp to 7. and thats what happened to linux too. these versions bring a lot of better icons, better kernel and hardware compatibility, but in reality they just growing and growing, without any real innovation. now if you want a debian with lxde, text editor, etc, you need 2 gbyte of disk space, and an 1 ghz cpu to have it working, while the 4-5 year old versions needed only 1 gbyte and a 600-700 mhz cpu to achieve the same things. at this rate, a working linux distribution with the basic packages will require minimum 8 gbyte disk and 2 ghz cpu by 2020. this is a kind of problem i dont want in dawn, so i actually limited the boot ,,fragment'' of the disks to fix 256 mbyte. i agree that writing a compiler is very hard, with the compiler alone was more work probably than the other parts of the os.

ataradov: folowing your logic, linux cant even called a real os, becouse it have no piano or chess built in while my system can. pointing some features important for you will not decide what an OS is, there are several os-es for different purposes, and i am pretty sure if i would hack the networking out from linux will still boot much slower than my system. also my system has networking, its just not tcp/ip based, so saying it not have networking is bullshit, becouse it technically has.

senso: the best advantage of subleq is the simplicity of the build-up. this makes it easyer to design, manufacture, as it discussed before. currently you can buy an fpga, and do a subleq design for your own. later i will contact to some subleq-oriented persons with related fpga developments, so there will be maybe available to just download one directly for the people who wants it. the tcp is overdiscussed, the 50 ghz was wrong data and i removed that section from the website.

[Wilksey]

Good luck with your "50GHz" clock 

In what way is your thing better than Windows, Android, Linux etc?

What are you emulating this thing on?

I can emulate a Nintendo on my machine, does that make it better? Nope, but I can go on EBay and buy a Nintendo, can I go on EBay and buy one of these single instruction things? Nope.

Can you let me know who convinced you in the first place about this, I need a really good bullshitter of a sales person, and they sound just the ticket!

[Geri]

What are you emulating this thing on?

do you even know what an emulator is?

[Wilksey]

Nope, no idea what an emulator is, presume it's some kind of mythical beast that exists beyond this mortal world 

[JoeN]

This is the best troll thread I have read in some time and to think I am only through the first page.   

New user, 1 thread, 28 hours, 79 posts himself on this one thread, garnering 128 replies, nothing but nonsense.

A prize should be awarded.

[cgroen]

cgroen: i guess the great way to profit from my *some* skills would be to do works directly for you, while you watching tv from the money i earn you. these times ending in the IT market nowdays.

Ehhh, nope, I would be busy figuring out why I hired you in the first place 
I have been in the software business for almost 35 years now (as a professional) and once in a while someone that claims to "Know it all" pops up. It always, I repeat, always ends the same way. This thread was "kind of" fun in the beginning, now its just sad 

[Tac Eht Xilef]

For your next greatest achievement, build an operating system using only 1 instruction...nop

[JPortici]

professionals ?

of which sector? 
here we have a guy that can apparently write an os but can't write a tcp/ip stack (if he's so paranoid about using other people's code)
what's going on next? no usb?

i thank whoever pointed to Temple OS, i totally forgot about that

[cgroen]

professionals ?

of which sector? 
here we have a guy that can apparently write an os but can't write a tcp/ip stack (if he's so paranoid about using other people's code)
what's going on next? no usb?

i thank whoever pointed to Temple OS, i totally forgot about that

 
Drop all interfaces completely, makes it much easier. You would also have no problems with users (as there would be none).....

[beenosam]

2017, march 16. - Compiler ang GUI bugs
Several bugs fixed in the compiler related to inline function calls and runtime library.
A bug sometimes caused missing re-rendering of texts.


now i have a bit time, and i will make now very short answers to the newer questions.

