Putting a 1.48GHz Tualatin CPU in an Xbox

An anonymous reader writes "A stock Microsoft Xbox has a 733 MHz Intel CPU with 128KB L2 cache. On Valentine's Day, Friendtech will launch the DreamX-1480, a modified Xbox with a 1.48 GHz Tualatin-core CPU with 256KB L2 cache, promising better framerates and more stable network gaming. FiringSquad has the review."

Comments

Re:so whats the deal.

[hawkstone]

Here are correct links:

FriendTech
The FriendTech DreamX XBox project

Isn't this kinda pointless?

[PedanticSpellingTrol]

If all of the games are designed for the stock hardware, is upgrading really going to make the slightest difference? Who cares about the framerate anyway? NTSC is only 60Hz interlaced refresh. Also, I'm pretty sure I heard about a company in Hong Kong that was already doing this.

Re:Isn't this kinda pointless?

[webscathe]

but i guess the cheapest and easiest way to get the 1.48ghz

Cheapest?? The article said this thing is going to be $500. That aside though, I can't see why anyone would actually want this. It increases the frame rate for a couple of games but because MPEG-2 playback is dependent on the 733 MHz clock the article says that most in-game movies will pause every 3-5 seconds, and most DVD play back has similar problems. Granted, it has a switch on the front to knock it back down to 733 but the article also said that mostly didn't work for video playback.

I applaud the folks at Friendtech for this hack, and their mod chip for the system sounds like it has some pretty cool features for turning the xbox into a media center, but I can't imagine it'd be worth blowing $500 for a partially functional xbox.

Hmmm

[Anonymous Coward]

Will this void your X-Box warranty?

Re:so

[AKnightCowboy]

so when are we going to see some consoles with 1MB level1 cache? that would totally improve performance, more than upping processor speeds

And put in a better processor! How about an AMD Athlon 64 FX-51? Up the paltry 10 gig drive to a 72GB WD SATA Raptor... granted all this is gonna get pretty hot so it'd need a bigger case and some more fans. Oh, let's not forget a good modern video card.. maybe an NVidia GeForce FX 5900. Get rid of the goofy controller though while we're at it and repace it with a nice keyboard and trackball. Also, it should run the full version of Windows XP Pro. I wish someone would make a version of the X-Box with those specs. I'd buy it in a heartbeat!

Re:so

[philthedrill]

You won't see 1 MB L1 caches, at least not in the forseeable future. The reason there is a memory hierarchy is to reduce access time due to address decoding and (especially) wire delays while keeping costs low since cache is expensive in terms of transistors... 1 MB is simply too large for the L1 to have a reasonable access time. L1 is performance critical, so a large L1 could hurt performance more than it helps, regardless of the hit rate. Plus, if you plan on implementing multiple processors (Xbox Next?), it may be a good idea to have inclusion (where data in L1 is guaranteed to also reside in L2) to shield the L1 from remote probe requests. This in turn means that your L2 should be much larger than your L1 (or else your L2 really doesn't serve much of a purpose).

Not all that it seems...

[sirmikester]

While upgrading the processor in an xbox sounds like a great idea, the artice makes it perfectly clear that it doesn't make everything better. I don't know if I could live with only some of my games being faster and with some not working properly. Plus, I think that Xbox live is one of the Xbox's major selling points, so taking that away would really make buying an xbox pointless. But its an interesting modification nonetheless.

Not good idea.

[polyp2000]

As soon as this type of thing starts happening, you can kiss goodbye to the concept of consoles as being a set/fixed architecture. Say hello to tweaking your FPS/performance settings a'la PC gaming.. Might as well just buy a PC and stick windows on it because in the long run there aint going to be a lot of difference.

I dont want to have the inconvenience of sifting through games in a store to check if my console meets the minimum gaming requirements, and neither do a great deal of console gamers. I really dont like the sound of this at all.

I wonder what the opinion of a game developer would be on this one...

