OpenGL Game OS

A gameoriented OS doesn’t need to be built from scratch.We could smiply take the linux kernel,remove all the useless drivers(like HAM for instance) fiddle with the scheduling to make it more generous towards certain processes(the games of course) with priorities(allready implemented),bigger quanta,etc. ,edit the initscripts so that only the absolutely necessary deamons,etc. are loaded and there you have it.Oh yeah you’ll have to find a cool name too.That can be a little harder.
But I don’t see how that would help us to make a revolution in the computer games world.Maybe it would be better to create some sort of organization,get a few coders/artists who want to create games and do it.We’d have to find some way of bypassing all those publishing companies and actually sell the games at a nice price.But then again coordination and management might prve difficult…

To LaBasX2 and zen, and eyeryone else out there.
Yes a whole OS dedicated to games are surely a thing for the future. Then we create a special “license” GPL sort of. Not allowing any overpriced games to run. But would anyone build games for such OS? If you gain speed and fps… with out a doubt. The problem is that I’m a game developer, I do the design of the game, story , characters , stuff like that. Sure I do know how to write c++, and I know what an bitch ASM could be. But that’s far from enough to write a whole OS. What’s needed are professional OpenSource programmers with knowledge enough to write this OS. Me and my team of dedicated developers would without delays write a version of our current RPG for that OS. And yes we have made a deal with the publisher so our game “Sion”, will be sold for a price around 20$. Let’s bring the large companies on there knees…

Remember Indrema? One of their goals was to create a Linux based gaming OS. I think they released an SDK at some time.

Originally posted by uzu_manga:
Then we create a special “license” GPL sort of. Not allowing any overpriced games to run. <SNIP>… Let’s bring the large companies on there knees…

The success of any gaming platform/OS depends on having those “large companies” porting their “overpriced games” to that platform.

Then we create a special “license” GPL sort of. Not allowing any overpriced games to run.

Or we could create a hardware chip that only allows cheap “trusted” games to be run…

How is that any different that what you are combating? Every one wants to be in control, but no one wants to be controlled…

Originally posted by rapso:
[b]“That’s what the article was about… no free software and no freedom… don’t worry the US government will never allow anything that infringes so heavily on personal privacy and capitalism to slide…”

if I were you, I wouldn’t be that sure. the US Goverment has a backdoor to all PCs with TCPA, the USA will even controll the Internet, why shouldn’t they support TCPA?
[/b]

See knackered’s point…

What scheduling overhead? All the os processes running on a Windows or Linux machine are idle almost 100 percent of the time. This means that they use 0 cpu! If a game wants 100 percent of the CPU, then it pretty much has it 100 percent of the time and only ever lets it go if there is another process like it which is actually running (why would you run SETI@Home while playing a game?) or if it makes a system call (which actually makes the system run faster because the CPU can do something else while waiting for IO).

Please learn how OS scheduling works. People who do not understand it are the ones who do things like write SETI@Home like apps (apps which use lots of CPU) and set the priority to the highest it can go. These people do not understand that even set at the lowest priority, a SETI@Home will get 100 percent of the CPU if it is the only program not idle. So, all setting it to highest priority does is make everything ELSE on the system run poorly.

Windows is the best gaming OS on PCs because it has not only the technology, but the actual market of people who go to Babbages to buy games. Who cares about performance when no one even hears about your game?

You people really need to read up on the TCPA! It is not designed to protect the system by keeping programs from running, it does so by protecting programs from each other. It is more like a firewall which keeps other programs from messing with your data if you choose to, including the OS. Any programmer can take advantage of it to create protected information which only the program can access.

A person with any thought process at all in their heads will realize that the Windows platform is what it is because there are millions (MILLIONS!) of programs written for it by everyone and his dog. Why would anyone suddenly decide that it is time for that to end? (HINT: It is not!). Not only is it stupid, but it is a blatant violation of antitrust law.

I am not saying so much about the TCPA as I am about the insane amount of misinformation in this thread. Go find some actual essays on the TCPA, especially a technical one that actually explains how it works.

Please do not let a love of paranoia and conspiracy make you start believing that people are going to do something as blantantly moronic as try to sell a PC that will not run any of the millions of existing programs already made available in the past 20 years since the introduction of the PC.

Think!

Nakoruru,

I think the concerns (at least for me) are the “trusted” issues. The hardware can prevent software from running and guess what you’re not in control of the hardware. The hardware (supposedly) goes out to see if you’re allowed to run the software and to see if you’re allowed to open a particular file. I’m not into the government, Microsoft, or anyone invading my privacy, if I’m downloading porn and writing a book that’s my business, no one else’s. The biggest issue is that I read that (don’t know if its true) this magical device will delete files that you’re not supposed to have. Ok, I decide what files I’m allowed to have, if the FBI wants to waste the money arresting me for an unreleased Sneaker Pimps mp3 while some whack job is picking people off in DC, fine, but I’ll be d*mned (don’t know if this board blocks profanity) if some jack *ss at the record company is going to do it…

John.