What i'm afraid of is someone will believe you and your "next best thing" ramblings and actually part with their hard earned real money.

what im afraid is that masses buying 400 usd x86 cpus with 200 usd motherboards and 600 usd graphics cards becouse millions of people fake-expertising them, and posing for them with little knowledge with they little bash scripts, routers, shitty games in forums, governments, corporations and chaining humanity to these ineffective, slow and dangerous technologies, becouse they refuse to learn, innovate, and make any notable technologic expansion in the last 25 years.

cgroen: i guess the great way to profit from my *some* skills would be to do works directly for you, while you watching tv from the money i earn you. these times ending in the IT market nowdays.

free_electron: from what viewpoint dawn is obsolote? its more modern in every aspect than android, windows, or linux.

helius:


DBecker: thats what is usually happening with the operating systems, and most of the software. there is a good and effective implementation, which boots instantly, works very fast with very small resources, and as more and more people starts to work on it, from time to time it starts to gain more ,,weight'', its becoming slower, people start to add random unnecessary things that making it even more slower, and in the end you need gbytes of disk space to gave. thats what happened to windows from 95 to 98, from 98 to xp, and from xp to 7. and thats what happened to linux too. these versions bring a lot of better icons, better kernel and hardware compatibility, but in reality they just growing and growing, without any real innovation. now if you want a debian with lxde, text editor, etc, you need 2 gbyte of disk space, and an 1 ghz cpu to have it working, while the 4-5 year old versions needed only 1 gbyte and a 600-700 mhz cpu to achieve the same things. at this rate, a working linux distribution with the basic packages will require minimum 8 gbyte disk and 2 ghz cpu by 2020. this is a kind of problem i dont want in dawn, so i actually limited the boot ,,fragment'' of the disks to fix 256 mbyte. i agree that writing a compiler is very hard, with the compiler alone was more work probably than the other parts of the os.

ataradov: folowing your logic, linux cant even called a real os, becouse it have no piano or chess built in while my system can. pointing some features important for you will not decide what an OS is, there are several os-es for different purposes, and i am pretty sure if i would hack the networking out from linux will still boot much slower than my system. also my system has networking, its just not tcp/ip based, so saying it not have networking is bullshit, becouse it technically has.

senso: the best advantage of subleq is the simplicity of the build-up. this makes it easyer to design, manufacture, as it discussed before. currently you can buy an fpga, and do a subleq design for your own. later i will contact to some subleq-oriented persons with related fpga developments, so there will be maybe available to just download one directly for the people who wants it. the tcp is overdiscussed, the 50 ghz was wrong data and i removed that section from the website.

Is that your master plan? Blind people with your chess and piano built-in so they go "oh this OS is crap, I can't even go online". Unfortunately for you, an OS isn't defined based on what programs that are built-in. You can't compare chess or piano programs to anything like the internet or browsers...and if you do then you clearly have your head shoved so far up your ass.  Maybe you should take it out.

Internet is a necessity, connectivity is a necessity. You're being a dictator moron claiming all this bullshit about x86 and then trying to enforce what YOU want on anyone stupid enough to use your "OS". Having the ability to go online in an OS isn't an option, it is a MUST. Just because you are too asinine to admit it, doesn't mean it isn't true.

[Zero999]

thats what is usually happening with the operating systems, and most of the software. there is a good and effective implementation, which boots instantly, works very fast with very small resources, and as more and more people starts to work on it, from time to time it starts to gain more ,,weight'', its becoming slower, people start to add random unnecessary things that making it even more slower, and in the end you need gbytes of disk space to gave. thats what happened to windows from 95 to 98, from 98 to xp, and from xp to 7. and thats what happened to linux too. these versions bring a lot of better icons, better kernel and hardware compatibility, but in reality they just growing and growing, without any real innovation. now if you want a debian with lxde, text editor, etc, you need 2 gbyte of disk space, and an 1 ghz cpu to have it working, while the 4-5 year old versions needed only 1 gbyte and a 600-700 mhz cpu to achieve the same things. at this rate, a working linux distribution with the basic packages will require minimum 8 gbyte disk and 2 ghz cpu by 2020. this is a kind of problem i dont want in dawn, so i actually limited the boot ,,fragment'' of the disks to fix 256 mbyte. i agree that writing a compiler is very hard, with the compiler alone was more work probably than the other parts of the os.

