If tablets are today's fad...

[Exophase]

I think today's desktop OSes hinder more than they help in resource usage vs libraries (for functionality). I also tend to think gaming (on both PCs and consoles) isn't as optimized as it used to be because of higher level languages and more capable hardware.

It's not the OS's fault, and running something "lighter" doesn't win you anything. It's the nature of abstraction and the APIs used, which are necessary in order to accommodate lots of different hardware configurations that are all different under the hood. Game developers aren't going to want to write to the metal for every type of GPU out there, especially for lower performance ones just to say they're doing it. This goes beyond vendor vs vendor, even generation to generation things change dramatically.


If you want fixed hardware you're basically just asking for another console.


And if you think that current gen console devs don't know how to optimize their code then you're sorely mistaken. There are probably more experts at it now than there were 20 years ago. Sure, there's less coding in ASM, but a lot of that wasn't really for performance reasons to begin with. I'm sure most SNES games were written in ASM, but some are not at all efficiently coding (am looking at you, Konami)

 

[Exophase]

The [Direct]XBox doesn't have to work through Microsoft's "Thickly abstracted " 3D API? That customized version of OpenGL the PS3 uses isn't a "Thickly abstracted " 3D API? Really?

Yes on both counts. Go ask John Carmack if you don't believe me: there's a reason why he says that PSV will have an abstraction advantage over phones despite likely running some permutation of OGL ES. NDS uses a "customized version of OpenGL" by some recknoning too, but you don't get the same assortment of penalties you get with typical OpenGL drivers on an abstracted platform.

Wing Commander 3 came out in 1994, requiring a 486 and a "graphics accelerator" while representing one of the stronger movements in early 3D. Note the Pentium showed up in 1993, and that not only was S3 founded in 1989, and had "graphical accelerators" out in 1991 even if they weren't really considered "3D accelerators" until 1998. The Playstation 1 came out the same year and would get the game in 1996. The Sega Saturn never got Wing Commander, but also came out in 1994 as a 32-bit system. The 3DO came out in 1993 and utterly flopped.

I said 3D acceleration, please recognize the distinction. VGA is considered "graphical acceleration" (look at its name) but its actual feature set is extremely limited, which shows that you can really call anything acceleration. It's not a change in nomenclature here, there's a big difference in what something like VGA does and what a triangle/quad rasterizer does. And in fact, Wing Commander 3 only requires VGA/SVGA. The capabilities are really a modest evolution of the CGA/EGA path and pale in comparison even to what 2D consoles did. I don't know what Pentium's release has to do with anything I said. The question is what was driving dedicated hardware for gaming, and I don't see how one can say PCs were over consoles when for the longest time PCs were the near antithesis of hardware acceleration.

From there we move on to Quake 1 and Quake 2, which represent the most influential 3D games and engines before the introduction of programmable shaders. The PS1's port of Quake 2 incidentally was noteworthy for not being up to the level of contemporary computers as "3D accelerators" instead of "graphical accelerators" came online and the OpenGL API took hold. So I'm not see how you're defending it drove things at that step along the way. The standard for programable shaders was basically set by IdTech4, and HDR lighting effects and physics by Half Life 2's Source and the less talked about CryEngine 1 so I'm going to tend to question the level of console involvement in the the development of the current generation of 3D that can fake lifelike reasonably well.

"Most influential" in what regard? Quake was a big move in software rendering techniques (which quickly became irrelevant). By that point, in 1996, there was no question that the industry embraced 3D gaming fully.. Whether or not shaders were driven by PC development is immaterial because shaders did not start the 3D revolution in gaming. This is a question of what made things take off, not later milestones. Note that I never said anything about current gen 3D being driven by consoles, clearly GPU tech influenced the 3D being done by GPU vendors. What I'm saying is that there's no reason to believe that it was the PC gaming scene that encouraged the development of 3D acceleration hardware instead of arcades, consoles, and big workstations.

This isn't helped by the fact the PS1's star game was in many respects Final Fantasy VII, which needless to say wasn't exactly what you'd call super advanced over Wing Commander in its 3D efforts.

