Slashdot Mirror
Transmeta Closing Up Shop
Ashutosh Lotlikar wrote to mention an article on the Business 2.0 site stating that chip producer Transmeta is going out of business. From the article: "The company's Crusoe family of microprocessors promised lower power consumption and heat generation, enabling the creation of laptops with longer battery life. Critics bashed the chips for being underpowered compared with Intel's latest and greatest. Transmeta struggled to find a market, and recently it sold off most of its chipmaking business for $15 million to Culturecom Holdings, a Hong Kong company better known for publishing comic books."
I wonder what would have happened if Transmeta had released the instruction set for the native VLIW instruction-set processor that runs the x86 emulation layer. Sure, it's probably very hard to code for, but may have offered a tremendous advantage for some applications.
Also, hopefully OQO and others have a backup plan so this doesn't put a kink in the handheld pc market.
HIV Crosses Species Barrier... into Muppets
If Transmeta does close shop, I hope they consider opening up their "Code Morphing Software". It's an interesting approach to X86 processing on non-X86 processors, for more info check here: http://www.transmeta.com/crusoe/codemorphing.html
Sadly, PS/2 was yet another victim of USB, which doesn't care what you plug into it, the electrical slut.
Comment removed based on user account deletion
When they came out, they definitely had the best MIPS/Watt for x86-compatible chips. I bought a Crusoe-powered laptop back in 2002 (Fujitsu P-series). It routinely got over 10 hours of battery life with the screen at full brightness and over 20 with the screen closed listening to MP3's. With the original batteries, it still gets 6-7 hours with the screen, and 15ish with it closed. It also doesn't get uncomfortably hot, and also has builtin wifi drawing power.
I've never seen an Intel-powered laptop that could come close to that. Granted, it is a dog (and was even then), but a similar Intel-powered notebook draws more power. If you were to scale-back Intel's current offerings to match the speed of my laptop, they'd probably beat it in MIPS/Watt. However, at the time there was nothing comparable.
If nothing else, Transmeta will have prodded Intel and AMD to make more power-efficient chips.
FWIW; I'm an embedded firmware and hardware developer amoung other things, and HAVE worked with their hardware:
I evaluated transmeta's chips in 2003, I think.. it was for a target product that needed a low power consumption. When we got their development kit and the heatsink was huge, I knew they were in trouble. I KNEW they were in trouble when we tried to return the multi-thousand-dollar kit to look at some other options they had.. and they wouldn't listen.
If you're working in the embedded world, you're probably in a well defined area:
- Low power, low speed micros. These are usually under 20mhz, sometimes faster. Cost a couple bucks and have everything under the sun integrated. Some have micro RTOS's developed for them, most don't. This market is mature and owned by people like Atmel, Microchip, Zilog, and a hoarde of other people making variants of chips like the 8051. Transmeta didn't stand a chance there. Those chips consume almost no power at all and cost nothing.
- Midrange micros for pdas and other appliances. This is where I thought transmeta had a chance, but then along came Intel with the XScale architecture and they made it work and work very well. This, not the pentium M, is what killed them I think. XScale is cheap, well supported, and very low power.
- Above-midrange; Transmeta might have had a shot here, but their power consumption and support was much worse than the x86 compatible Nat Semi Geode (now owned by AMD?), and offerings from Via (C3 MiniITX). Price? No competition.
- Notebooks. Pentium M ended this one. So did the G4 chip from Motorola.
- Desktop high end CPUS. Nobody ever expected them to be competitive.
Looking back, it seems like their market ran away from them whereever they looked. Unfortunate, but not unforseeable IMO.
..don't panic
Sun's MAJC: dual core VLIW FP monster...gone
Transmeta: also VLIW...going
Intel: Itanium VLIW FP monster...stagnant once HP's base converts from PA-RISC and Alpha
It seems that no VLIW architecture to date has really been successful against PowerPC, SPARC, and AMD64. Is it the compilers? Too nontraditional?