Fortunately that's not been so much of a problem over the past ten or so years. In the past software and operating systems and software has grown at the same pace as hardware but over the last ten years, software has grown very little, yet hardware has continued to grow.

With modern operating systems, it's indeed possible to use a ten year old computer. A modern OS such as Windows 10 or a typical Linux distribution will work perfectly well on a ten year old computer. At home I use PCLinuxOS on a PC made in 2006, with 1GB of RAM and a single code 2GHz 64-bit CPU. It works perfectly for what I use it for. Of course I can't play the latest games but I don't want to do that.

This wasn't the case ten years ago. Try running Vista on a computer made in 1997 and you'll see what I mean.

[senso]

For all the bad things that Windows10 as, I find it amazing that I can get a random core2duo laptop, install it, and everything works, even the Fn combos for brightness and sound control.

You didnt really answer me, why do you think that SUBLEQ is superior to everything when every instruction must be emulated by dozens/thousands of subleq instructions?
It will be slower than a z80 running CP/M :/

[CM800]

Agreed,

I think he needs to get himself an FPGA and try see how it works out...

[senso]

I will say that out of curiosity I download the compressed thing, uploaded it to virustotal, and ran the exe inside a VM, and its very,very,very slow(in a i7-4700mq in a laptop with an SSD), I tried to use it for 20 minutes, but its way to slow to even try out.

[beenosam]

I will say that out of curiosity I download the compressed thing, uploaded it to virustotal, and ran the exe inside a VM, and its very,very,very slow(in a i7-4700mq in a laptop with an SSD), I tried to use it for 20 minutes, but its way to slow to even try out.

That's because your processor isn't 500000 GHz! And the TCP/IP stack in your virtual machine is slowing it down by at least 1000 times!

[Kalvin]

Here's a nice tutorial on operating systems basics:

https://www.tutorialspoint.com/operating_system/

The OS needs to implement quite a many features to be considered as secure and usable, and the hardware needs to be able to provide the support and enforce those features. There will be a compromise between simplicity and performance, so you could get away with a simpler implementation which will have a slight penalty compared to a more complex yet higher performing system. Without networking the OS is pretty much useless. Without good selection of software and applications nobody wants to use it. Although you want to keep your system simple - which is a good thing - the system still has to provide the features that the users need and expect from a modern system and it should perform well enough compared to the existing systems.

There are some simple yet complete operating systems available, Oberon being one of them implementable in an FPGA:
https://en.wikipedia.org/wiki/Oberon_(operating_system)

[newbrain]

(S)He needs much more than an FPGA, from several points of view, both technical and non-technical.

A number of things made me immediately skeptical :

• The conspiracy theory angle on every consolidated technology and even open standards.
• The wild numbers, thrown around without justification.
• The complete lack of information about the design of this "OS":
  
  • File system?
  • Process and threading model?
  • Homebrew C compiler? Validation? Supported standards?
  • HW interfaces?
    The only known thing is the CPU, which is usually immaterial (not in this case, though ) and the flat memory model.
  And no, I'm not going to install it to see the documentation inside.
  If one wants people to be interested, the information must be accessible with minimum effort.
• Also, I wonder whether any open source code has been used.
  Maybe not, it might be spying on the user.

But really, one of the most amusing threads since longtime.

For all the bad things that Windows10 as, I find it amazing that I can get a random core2duo laptop, install it, and everything works, even the Fn combos for brightness and sound control.

Quoted for truth: when I'm in Italy, I have a 2009 vintage desktop (i5-750) and laptop (T4400).
Windows10 runs fast and flawlessly on both. Decent gaming, too, on the desktop.