Nor is it what anyone would call a stellar example of what the PS1 was capable of graphically, when compared to several other titles. Just because the game was popular (and daresay, good) doesn't mean that it somehow represents the graphical achievements of the time.

 

[sporkbox]

I think today's desktop OSes hinder more than they help in resource usage vs libraries (for functionality). I also tend to think gaming (on both PCs and consoles) isn't as optimized as it used to be because of higher level languages and more capable hardware.

It's not the OS's fault, and running something "lighter" doesn't win you anything. It's the nature of abstraction and the APIs used, which are necessary in order to accommodate lots of different hardware configurations that are all different under the hood. Game developers aren't going to want to write to the metal for every type of GPU out there, especially for lower performance ones just to say they're doing it. This goes beyond vendor vs vendor, even generation to generation things change dramatically.


If you want fixed hardware you're basically just asking for another console.


And if you think that current gen console devs don't know how to optimize their code then you're sorely mistaken. There are probably more experts at it now than there were 20 years ago. Sure, there's less coding in ASM, but a lot of that wasn't really for performance reasons to begin with. I'm sure most SNES games were written in ASM, but some are not at all efficiently coding (am looking at you, Konami)

Not the OS's fault that it takes up numerous gigs on the HDD (in Win7's case) and takes up more RAM idle than my Linux system does with a dozen apps running? It's simple math. For the sake of simplicity, if your system has 100 megs of RAM and the OS is taking up 20, 5-10 of which are important for gaming, then you're missing out on 10-15 megs that a game could be using, if not for the OS getting in the way. Then there's CPU cycles that are being spent on OS-stuff.


It's that issue of that forces PCs to make larger leaps in technology to get the same gains that consoles do every 5-7 years. The games have to compete with the OS (and any other apps currently running). Dedicated consoles don't have that problem. That's why I think it'd be great to have something like a dedicated game mode on Windows or something, where only the necessary libraries are loaded. It would open up some resources so games can bring out the true potential of the hardware. I'm positive it can be done. There are obviously technical hurdles to overcome to make it happen, but I don't think software would be where it is today if those hurdles weren't tackled.


If you're gonna try to make an appeal to pragmatism and say it's not worth the effort, I'd like to know what sort of proof is out there to support it. To my knowledge, no OS vendor with major games on it has attempted a lightweight "gaming" mode.

 

[Exophase]

Not the OS's fault that it takes up numerous gigs on the HDD (in Win7's case) and takes up more RAM idle than my Linux system does with a dozen apps running? It's simple math. For the sake of simplicity, if your system has 100 megs of RAM and the OS is taking up 20, 5-10 of which are important for gaming, then you're missing out on 10-15 megs that a game could be using, if not for the OS getting in the way. Then there's CPU cycles that are being spent on OS-stuff.

This comparison isn't fair until you can determine what of that claimed memory is not recoverable by programs that need it. The HDD space is barely applicable. The belief that the OS spends a bunch of CPU time in the background is a myth, the actual amount of time spent in kernel is vanishingly small.

It's that issue of that forces PCs to make larger leaps in technology to get the same gains that consoles do every 5-7 years. The games have to compete with the OS (and any other apps currently running). Dedicated consoles don't have that problem. That's why I think it'd be great to have something like a dedicated game mode on Windows or something, where only the necessary libraries are loaded. It would open up some resources so games can bring out the true potential of the hardware. I'm positive it can be done. There are obviously technical hurdles to overcome to make it happen, but I don't think software would be where it is today if those hurdles weren't tackled.

If you're gonna try to make an appeal to pragmatism and say it's not worth the effort, I'd like to know what sort of proof is out there to support it. To my knowledge, no OS vendor with major games on it has attempted a lightweight "gaming" mode.

So you're going to be ignoring what I said about abstraction being the issue? Got it. Well, believe what you want. Over the years I've seen tons of people clamor for PCs with no OSes or light OSes in order to run things like games or emulators faster. I think the burden of proof of usefulness lies more on the people making the claims that it is, because I'm not aware of any evidence to this case. Lots of extremely lightweight OSes exist for PCs - hobbyists make them all the time - and they've never offered any kind of performance advantage. But I suppose it's difficult to make a fair comparison when none of these can offer support for modern hardware and have to fall back on legacy hardware interfaces like VESA.


