Slashdot Mirror
Xbox 2 SDK Released On Mac G5?
Espectr0 writes "The Inquirer reports rumors, courtesy anonymous sources, that Microsoft has released the Xbox 2 SDK to select videogame developers, and they are using 'dual Apple Power Mac G5 systems running a custom Windows NT Kernel.' This ties up with earlier rumors which mention that the XBox 2 will be powered by the IBM chip, and ATI will be providing the video chip." The report also notes: "Interestingly, the SDK apparently also features an Apple logo on a side bar within the application."
Several years back, Microsoft wrote a version of Windows NT for PowerPC chips.
Here's a link.
Paizurishitetai desu ka?
There is very little other than the processor and BIOS that is different about a Mac and a PC. They use the same memory, same system busses, same architecture, same periphals. The periphals are, in most cases, totally interchangeable, needing only appropriate OS drivers.
Also, NT has been on PPC. NT 4.0 was available for x86, Alpha and PPC architectures. However the target of NT 4 PPC was IBM systems, not Macs. Neither it nor the Alpha version did well, so they were discontinued with Windows 2000.
The NT architecture is highly portable and was designed that way from the beginning. Porting it to a new chip is not a major feat. For something like the X-box, even easier since it runs a much stripped down version of Windows.
Ummmm, the 64-bit IA-64 version of Windows XP has been available for a long time. Full release version, not a demo.
About six or seven years ago, MS bought a small amount of Apple preferred stock (which means non-voting shares). It was intended as a public show of confidence in Apple's future. Apple's stock price went up quite a bit in the time after that, so MS sold the stock for a nice profit. MS never had ANY control over Apple, contrary to what people mistakenly repeat.
No I mean like the IA-64 version was out years before the x86-64 version. It came out a few months after the Itainum 1. It's pretty much been a non-story, given the Itainum's less than stunning success, but it's been around.
Will Darwin or Linux be ported to Xbox first?
Just a tidbit of info. My version of Safari seems to be v1.2. IE is still included with the freshest build of Panther, but it is no longer the default browser.
No one's turned the GameCube into a Mac.
Nope. Because the GameCube has very little in common with a PC. It has a G3 CPU, then a massive all in one control chip that contains the ArtX designed GPU and other components.
The XBox on the other hand was very much like a computer.
Let's see here:
For Nintendo:
SNES: Success, not backward compatible with NES.
N64: Success, not backward compatible with SNES or NES.
Gamecube: Success, not backward compatible with N64, SNES, or NES.
For Sega:
Genesis: Success, not backward compatible with SMS.
Saturn: Failure, because of lack of 3d. The Saturn was designed to be the ultimate 2d console, which it was. 3d was an afterthought, and never worked well.
For Sony:
Playstation: Success, not compatible with ANY other system, being their first.
So it looks to me like backwards compatibility is a nice feature, nothing more. Nintendo, being the oldest company, is the most shining example. NONE of their consoles have been compatible. They thought about it with the SNES, hence the use of the 65C816 (which has a compatibility mode for 6502 code) but didn't end up doing it. The rest of the consoles aren't even remotely compatible. None the less, each has been a success.
What makes a console succede? Two main things:
1) Having flashy graphics to attract people, and the marketing to let people know about them. You may not, but most people gravitate towards pretty graphics.
2) More importantly: Good games. This is REALLY what makes or breaks a console. If your system has the games people want to play, they'll buy it. If it doesn't they won't. This is also a positive feedback loop since the more good games you have the more you sell and the more consolse you sell the more developers that will want to release for your console.
That's why Sony succeded, despite being new to the market. They released the games people wanted to play, and had stunning graphics for the time. Combine that with good marketing, you've got a winner.
As of late IBM's chip division has kind been seen only as a high-end server/supercomputer thing. Thier midrange market is almost non existant and their embedded market has been shrinking. Well the 970 is a serious midrange contender and I'm sure they want people to know it.
"midrange" is a very subjective term. IBM has been making some of the G3s for Apple for some time. However, it is true that IBM processors generally go into their servers (This isn't limited to just "high-end" and supercomputers). You can buy a IBM p615 with a real POWER 4+ processor for a relatively affordable (depending on what you consider relatively affordable) price of $5745. But I do agree that their consumer processor business as well as embedded market has been shrinking.
but they have turned the gamecube into a linux machine.
This reminds me of when the Apple IIgs was used as a development platform for the Super Nintendo, since they both used Motorola's 65C816 chip.
:-)
If the Inquirer's report is true, maybe Apple should take a hint and revive the Pippin program. Others seem to think Apples are good for game programming.
Actually considering that IBM is manufacturing/assisting with the design of CELL for Playstation 3 - designed & built the Gecko processor for the Gamecube - and now finds its way into Xbox 2 would indicate that:
"Well the 970 is a serious midrange contender and I'm sure they want people to know it. Being the chip in one of the big 3 consoles certianly goes a long way for that."
IBM doesn't just have 'a chip' in one of the big three consoles - it is making all the chips in all the consoles.
As well process innovations in fabrication will allow IBM to reduce the cost of production on the 970s over the lifespan of the X-Box 2 - in addition - economies of scale should kick in to a large degree considering the scale of the resource commitment for the cpu's in a console.
No, long ago AMD had a cross-IP licensing agreement with Intel, because AMD got a big start helping Intel supply x86 chips. There was a court case brought by Intel against ADM when AMD started making their own compatible chips, which AMD won, around '486 times, that enabled AMD to continue to use said IP up to a certain tech level, essentially in perpetuity. So, no, AMD does not have to pay licensing fees to Intel for some of the X86 stuff that they use in their chips.