[Alex Eisenhut]

[free_electron]

it is obsolete because
- there is no hardware to run it (only emulation)
- nobody will ever build the subleq cpu in a commercial setting
- the writer of the OS does not allow for connectivity to other machines.
- the writer of the os does not allow anyone else to write code for it.

so yeah. it is obsolete.

[Monkeh]

it is obsolete because
- there is no hardware to run it (only emulation)
- nobody will ever build the subleq cpu in a commercial setting
- the writer of the OS does not allow for connectivity to other machines.
- the writer of the os does not allow anyone else to write code for it.

so yeah. it is obsolete.

That's not obsolete, that's stillborn.

[Zero999]

The conspiracy theory angle on every consolidated technology and even open standards.

Which I thought was odd and very hypocritical too, as the OS isn't even open source.

The website talks of freedom, which is very misleading, given the proprietary nature of the operating system.

[bitwelder]

The conspiracy theory angle on every consolidated technology and even open standards.

Which I thought was odd and very hypocritical too, as the OS isn't even open source.

The website talks of freedom, which is very misleading, given the proprietary nature of the operating system.

And then (s)he proposes to those who say that writing a TCP/IP stack is not difficult to write one for the Dawn OS.
 
(which will be never include in the main code anyway, because of trust issues)
 

[beenosam]

Has there been any updates to this or did the thread starter come to their senses? Heh.

[Zbig]

Has there been any updates to this or did the thread starter come to their senses? Heh.

He/she keeps babbling at http://forum.osdev.org/viewtopic.php?f=2&p=272439. Apparently tried to write a bare metal emulator for his/her OS. Failing that, Geri has concluded that PCs are unable to run any operating systems other than DOS, Windows and Linux because of "hardware imperialists". Then some more pseudo-political, anti-establishment drivel and more large arbitrary numbers being thrown left, right and center. I must admit I was checking this thread for some more comedic guilty pleasure but now, that it's clear to me the person has some serious mental health issues, it just doesn't feel right anymore. The whole story is kind of sad.

[MK14]

Has there been any updates to this or did the thread starter come to their senses? Heh.

While this thread was active, elsewhere on the internet (it seems to have gone now), I found that on questioning by others (another forum), Geri admitted that the plan was to release the OP's opening post, across internet forums on the 1st April 2017. On purpose to apparently cause a stir etc.
I will let you draw your own conclusions on their motivations.

tl;dr
This thread may have been an early test run, for the April 1st launch date.

EDIT: Found source, please see new post below:

[Zbig]

While this thread was active, elsewhere on the internet (it seems to have gone now), I found that on questioning by others (another forum), Geri admitted that the plan was to release the OP's opening post, across internet forums on the 1st April 2017. On purpose to apparently cause a stir etc.
I will let you draw your own conclusions on their motivations.

tl;dr
This thread may have been an early test run, for the April 1st launch date.

That would mean I've been played like a fiddle. But then again, I'm not sure if making people wonder whether you're joking or crazy counts as a successful prank...

EDIT:
Now I can see how (s)he could possibly be envisioning all the "people duped into paying crypto-currency for a software for a non-existing CPU" headlines on The Register and whatnot and thinking it could actually be a good idea.

[MK14]

My post above this, was from memory.
After much digging, I found my source for the April 1st remarks.

https://arstechnica.com/civis/viewtopic.php?f=14&t=1376533

Wait, what? Is it April 1st already?

Geri
Smack-Fu Master, in training
Registered: Mar 14, 2017
Posts: 11
Reply with quotePosted: Wed Mar 15, 2017 5:00 am
CommanderJameson: originally i thinked on releasing at april 1, so people would think that is a joke, and they being shocked to see that its actually real. this will actually happen, becouse i will register it to a lot of places on april 1.

[MK14]

That would mean I've been played like a fiddle. But then again, I'm not sure if making people wonder whether you're joking or crazy counts as a successful prank...

