MIPS to go after the cellphone market -- really

Wishful thinking?
To judge whether the MIPS people aren't just caught up in wishful thinking, it's necessary to ask whether MIPS already has the right stuff, like low-power cores for mobile handset chips. The bigger question is: What could possibly tip mobile OEMs to ditch ARM and go for MIPS?

Joseph Byrne, a senior analyst with the Linley Group, is another analyst who was caught by surprise at MIPS' intended plan to go after cell phone application processor sockets. He cautioned: "MIPS faces an uphill battle."

With processors, the key success factors are software, price, performance, power, and price/performance and performance/watt.

First, let's examine the technologies behind those competing cores. How does MIPS stack up against ARM?

Linley Group's Byrne said that the MIPS 74K is not as power efficient as the ARM Cortex-A9 or Cortex-A8 in terms of mW/MHz.

"Normalizing for actual application performance, the 74K may be more efficient than the A8, but is unlikely more efficient than the A9," he added. "Part of the reason is that MIPS hasn't focused as much on power efficiency."

Further, he said: "It's possible that minor tweaking (e.g., to the distribution of clock buffers) would yield an improvement in power consumption of the 74K and put it at parity. Theoretically, the 74K is likely to scale to higher clock rates than either Cortex, but only ARM is pointing to a 2GHz, albeit hand-tuned and hardened, implementation."

Jim McGregor, chief technology strategist at In-Stat, acknowledged that there is no one clear answer. Both architectures -- MIPS and ARM -- offer synthesizable cores that could be used in cell phones, smartphones, and many other mobile applications, he said.

"They are both viable building blocks for an SoC, which is the final processor solution for all these applications," he said. "The performance, power consumption, and features of the final SoC depend upon the chip designer, other IP, and manufacturing process."