I would have to agree. Transmeta developed chips years ago for a market that didn't exist, and still doesn't. The Crusoe (sp) was originally developed to fill the _Ultra-low-power-but-still-full-featured_ niche that was being researched for handheld PC's.
The problem is that Handheld PC's never happened. There are essentially three markets for computers right now: Server, Desktop, and embedded, with Crusoe hoping to land in the region between the latter two.
What it boils down to is this:
Unfortunately for Transmeta, people were never willing to shell out for a hybrid device that looks like a handheld, but works like a PC. People would much rather spend a few extra bucks and get a slightly bigger and heavier battery to run a much faster processor, or just take the plung and get an embedded device. They gambled and lost, or at least they haven't won yet.
Maybe they will have a market in a few years, when handheld units require more processor than a Strongarm\Mips can provide.
What makes this really interesting is the fact that Intel already has quite a bit of the OEM server market in the bag.
HP's new 8800 proc, was designed to be both bus and pin compatible with Itanium. That had to of been a deliberate act on HP's part quit a while back during the design process. Also, with Compaq dropping Alpha and acquiring HP, it seems as though at least a few key players have been betting on Itanium's success.
That's what the 'explicit parallel' architecture they're referring to is all about: completing several operations at the same time or during the same clock cycle. That was all that older computers and OS's could do (Original DOS for instance). Back then, a megahertz rating on a computer meant everything simply because of the fact that the computers were explicitly serial (only able to do one operation at a time). Now we're getting into computers that operate more like the human brain, where a 'speed' rating is absolutely useless. When was the last time you heard an I.Q. rating related to the physical chemical speed of a persons brain?
Intel is starting into this field with SIMD, although, you can get the same thing on a different scale with multiproc system and the right OS.
I'm not saying the figures are right, but that's what they're pointing at.
Slashdot Mirror
"If you come off as paranoid, nobody's going to put you in charge of stuff."
:)
Not true.
You should see some of the Paranoid network-nazis I work with.
I would have to agree. Transmeta developed chips years ago for a market that didn't exist, and still doesn't. The Crusoe (sp) was originally developed to fill the _Ultra-low-power-but-still-full-featured_ niche that was being researched for handheld PC's.
The problem is that Handheld PC's never happened. There are essentially three markets for computers right now: Server, Desktop, and embedded, with Crusoe hoping to land in the region between the latter two.
What it boils down to is this:
Unfortunately for Transmeta, people were never willing to shell out for a hybrid device that looks like a handheld, but works like a PC. People would much rather spend a few extra bucks and get a slightly bigger and heavier battery to run a much faster processor, or just take the plung and get an embedded device. They gambled and lost, or at least they haven't won yet.
Maybe they will have a market in a few years, when handheld units require more processor than a Strongarm\Mips can provide.
What makes this really interesting is the fact that Intel already has quite a bit of the OEM server market in the bag.
HP's new 8800 proc, was designed to be both bus and pin compatible with Itanium. That had to of been a deliberate act on HP's part quit a while back during the design process. Also, with Compaq dropping Alpha and acquiring HP, it seems as though at least a few key players have been betting on Itanium's success.
That's what the 'explicit parallel' architecture they're referring to is all about: completing several operations at the same time or during the same clock cycle. That was all that older computers and OS's could do (Original DOS for instance). Back then, a megahertz rating on a computer meant everything simply because of the fact that the computers were explicitly serial (only able to do one operation at a time). Now we're getting into computers that operate more like the human brain, where a 'speed' rating is absolutely useless. When was the last time you heard an I.Q. rating related to the physical chemical speed of a persons brain?
Intel is starting into this field with SIMD, although, you can get the same thing on a different scale with multiproc system and the right OS.
I'm not saying the figures are right, but that's what they're pointing at.