EDIT:
Now I can see how (s)he could possibly be envisioning all the "people duped into paying crypto-currency for a software for a non-existing CPU" headlines on The Register and whatnot and thinking it could actually be a good idea.

I'm still not clear as to if it was to get money on the crpto-currency coins, mental health issues, viral advertising (dogbitcoin thingies), April 1st prank, or just plain old Trolling.

[ataradov]

Geri has concluded that PCs are unable to run any operating systems other than DOS, Windows and Linux because of "hardware imperialists"

Oh, that thread over there is a goldmine:

basically x86 computers can only run windows, dos, and linux properly, they are not capable of hosting new operating system technologies any more - its true that platforms like x86 and arm are enemy of freedom - and thats why dawn operating system is born aniway - to take back the control from the hardware-imperialists.

So I guess we have successfully evicted a troll form here. Good job, everyone

EDIT: And it gets even better:

2 billion transistors of imperialism.

[beenosam]

Geri has concluded that PCs are unable to run any operating systems other than DOS, Windows and Linux because of "hardware imperialists"

Oh, that thread over there is a goldmine:

basically x86 computers can only run windows, dos, and linux properly, they are not capable of hosting new operating system technologies any more - its true that platforms like x86 and arm are enemy of freedom - and thats why dawn operating system is born aniway - to take back the control from the hardware-imperialists.

So I guess we have successfully evicted a troll form here. Good job, everyone

EDIT: And it gets even better:

2 billion transistors of imperialism.

That is the most amazing thing I've ever heard.

[Wilksey]

One of the "features" of this "revolutionary" OS is...wait for it...CPU Hotplug! 
Oh, transparent screens, they are a must these days, and don't forget the peer to peer wireless networking using geolocation!  All good and clever stuff!

They state it is "easy" to create hardware for this OS, yet they can't find any. 

[Geri]

beenosam: was not much interesting updates recently, only minor bugfixes nd optimizations in the OS itself, that can be read on the webpage:

2017, march 29. - Optimizations and bugfix
The error line handling in compiler is improved.
Imrpovements to fix mouse and network packet losses on low rendering speed
Operating system speed is improved by 10%


2017, march 21. - Optimizations and license
A restriction is lifted from the License.
5% performance gain in GUI.

yes, i did some experiments related to create a bootable dawn-compatible emulator on x86, as someone alreday mentioned, but i am not satisfyed with the results so i made some minor optimisations for the existing emulator for now. (i have talked to someone who maybe tryes to make a bootable emu, i am not sure if he really will put the efforts in, or not)

now nothing spectacular will upcome for a while, only important bugfixes. as i find them, i fix them.

[Geri]

My post above this, was from memory.
After much digging, I found my source for the April 1st remarks.

https://arstechnica.com/civis/viewtopic.php?f=14&t=1376533

Wait, what? Is it April 1st already?

Geri
Smack-Fu Master, in training
Registered: Mar 14, 2017
Posts: 11
Reply with quotePosted: Wed Mar 15, 2017 5:00 am
CommanderJameson: originally i thinked on releasing at april 1, so people would think that is a joke, and they being shocked to see that its actually real. this will actually happen, becouse i will register it to a lot of places on april 1.

shit, i almost forget my masterplan

[MK14]

shit, i almost forget my masterplan

Assuming that you here (Geri), and the 'Geri' over on the quoted/linked to bits, are the same 'Geri'.
Then I'm just repeating/quoting what YOU yourself actually said there.

If I'm misinterpreting it, then please feel free to explain ...

[Geri]

i really was forgot that i want to start a marketing campagin at april 1. i felt like days that i am actually forgotting some very important with dawn os, but it not jumped in, until i seen the quote. you are a hero.

[MK14]

I've tried it today, and think it is bad.
E.g. The calculator thinks 2^3 is some weird negative? number.