I don’t care about warez meltdown. Or the end of mp3 sharing. These are non-issue to me. But that’s not what TCPA is about.

And to the “the american government surely wouldn’t allow such a thing” people: there is such a thing as the DMCA, remember? And the TCPA/Palladium thing is also already signed. See Micros~1 press releases.

And to the “let’s make our own OS” people: Non-certified OSes won’t even boot. Period.

You’re all a little late in the game.

This is the beginning of the end.

Originally posted by zeckensack:
[b]Non-certified OSes won’t even boot. Period.

You’re all a little late in the game.

This is the beginning of the end.[/b]

Sorry I didn’t realize that Sun, SGI, and AMD all signed up as well… oh, wait they didn’t…

AMD did.

[edit]
TCPA homepage: http://www.trustedcomputing.org/

Partner list (PDF): http://www.trustedcomputing.org/docs/TCPA111999REL.pdf

[edit2]
And I don’t see how SGI/Sun would matter. Sun doesn’t make x86 hardware themselves. But Solaris/x86 would still have to be certified to run (if it is even still maintained).

SGI is just like Dell. You buy/resell x86 hardware, you have TCPA.

Keep in mind that this Palladium stuff redefines the x86 PC platform to incorporate this tripe in hardware. There will be no way around OS/software certification.

[This message has been edited by zeckensack (edited 10-23-2002).]

[This message has been edited by zeckensack (edited 10-23-2002).]

> […] There will be no way around OS/software certification.

hmm… judgeing by the record of the companies involved, I’m not afraid of there being no way around, I’m rather afraid of proggies using a way around (willingly, or through bugs) and then, f.i., sabotaging your certificates. The manufacturer return would be the only solution to get it running again…

None, none, of the consumer, government or big-company sponsored security technologies has proven “secure”, from DVDs to Playstations, from Clipper chips to Bluetooth, or WinXP to Oracle.
As soon as they were announced “secure”, they were breached (if not before) and a never ending stream of flaws uncovered.

Finally, security built into the hardware can’t be patched.

Since when is this about security? I tell you it’s not. That’s what uncle Bill will try to tell you but so far he has failed to bring up a valid example.

This is a system designed specifically to the purpose of preventing code from running, on a per-application certificate basis. How can this improve security? Say, Outlook gets certified and still stupidly executes every vb script it can get its hands on. Your freeware virus/worm scanner won’t run anymore, because certification costs money. Did security just increase? IMO no. The opposite is true.

Document Encryption/decryption is mentioned in the context but really doesn’t require TCPA at all. Neither do secure network sessions.

All that TCPA has to offer is more cost and less choice for users. And it’s already too late to unmake it. Cheers.

Zeckensack,

Ok, I read the faqs and models at: http://www.trustedcomputing.org/

I would be a hypocrite if I didn’t… Anyway the docs were only about security, online transactions,… how are they going to stop software from running, I must be missing something…

Nakoruru:it’s funny how you can deduct that someone doesn’t know about multitasking/timesharing(or OS in general) from a few lines…Anyway besides the fact that there’s not only the overhead of other processes running(wich might be needed in a game OS too) but also of the time spent in kernelspace every n ms(usually n=10ms) and of the context switches wich are needed every time the to-be-run process changes.But I do not disagree with you that all this might be negligible,that is I don’t disagree because I haven’t looked into it yet but if you would design a gamespecific OS,which I didn’t propose and which purpose wouldn’t be eliminating overhead(just read previous posts),well if you would design such an OS wouldn’t you tune the scheduling with respect to its specific goal?Would you just implement general purpose scheduling.I just said that an OS can be made from a tailored Linux kernel and that I don’t think it would be of much help.
As for the GNU-like license,I also disagree.I don’t see much freedom in forcing someone to be gnu-compatible in order to use your ‘free’ product.
And regarding TCPA,a CPU(the computer actaully) is a tool,it’s purpose is to execute specific instructions wich the tool user provides.Anything less than that,no matter how ‘well ment’(!) in order to ‘protect me’(again !),is not acceptable (by me at least).

ho, ho, ho. all i can say is, nothing truly crap can ever survive, cos if people don’t want something, it’ll die away. and if they don’t care enough, they deserve it. maybe at some point in the future this sad little species will learn its own pitfalls and fix them - honestly, of what benefit to this society can consolidated monolithic companies be?

a bit deep, but hey, i’m just tired of seeing big bastards do anything to get bigger anyone remember ZX Spectrum’s intro tape, “The Horizons?” i’m thinking foxes and rabbits

[do note the capitals ]

Responding to the first post, I would have to say that a “gaming OS” is a silly idea.

Originally posted by uzu_manga
Outperforming both Windows and Linux in framerate… Then implementing a “directx” look alike media library and opengl.

