Transmeta

Transmeta Nasdaq: TMTA develops computing technologies with focus on reducing power consumption in electronic devices. Transmeta was founded in 1995 by Bob Cmelik, Dave Ditzel, Colin Hunter, Ed Kelly, Doug Laird, Malcolm Wing, and Greg Zyner as a US-based corporation that designed VLIW code morphing microprocessors. To date, it has produced two x86 compatible CPU architectures: the Crusoe and Efficeon processors. These CPUs have appeared in ultra-portable Laptops, Blade servers, Tablet PCs, and a silent desktop, where low power consumption and heat dissipation are of primary importance.

History

Throughout their first few years, little was known about exactly what Transmeta would be offering. Their webpage went online in mid-1997, but for around two and a half years displayed nothing but the text "This web page is not here yet." Information gradually came out of the company, suggesting of a VLIW-based design that translated x86 code into its own native code. As Intel's then-forthcoming "Merced" processor was also a VLIW design which could translate x86 code, much wild and ultimately inaccurate speculation whipped up (though none of it by Transmeta themselves) suggesting that Transmeta's product could have supercomputer-level processing power while actually being cheaper to manufacture than any offering by Intel, AMD or Cyrix.

In fact, Transmeta marketed their micro-processor technology as extraordinarily innovative and revolutionary in the low-power market segment. They had hoped to be both power and performance leaders in the x86 space. However initial reviews of the Crusoe indicated the performance fell significantly short of projections. [1] Furthermore, the market did not stand still during Crusoe development -- Intel and AMD had significantly ramped up speeds and began to address increasing concerns about power consumption -- and so the initial offering, Crusoe, was rapidly cornered into a low-volume, small form factor (SFF), low-power segment of the market.

This forced a rapid re-design of the technology, marketed as the Efficeon processor. The Efficeon claimed to have twice the performance of the original Crusoe CPU at the same frequency. However, the performance was still weak relative to the competition, and the complexity of the chip had increased significantly. This increased size and power consumption may have diluted a key market advantage Transmeta's chips had previously enjoyed over the competition.

Transmeta has employed a number of industry luminaries such as Linus Torvalds and Dave Taylor. Initially, its purpose was kept secret, but partially because it had such talent amongst its staff, the industry was constantly abuzz with rumors in addition to 'conspiracy theories' resulting in excellent press relations (PR).

Linus Torvalds left Transmeta in June 2003 to dedicate himself to the further development of the Linux kernel. The most valuable asset held by Transmeta at the present time is their patent portfolio, and rumors persist that the company might be bought by a larger player such as AMD or even Microsoft for this reason.

As an example of technology media hype, the company was once named as the Most important company in Silicon Valley. Less well reported was that the company was never profitable while it was a chip vendor. In 2002, the company had a loss of $114 million dollars, in 2003 a loss of $88 million, in 2004 a loss of $107 million.

As of January 2005 the company announced a strategic restructuring away from being a chip product company to an intellectual property company. That is, instead of selling chips the company will now sell technology for use by other chip makers. In February 2005, there was wild speculation that AMD might buy out Transmeta. In March 2005 the company announced that it was laying off 68 people, leaving 208 employees. About half of the remaining employees were to work on propagating the LongRun2 power optimization technology within Sony products. Sony was reported to be a key licensee of this Transmeta technology.

On May 31, 2005, Transmeta announced the signing of asset purchase and license agreements with Hong Kong's Culture.com Technology Limited (Culturecom; 文化傳信) led by Chu Bong-Foo, the inventor of the Cangjie method and one of the founding fathers of modern Chinese computing. However, due to delays in obtaining the necessary technology export licenses from the U.S. Department of Commerce, the parties announced the termination of this agreement on February 9, 2006.

On August 10 2005, Transmeta announced its first ever profitable quarter.