Re:Not good idea.

[hawkstone]

I don't see this ever becoming common.

First, developers would hate it. The beauty of consoles has always been that you know, as a developer, whether the game will work on everyone's system based on a single data point: whether it works in the exact same console you have sitting in your office. They don't want to have to worry about varying systems, and gamers don't want to have to worry about system specs.

It is absolutely in the developers best interest to target the lowest common denominator (i.e. the base system) to have the widest market, even if there ever did get to be a significant percentage of people with upgrades in the wild.

Not only that, it most certainly would void your warranty with MS, so I don't see the average customer ever being interested in it. How many console owners (including XBox owners) are less tech savvy than TiVo owners, and what percentage of TiVos have been upgraded? My guess is that the percent of modded TiVos is actually quite small, and the percentage of modded XBoxes (where the benefit of upgrading is even less apparent) will be even smaller.

Re:Not good idea.

[hambonewilkins]

Insightful? This isn't Microsoft doing this, it's a hack. This isn't a trend toward customizable consoles, where you'll have to make sure you meet specs. The Sega Dreamcast and Atari Lynx were also speed-hacked. But, again, not by their manufacturers. Manufacturers of consoles enjoy economies of scale when they stick to ONE console with NO options. It won't change and how you got that it would from this article blows my mind.

Re:Not good idea.

[Lightwarrior]

> The beauty of consoles has always been that you know, as a developer, whether the game will work on everyone's system based on a single data point: whether it works in the exact same console you have sitting in your office.

Except, of course, that console developers get special Development Kit versions of the consoles, that are more powerful and capable of outputting higher resolutions.

I'm sure they have access to the "normal" consoles as well, but it's an important distinction to make.

I always thought the "consoles are better for developers because it's a single target" was specious reasoning. Not only do the majority of games come out for multiple platforms, but they have several built-in limitations that are impossible to get around without forcing the player to spend more money (Final Fantasy:Crystal Chronicles, for an example). Consoles have limited graphics capability (due to the extremely poor resolution of the standard television) and a very limited control set (even the Xbox's 4-axis 10 button controller pales in comparison to a mouse with a wheel, and a 102+ keyboard), as well as a nonexistant mod base.

That's like saying Children's Books are better for Authors because children have a lower common denominator.

-lw

Re:Not good idea.

[hawkstone]

> Except, of course, that console developers get special Development Kit versions of the consoles, that are more powerful and capable of outputting higher resolutions.

You're right, I glossed over that point. The development consoles need to be different, at least to handle debug code at close to a reasonable speed. The point still stands, though, that they have a single target, even if intermediate development is on a different platform. Even with three consoles, three targets is an easier thing to deal with than, say, 5 video cards speeds X 5 cpu speeds = 25 "speed" configurations.

And even then, I've simplified the picture for PCs. Do we keep using DX8, or switch to DX9 now that it's out? Am I going to have an issue with ATI cards if I use pixel shaders? What glitz can I drop for a slow video card -- can the CPU take over some of that work? How about giving more work to the graphics card for a machine slow CPU? When the game is finished a year from now, what will be the base CPU speed we can allow? Not only do I have to worry about different speed configurations, I need to worry about different brands now, too!

> Consoles have limited graphics capability (due to the extremely poor resolution of the standard television)

True, but the limited resolution makes it easier to develop for, not harder. You may be able to do less with that resolution, but damnit, you know what that resolution is going to be and you don't have to worry about someone who wants to run at 2048x1440 and someone who wants to run at 640x480! Anyway, that is passing with HDTVs becoming more common. Many PS2/XBox games have support for progressive scan (e.g. 480p/720p).

> very limited control set (even the Xbox's 4-axis 10 button controller pales in comparison to a mouse with a wheel, and a 102+ keyboard)

That's true. However, there are many games -- entire genres -- where a mouse+keyboard doesn't make sense. Even so, that is simply another hardware restriction, which makes it easier to handle because there is less variety. You don't have to worry about supporting mouse+kbd and a joystick.