Exactly how would this new OS outperform Windows or Linux? The multitasking overhead may have been significant on a 486, but it is not on today’s processors. The OS doesn’t determine the speed of a game; it is the hardware. So what you would have is an OS that has no advantages over any current OS and is good for nothing but playing games. Great idea.

Originally posted by john_at_kbs_is:
[b]Zeckensack,

Ok, I read the faqs and models at: http://www.trustedcomputing.org/

I would be a hypocrite if I didn’t… Anyway the docs were only about security, online transactions,… how are they going to stop software from running, I must be missing something…[/b]

trustedcomputing.org of course concentrates on a friendlier view. The real issues start with the TCPA implementations, spearheaded by Microsoft (“Palladium”) and Intel (“Fritz chip”).

Maybe this will prove useful: http://www.cl.cam.ac.uk/~rja14/tcpa-faq.html

Fritz is a hardware dongle on the motherboard, which will do what I described: check certificates for boot rom and all software and deny execution. In the next phase this will move into the CPU itself.

Be afraid.

Zeckensack and all,
please correct me if I’m wrong,I’m not sure how accurate that FAQ really is, but I’ve done some reading on slashdot and I tend to agree with Nakoruru’s second point that there seems to be some misinformation reagrding TCPA,rather TCPA/Palladium.The actual purpose of TCPA seems to be protecting ‘content companies’ and their money.That is achieved(?) by implementing a safe(?) mechanism for anyone to distribute encrypted/protected media to ‘safe’ clients.Regardless of the actual mechanism to achieve this and according to M$ you can still generate content yourself and distribute it freely and anyone can use it but if, for example, you code an audio player you can’t expect it to play Palladium-protected media unless you get it certified.And then of course whenever you do a bugfix and recompile it you’ll have to recertify it.Neat huh?But I doesn’t seem to keep you from writing software,not even an OS although this OS won’t be certified and ‘nice,big,lawful content companies’ won’t like it,err trust it.If this is all correct then nothing stops me from recording a song and distributing it freely but if I (legaly)own a protected song then I can’t rip it to play it on my (non-palladium)stereo.Of course this all can be cracked on the HW level(not too hard from what I’ve read) and this’ll propably end in another copy-protection fiasco.
Yet that doesn’t seem to make it less dangerous.Something like the TCPA you’ve been describing here couldn’t be enforced,err… I mean,offered to the people right away you’ll have to do that in steps so that they can get used to it over time.
Besides did anybody ask us if we want to have such unneeded(and worse unwanted) complexity in the CPU and chipset and pay for it?Bah, the x86 architecture sucks badly anyway and since sparcs and alphas costs,these VIA CPUs are starting to look much better…

BTW FAQ #18:

  1. Does code, applets or drivers used on a TCPA-enabled system need to be
    signed to run?
    No. The use of signed components depends upon the operating system environment in
    which the Subsystem operates.

Last time I checked ‘no’ was quite different from ‘depends’…

Originally posted by Aaron:

The multitasking overhead may have been significant on a 486, but it is not on today’s processors. The OS doesn’t determine the speed of a game; it is the hardware. So what you would have is an OS that has no advantages over any current OS and is good for nothing but playing games. Great idea.

There is always some overhead in multitasking, specially on a single cpu. In an ideal situation, each process, each thread would run on it’s own CPU, each cpu with there own cache and perhaps even memory.

Seriallization good, parallel better.

A tuned up OS might run better, but the effort would be a waste.

PS: I think all you guys are nuts. Except for Nutty

V-man

What’s all this about multiple CPUs? I thought we were talking about an OS not a hardware platform.

There is always some overhead in multitasking, specially on a single cpu. In an ideal situation, each process, each thread would run on it’s own CPU, each cpu with there own cache and perhaps even memory.
The overhead increases with the number of CPUs. If several CPUs share the same memory but have their own caches, you have cache syncronization overhead. If they have their own memory, you cannot directly access data from another thread running on another CPU.

My point was that if you could run Windows95 (a preemptive multitasking OS) on a 486 without problems, then the overhead will be extremely small on a 1.5 GHz processor. If the overhead were, say, 10% of the total CPU time on a 486, then the overhead on a processor that is 10 times as fast would be about 1%. Do you think you would be able to perceive the 1% increase in frame rate if this overhead were eliminated entirely? Windows and Linux may not be the most efficient OSes possible, but the speed enhancements of a “gaming OS” wouldn’t be noticable. Sure it would be nice if installing a new OS would make Quake 3 run at twice the framerate, but that isn’t going to happen.

And one thing I don’t think has been mentioned here before is that in order to play games made for the proposed OS, you would first have to install it. How many game players (not game developers) would install a new OS on their computer just to play a game? I get annoyed when a game says I have to update my drivers . Games released for this OS would have a very narrow market.