If you're a hardware manufacturer, you can build your own PowerPC 970FX-based machines and put Linux on them. There aren't any such motherboards for sale now, though (the Momentum board does not count because it is an eval board).
So in the glorious IM2000 future my computer will pull down gigabytes of spam from random trojaned PCs whose owners say "what?" when you accuse them of spamming...
Either way, two versions of the same console would be gimmicky at best. It'd have the same problem all add-on hardware does: developers can't count on possibly expanded functionality being there on every box, so they don't spend time/money to leverage it in a meaningful way.
You mean like the PS2 and PSX? Developers aren't supposed to use the expanded functionality in the PSX, so your argument is moot.
So there would be no reason for the average consumer to buy a psx+ps3 instead of just buying a ps3 and a tivo.
Yeah, but the PS3X* will probably be cheaper than a PS3 + TiVo, because the PS3X is all in one box.
*The version of the PS3 with a built-in PVR and Blu-Ray-RW drive.
The only problem with the approach that I can see is the probability that this utility will have no incentive to run in anything but a maintenance mode. They're not going to roll fiber out to every home, for example.
One solution that has been suggested is to offload the risk to the customer. Want fiber to your home? Pay the $2,000 for the utility to install it.
If your ISP is different than the company that maintains the wires, they always point fingers at each other when there's an outage. People have discovered this over and over with DSL, and there's no reason to think it would be any different with cable.
In Austin there are three cable "ISPs": RoadRunner, Earthlink, and a local one whose name I forgot. Since they all use RR's physical plant, I choose RR since there's only one company to call and one company to blame.
The only solution to this appears to be structural separation (where the company that owns the wires is not allowed to be an ISP), but this has its own problems (like it would probably be more expensive).
SplitStream and Bullet are really more like P2P streaming systems; you can presumably get better efficiency using a bulk swarming system like BitTorrent or Slurpie.
If you care about performance or power, you need to be using the latest 90nm fab processes. If you are using 90nm, a mask set costs almost $1M. If you are going to spend $1M just on manufacturing set-up costs, you might as well spend a few million designing the chip.
Maybe they'll release the ApplePI docs right after Intel releases the Pentium 4 and Itanium 2 FSB protocol docs...
Re:more info on PowerPC 970 vector processing
on
A History of PowerPC
·
· Score: 1
Traditional vector processing (as you described in your parent post) is quite different from the SIMD that the 970 supports. AFAIK no PowerPCs support traditional vector processing.
Slashdot Mirror
It's pretty unlikely that a firmware upgrade could turn a single-channel radio into a multi-channel radio.
Maybe they're trying to hide the fact that most "RAID" these days is actually just software RAID implemented in the driver.
If you're a hardware manufacturer, you can build your own PowerPC 970FX-based machines and put Linux on them. There aren't any such motherboards for sale now, though (the Momentum board does not count because it is an eval board).
The Linux and Darwin kernels are not at all alike. Apple cannot use Linux code in Darwin because the licenses are different.
It's here: IBM SDK for 64-bit iSeries/pSeries.
Yes, there is fan control. You can even read temperature and power data out of /proc.
So in the glorious IM2000 future my computer will pull down gigabytes of spam from random trojaned PCs whose owners say "what?" when you accuse them of spamming...
Looking Glass looks like a perfect example of how eye candy is not the same as usability. Where are the productivity studies for Looking Glass?
When Sony starts making EE+GS chips on their 90nm fab, they can drop the PS2 price. It shouldn't be long now.
Either way, two versions of the same console would be gimmicky at best. It'd have the same problem all add-on hardware does: developers can't count on possibly expanded functionality being there on every box, so they don't spend time/money to leverage it in a meaningful way.
You mean like the PS2 and PSX? Developers aren't supposed to use the expanded functionality in the PSX, so your argument is moot.
So there would be no reason for the average consumer to buy a psx+ps3 instead of just buying a ps3 and a tivo.
Yeah, but the PS3X* will probably be cheaper than a PS3 + TiVo, because the PS3X is all in one box.
*The version of the PS3 with a built-in PVR and Blu-Ray-RW drive.
The PSP has 30MB of RAM. PSP games are 1.8GB.
The best way to do that is fair queueing.
The only problem with the approach that I can see is the probability that this utility will have no incentive to run in anything but a maintenance mode. They're not going to roll fiber out to every home, for example.
One solution that has been suggested is to offload the risk to the customer. Want fiber to your home? Pay the $2,000 for the utility to install it.
Start a cable ISP that does customer-to-company encryption.
You mean like the encryption in every DOCSIS cable modem?
If your ISP is different than the company that maintains the wires, they always point fingers at each other when there's an outage. People have discovered this over and over with DSL, and there's no reason to think it would be any different with cable.
In Austin there are three cable "ISPs": RoadRunner, Earthlink, and a local one whose name I forgot. Since they all use RR's physical plant, I choose RR since there's only one company to call and one company to blame.
The only solution to this appears to be structural separation (where the company that owns the wires is not allowed to be an ISP), but this has its own problems (like it would probably be more expensive).
SplitStream and Bullet are really more like P2P streaming systems; you can presumably get better efficiency using a bulk swarming system like BitTorrent or Slurpie.
Granted, Intel has better documentation.
I was just pointing out that IBM isn't the only one keeping their bus protocols a secret.
.Mac is no better than an ISP; they can change the price or terms of service at any time, or just shut it down.
That's easy; just register your own domain. The probability that VeriSign will steal it from you is lower than the other things that could happen.
If you care about performance or power, you need to be using the latest 90nm fab processes. If you are using 90nm, a mask set costs almost $1M. If you are going to spend $1M just on manufacturing set-up costs, you might as well spend a few million designing the chip.
Maybe they'll release the ApplePI docs right after Intel releases the Pentium 4 and Itanium 2 FSB protocol docs...
Traditional vector processing (as you described in your parent post) is quite different from the SIMD that the 970 supports. AFAIK no PowerPCs support traditional vector processing.
Backports are evil for exactly the reason you describe. Luckily Fedora has a policy of avoiding backports wherever possible.
How exactly is it crippled?
They can't include MP3 support precisely because Fedora is non-commercial. (Who would pay the per-copy license fees?)