Technology

The actual TM processors are in-order VLIW cores. To execute x86 code, a pure software-based instruction translator dynamically compiles or emulates x86 code sequences, using execution-hotspot guided heuristics. While similar technologies existed (WABI for Sun, FX!32 for Alpha) in the early 90s, the TM approach has set a much higher bar for compatibility—able to execute all x86 instructions from initial boot up to the latest multimedia instructions—while retaining most of its core performance.

Transmeta claims several technical benefits to this approach:

1. As the market leaders Intel and/or AMD would extend the core x86 instruction set, Transmeta could quickly upgrade their product with a software upgrade rather than requiring a respin of their hardware.
2. Performance and power can be tuned in software to meet market needs
3. It would be a relatively simple matter to fix hardware design or manufacturing flaws in the hardware using software workarounds.
4. More time could be spent concentrating on enhancing the capabilities of the core or reducing its power consumption without worrying about 16 years of backward compatibility to the x86 architecture.
5. The processor could emulate multiple other architectures, possibly even at the same time. (At its initial Crusoe launch, Transmeta demonstrated pico-Java and x86 running intermixed on the native hardware.)

Prior to Crusoe release, rumors indicated Transmeta was relying on these benefits to develop a hybrid PowerPC and x86 processor. However, Transmeta would initially concentrate solely on the extremely low-power x86 market.

Viability

Transmeta lost much credibility and endured significant criticism due to the poor initial Crusoe showing with large discrepancies between projections and actuals for both performance and power. On one hand, the power numbers showed a reasonable improvement over the Intel and AMD offerings. However the end user experience (i.e. battery life) only showed a marginal overall improvement. [2] First, the Code Morphing Software (CMS) combined with cache architecture artificially inflated comparisons between benchmarks and real-world applications. This is due to the repetitive nature of benchmarks and their small footprints. The CMS software overhead may have actually been a key cause of much lower performance for many real-world applications; the simple VLIW core architecture could not compete on computationally-intensive applications; and the southbridge interface was limited by its low bandwidth for graphics or other I/O-intensive applications. Some standard benchmarks even failed to run, questioning its claim of full x86 compatibility. [3]

The Efficeon addressed many of the Crusoe shortcomings and showed roughly a 2x real-world improvement over Crusoe. In addition, its die size was considerably smaller than either the Pentium 4 or the Pentium M, when compared in the same process technology. Efficeon's die size in 90nm is 68mm^2, which is 60% of the Pentium 4 in 90nm, at 112 mm^2, with both processors possessing a 1MB L2 cache.

The notion of selling a product into a specific thermal envelope was typically not understood by the mass of reviewers, who tended to compare Efficeon against the gamut of x86 microprocessors, regardless of power consumption or application. One such example of this criticism suggests the performance still significantly lagged Intel's Pentium M (Banias) and AMD's Mobile Athlon XP. [4]

For the thermal envelope in which Efficeon was designed to compete (7W and 12W), there are unsubstantiated claims that its frequency far exceeded anything else in the market, at 1.5 GHz and 7W, while the Centrino at the time could only operate within the 7W envelope when its frequency was reduced to 1.1GHz. This claim also admittedly considers only CPU frequency and does not consider other very significant factors in overall performance, such as core CPI (clocks per instruction), or memory performance and bandwidth, which have varying impact on different benchmarks and system configurations.

Unfortunately for Transmeta, other components within a laptop computer also consume power, such as the LCD display and Hard disk drive. Since laptops with Transmeta CPUs share these components with regular laptops, the net increase in battery life was not large enough to make much difference to customers.

See also

• Crusoe
• Efficeon

External links

• Transmeta
• Crusoe Japan
• Linux on laptops and notebooks with Transmeta CPUs
• CPU-INFO: Transmeta, indepth processor history
• Transmeta Announces Processor Technology License And Sale Agreement For The Chinese Market: Agreement to Deliver Transmeta Crusoe and 130-nanometer Efficeon Processor Technology to Culturecom
• Crusoe Transmeta CMS Bug