Probably the reason OS vendors aren't shipping with "lighter" versions is precisely because there's no benefit to have. Do you think MS, with its stake in gaming performance via DirectX, wouldn't be receptive to a "gaming mode" that could improve performance?

 

[EvilDragon]

Most of the time, it's not the bare OS that slows down stuff.


It's the ton of background apps a user installs, like virus scanning tools (which REALLY slow down the system a lot), Adobe Updater, and so on.


So yes, a dedicated gaming mode MIGHT help - a mode that switches off LAN / WLAN and doesn't load any of the background apps.


It would help a bit. However, there are A LOT of LAN games out there, therefore, you need to keep the virus scanner, etc.

 

[Batou456]

Exophase, you claimed that:


-3D gaming was driven by arcades and home consoles with the home/personal computer market having questionable influence.


-3D first went to 5th generation home consoles while making a point to emphasize the utter commercial failures that was the 3DO and final Atari console, and hence kind of can't be legitimately flagged as driving the development of anything.


-That PC 3D hardware lagged significantly behind that of consoles, and surpassing it is a recent novel concept.


I was nice and mainly emphasized Wing Commander 3 in 1994 running off a 486 with graphical accelerator.  Hardware that was rather old by modern standards of computer development at the time of its release.  To remind you where things stood the Intel Pentium came out in 1993, Motorola 68060 in 1994, the PowerPc in 1992, and the DEC Alpha in 1992.  The SNES came in from 1990-1992 depending on your particular region with everyone but Japan getting it after the first half of 1991, meaning in terms of IBM PC compatible development the SNES is directly competing with the kind of computers Wing Commander 3 was aimed at.


But let's go even earlier, given you made the point to drag it that way.  Atari's Battlezone is from 1980 where it debuted with Arcade machines, and was arguably the first 3D game.  The Commodore PET, Apple II, and Atari 2600 entered the market in 1977.  Commodore however was releasing the VIC-20 around the time Battlezone came out, and would very rapidly replace that with the Commodore 64 in 1982.


So let's do this.  The original Arcade Battlezone verse the

https://www.youtube.com/embed/NSAO_08-vI8?feature=oembed which literally looks like a 8-bit game.  That is effectively what you're saying when you stressed how wonderful the SNES was combined with your hardware claims, unless you want to bring forward some stellar 3D game from the SNES catalogue that even approaches the 3D sophistication of WC3, or WC4 for that matter given the CPU requirements didn't change.  Of course, you're not exactly rushing to pull out a Playstation 1 game that beat the 3D sophistication of SNES era computer hardware.

Although really, Final Fantasy VII was released in late 1997 and was at the time was considered a rather sophisticated 3D game so I'm not sure why you're whining about it.  Sure you can appeal to something like Final Fantasy VIII from 1999 having better graphics, but you don't get some kind of magic get out of jail free card.  Wing Commander 3 for DOS wasn't exactly optimized for the 486 or a specific graphics card to even the degree of OS/2 and  if you're going to obsess on that the question is whether it is better to statically optimize or move to newer hardware.  Given Quake 3 with it's early shader system would be the title going up against Final Fantasy VIII, that sticks to the conventional rule that's not in favor of the fixed system approach and why you're stuck relying on sophistry.  Aka in the time it takes you to optimize the software to a fixed set of hardware hardware, newer better hardware will have come out that invalidates the effort making doing so a wasteful approach.  That's without considering the resultant absolute lack of mobility leading to requiring sophisticate emulators that consume an order of magnitude more processing power to attempt backwards compatibility.