The project tells me to write a better Emulator, and produce the hardware for it.
(From the text file with it, attached below)

People, please do your own emulators and hardware to run Dawn.

OP/Geri PLEASE STOP.

[Wilksey]

We should create hardware for it, then patch the hardware to make the software work.   

[timb]

We should create hardware for it, then patch the hardware to make the software work.   

It would be funnier to patch the hardware to *prevent* his OS from working and, instead, only allow it to boot DOS, Windows and Linux.

That would really be a fart in his soup.

50GHz of single instruction imperialism.

[Geri]

E.g. The calculator thinks 2^3 is some weird negative? number.

added to buglist

[beenosam]

I've tried it today, and think it is bad.
E.g. The calculator thinks 2^3 is some weird negative? number.

The project tells me to write a better Emulator, and produce the hardware for it.
(From the text file with it, attached below)

People, please do your own emulators and hardware to run Dawn.

OP/Geri PLEASE STOP.

Well, 2^3 being a negative number is the actual true value. Only the creations of imperialists would ever return the value 8!

[MK14]

Well, 2^3 being a negative number is the actual true value. Only the creations of imperialists would ever return the value 8!

2^3 apparently gives the following answer:

[ataradov]

That's one funky calculator. How do I calculate "-12345 - 56789"?

[Geri]

-the bug mk14 pointed out is fixed

ataradov, that should work fine

2017, march 30. - Various bugfix
Toolset control in paint and text editor is improved.
Bugs fixed in pow, log, exp.
Bugs fixed in calculator.

[ataradov]

ataradov, that should work fine

How do I enter -12345, if there is no "+/-" button? Not that it really matters.

[Geri]

simply press - before you start entering the number

[CatalinaWOW]

As with any minimalist system there are work arounds.  I assume large numbers are entered something like 1.234 * 10^3.  Tangent by dividing sin by cos.  Key arrangement is apparently based on the typewriter concept.  (Which made the layout terrible to slow typists down enough to avoid jambing the typebars).  Here you are forced to look, avoiding silly errors from using muscle memory.  Could only be improved by randomizing the digit arrangement.

It is features like this that really make people flock to a concept.

[Wilksey]

Jeez, there's slow and then there's SLOW!
I don't have a slow machine, it's a decent i7, but it was *almost* as bad as using Windows ME.

I am curious, if one were to create such a beast that could run an entire single instruction to solve the world's problems, how would one get the thing to boot using the disk image(s)?

[Geri]

Wilksey: on the disk, the first 256 mbyte is the os itself, the rest is the file system.  os image goes to memory address 0, execution starts at 0.

[Wilksey]

What interface? SATA, IDE, USB Flash, SD?

[Geri]

hardware interface implementations are not part of the specifications in any way, but i suggest SD card. start - help - hardware.txt for more information.

[Monkeh]

hardware interface implementations are not part of the specifications in any way, but i suggest SD card. start - help - hardware.txt for more information.

You funny.

So much faster than x86! SD card for storage.

[MK14]

hardware interface implementations are not part of the specifications in any way

 

Well I'll accept that one, as it is April 1st, in some parts of the world.

[helius]

UHS-II supports data rates up to 312MB/s, and twice that in the next version of the specification. That puts it roughly on par with SATA and has the same queuing and reordering features.

[Monkeh]

UHS-II supports data rates up to 312MB/s, and twice that in the next version of the specification. That puts it roughly on par with SATA and has the same queuing and reordering features.

SATA is a slug by modern standards, and those UHS-II cards cost at least twice as much as substantially faster SATA SSDs.

[Geri]

sd cards seems to be the easyest, cheapest, and the most realible and most generic solution to deal with, i would not suggest to try doing sata aniway, becouse the disk handling system is not much fine tuned yet, it will not yet able to do beyond generic sd card speeds aniway.

[Geri]

2017, march 31. - Disk handler bugfix
Disk read/write stall glitch fixed when using fread/fwrite .
A bug fixed in the task handler.