> That's like saying Children's Books are better for Authors because children have a lower common denominator.

Not sure I see the analogy. A book for a 2-year old and a book for a 6-year old bear absolutely no resemblance to one another, and trying to make one book (game) to appeal to (work on) all children (configurations) is virtually impossible.

Re:Not good idea.

[Lightwarrior]

> Even with three consoles, three targets is an easier thing to deal with than, say, 5 video cards speeds X 5 cpu speeds = 25 "speed" configurations.

Except the three consoles have vastly dissimilar architecture. Coding for a PS2 != GCN != Xbox. And once you make an Xbox version, there's not much stopping a developer from jumping to the PC. Coding for different _types_ of hardware requirements (CPU speeds, graphic cards) is significantly different than coding for entirely different hardware.

> And even then, I've simplified the picture for PCs.

You paint a much bleaker picture than is actually the case. DX8 code is forward compatible with DX9, and DX has gotten significantly more cleaned up over the years. It's much easier to support _additional_ functionality now, including detecting the capabilities of the graphics card and offloading anything that it can do in hardware off from DX's software.

It's as easy as coding for a baseline - like picking a target console - and then adding additional effects as you see fit.

> ...the limited resolution makes it easier to develop for, not harder.

We're looking at the same thing from different angles. The number of pixels you can display information on in 480i is _significantly_ lower than the number of available pixels in 1600x1200. PCs are able to present a much, much higher quality image - with the baseline of 480i, you have the additional headache of trying to present your image in an extremely limited environment.

Regardless, the number of pixels available is only part of the problem. The other part is the limited hardware - how many polys with different effects you can display, and how good they look. There's also multisampling (AA), which consoles have only barely scratched the surface of (Xbox supports 2xAA).

> Many PS2/XBox games have support for progressive scan (e.g. 480p/720p).

The PS2 & GCN support 480p, the Xbox supports 480p, 720p, and 1080i. There are very, very few games that support 1080i (3?), and a mere handful that supports 720p (10?). HDTV is not even remotely standard, and that is highly unlikely to change in the immediate future. I think that if the next round of consoles assumes HDTV will be standard, that they are in for a rude surprise.

> ...[fewer controls is] simply another hardware restriction, which makes it easier to handle because there is less variety.

Except fewer controls means that it is more difficult to require certain tasks of the user. FPSes are a great example of a really rough transition onto consoles. With the Xbox and Halo, your traditionally button-pressing right thumb is stuck on the right thumbstick, halving the number of buttons available to you. The PS2 has four shoulder buttons, so it's _slightly_ better. Still, 4-6 non-movement buttons for a FPS?

And don't even start with RTSes.

More options does not equal more restrictions - it's the opposite. And since functionality is often duplicated amongst the multiple control types (mouse, keyboard, joystick), we're not talking about huge amounts of effort. Shoot = mouse_1, joy_1, Ctrl.

> Not sure I see the [Children's Books] analogy.

I'm saying that targeting a simpler audience doesn't necessarily make it easier for the targeter to tell a really great story. We have books for adults as well as kids because adults have different, more complex tastes. I'm not reading Green Eggs and Ham in my spare time.

Developing games for a console has its own problems, quirks, and issues. It's not as simple as "consoles have set hardware, therefore it's easier". That's just one tiny part of the whole picture.

-lw

I'm curious

[ooby]

Has anyone ever been sued into existence?

Re:I'm curious

[rmull]

I think SCO is trying.

But does it support

[sharkey]

The Green Screen of Death?

But then again

[DumbWhiteGuy777]

72GB WD SATA Raptor ~$100

NVidia GeForce FX 5900 ~$300

AMD Athlon 64 FX-51 ~$750


Making your little brother cry until he wets his pants because you owned him at Halo on your suped-up Xbox - Priceless