But let's continue with following the curve from 8 bit vector graphics that tended to have single color textures and talk about the 2.5D sprite based 3D engines.  Catacomb Abyss was released in 1991, Id software's Wolfenstein 3D in 1992, and Wing Commander 1 in 1990 capable of all being handled by a 80286.  This in turn led to more sophisticated games like Betrayal at Krondor, Descent, X-Wing and DooM all designed to be handled by the 80386 line, and we've already covered the 80486's poster child, Wing Commander 3 & 4.  Id Software however had a very special role to play in all this.  You see they had this revolutionary concept of selling their engine for other people to use, and then when it was obsolete Open Sourcing the engines.  Because of that and Quake 1 & 2's codebase contributions, Id's early entries were both very influential and relevant to the development of 3D engines.  Stuff with Id code encompasses quite a bit, and they lost ground primarily because other Engine developers took on a similar approach of licensing their engines.  That you would brush them off as of minor consequence is telling.


The final nail in this coffin is when we take this to modern terms CryEngine 2 released in 2007, a mere 1-2 years after the current consoles were released, is one of the most advanced engines developed furthering AI programming, physics, lighting, and other sophisticated computational issues related to 3D engines.  Instead of CryEngine 2 becoming the new gold standard consoles resulted in EA basically neutering it to the level of Source so it could be forced to run on lobotomized PowerPC hardware.  That's not what you call helping development, Exophase.

 

[Exophase]

If that's what you got from it then let me revise it to say what I really meant - 3D hardware was driven primarily by arcades then consoles, not PCs. There were a lot of 3D games with software rendering on PCs and other computers for a long time, but the important distinction is that the PC hardware itself didn't drive these developments. Regardless of what you think Wing Commander 3 did with hardware it didn't actually utilize any meaningful graphical acceleration on a PC. Maybe this is too subtle a point, I don't really know, but there's a huge difference between what drives hardware and what drives software, and PC hardware was most definitely not driven by games until well after console hardware was.


I don't think 3DO and Jaguar were necessarily failures per se, they just got quickly supplanted by bigger and better consoles from bigger and better companies. And they weren't the first offerings.. well regardless of what you think about Starfox looking like an "8-bit game", but it's true that this is the same kind of 3D you get when you don't have dedicated hardware for it. PCs of the time were better suited than the kind of ridiculous SuperFX processor, but getting high end processors on their gaming machines wasn't Nintendo's strong suit at the time.


Also good to see you're back to your more ridiculously loaded argument style, I was starting to miss the old Jebe ;p Have you ever thought about going into marketing? And I'd like to know more about this "early shader system" Quake 3 had..

 

[TrashyMG]

And I'd like to know more about this "early shader system" Quake 3 had..

Quake 3 Shader Manual

 

[WizardStan]

Not the OS's fault that it takes up numerous gigs on the HDD (in Win7's case) and takes up more RAM idle than my Linux system does with a dozen apps running? It's simple math. For the sake of simplicity, if your system has 100 megs of RAM and the OS is taking up 20, 5-10 of which are important for gaming, then you're missing out on 10-15 megs that a game could be using, if not for the OS getting in the way. Then there's CPU cycles that are being spent on OS-stuff.

There's two assumptions you're making, myths that keep getting repeated.


First, you seem to be arguing that if there was more RAM that the application would somehow use this to do whatever it does better. Unless the program has been designed to use more RAM as it becomes available to improve its task (and most aren't, they just request RAM as they need it) then it doesn't matter how much free RAM you have, the application won't care: if it runs fine when you've only got 80MB free, it'll run basically the exact same as when you've got 100MB, or 1000GB free. When you start running applications (especially multiple applications) that butt up against the extents of RAM and start into the swap file then you may have an argument, but my counter to that is that I haven't really used a swap file in years now, even on my Vista laptop.


Second, you assume that the only usable RAM is free RAM. A lot of OS's will use RAM as cache for files or other entities which might get reused often. To the observer, this is used RAM, cutting into the free stuff that applications supposedly have access to, but an OS which does this will intelligently (and it's actually kind of hard to do it stupidly accidentally) dump this cache as it notices RAM becoming scarce.

 

[Exophase]

Quake 3 Shader Manual

These "shaders" are by and large way more high level than and not much of a direct analog with what we call GPU shaders today..