Otherwise, there would be no point in AMD making chips.
Which is why AMD has a different set of X86 extensions to do mostly what MMX does compared to Intel, which was some of the first new tech on X86 chips that AMD could not duplicate (terms of the lawsuit judgement).
Intel tried to first kill AMD with MMX, but that did not really work as well as Intel wanted it to, as Microsoft kept their options open and also coded Windows 9x to use AMD's multimedia extensions as well.
At that time, while 99.5% compatibility between K5 and K6 chips with Intel-equivalents was there (which should have been good enough for everyone), there was quite an anti-AMD bias amongst those who bought computers in large quantities, because the occaisional problems with K5/K6-based computers at the time wasn't necessarily the CPU's fault, but crappy drivers and bad motherboards.
As it stands, how many companies (esp. Fortune 500) have tried to save a few thousand dollars on their big computer upgrade cycles by getting AMD-based instead of Intel-based computers for their Windows users? Not too many.
Yet somehow AMD has managed to find a niche in the computer market, and enough collective mindshare to survive to the point that now Intel is reacting to AMD's moves, not the other way around.
I know that the next x86-based computer I get/build will probably be AMD-64 based.
As much as I want one, I can't quite justify getting a new Macintosh. I don't do Photoshop, and too many of the things I do/like do not provide enough options for Macs.
Heh, you obviously didn't follow my link *to the giri giri FAQ*. Look under headings like "Why is SSF / Giri Giri so slow?" and "Why doesn't giri giri play audio?" to understand why I stated that there was still no decent Saturn emu. Though this is by far the best out there, and hats of to them!
Wouldn't XP be Windows NT 6.0? I though Win2k was 5.0
Windows 2K is NT 5.0 and Windows XP is NT 5.1. Server 2003 is NT 5.2. These are the official version strings from Microsoft.
There is very little under-the-hood change between Windows 2000 (NT 5.0) and Windows XP (NT 5.1) aside from GUI modifications. From an application's point of view (or even a power user who makes all of his settings via the registry) there is not much different between the two. A lot of little things have been refined and updated, yes, but nothing huge.
And really... Win2K (NT 5.0) isn't a whole lot more than NT 4.0 SP4/SP5 + modern version of Direct X + modern drivers + light GUI polish. When NT 4 first shipped it did not come with Internet Explorer... later versions included a standalone version of IE... and still later versions included the deeply-rooted IE that we know and hate today.
NT 3.x was pretty archaic *looking* (Windows 3.x GUI, ugh!!) but still had most of the guts that NT 4 later used....
Call it what you will, but NT was the best thing Microsoft has ever done. We could all be using a heavily patched version of Win95 running atop DOS 7. "Win98 Seventh Edition!"
Get a Family Pack. Only $199 for up to five licenses.
Actually, the new G5 tower has a completely different system architecture which uses HyperTransport technology co-designed with AMD. Each processor has it's own point to point bus to the system controller.
http://www.apple.com/powermac/architecture.html
I believe that only AMD Opteron systems have a similar architecture.
While both PCs and Macs support the PCI, USB and Firewire standards, the connections between the chips that implement these standards is quite different.
not out of the question, but NVIDIA is probably going to make damn sure that ATI doesn't get details about the Xbox graphics implmentation.
Splinter Cell is one game that uses NVIDIA proprietary stuff. If you don't use an NVIDIA >= GeForce3 (excluding the 4 MX series), you'll notice that the shadows in the PC version will run differently to the Xbox one, even with a Radeon 97/800 something.
Atomic MPC ran an article on how Splinter Cell was ported across four platforms. Very interesting read. The PlayStation 2 ended up being the worst of the lot.
> everyone knows Xbox and PS2 and all those game systems are sold at a loss.
no everyone doesn't know that.
some informed people repeat it mindlessly, but otherwise everyone else ignores it.
The last console manufacturer to sell the consoles at a loss apart from Microsoft was Sega.
remember Sega?
Er, not really. the API is a construct of the software, not the hardware. It would probably be childs play to get an NT kernel with DirectX running on PPC, all you'd have to do is recompile the app to the new system (after some endian and minor tweaking).
The whole point of an API is that you have an abstraction AWAY from hardware to either make coding easier or code more portable.
"Sometimes, I think Trent just needs a cup of hot chocolate and a blankie." -Tori Amos on Nine Inch Nails
Components are similar, but the architecture (beyond general concepts) is not.
The current Dual-2GHz G5 utilises a 1 GHz bus, with no contention between the CPU's. The G5's use a 0.5xCPU-clock bus, so the upcoming 3 GHz machines will be using a 1.5GHz bus.
An overview is here
Anyone who claims Apple have no microprocessor design expertise needs to look at the system controller. It's an all-Apple part, and it's beautiful. Dedicated buses for each CPU, AGP-8X, PCI-X, Serial-ATA, 400MHz RAM interface...
The PPC970's are slick, but it's the architecture that really sets the G5 apart.
Cecil,
The universe is always broken in this example, as there's nothing there to overwrite pChars[0]. pInts[1] == pChars[0] + sizeof(int). Since on no platform is zero the size of an int, pChars[0] (which was memset back to 0) will always be zero.
Endianness refers to the internal representation of the numbers, i.e. what's happening in pChars[3] through pChars[7]. The ints themselves still advance normally.
--Dan
www.doxpara.com