Intel preps Clover Trail SoCs for Win8 tablets
Intel is hoping to claim significant tablet market share with its dual-core x86 Clover Trail SoCs - which could power a new generation of Windows 8 devices.
Clover Trail is a follow up to Intel’s Medfield chip, which is targeted primarily at smartphones. Indeed, some industry sources believe Medfield is little more than a placeholder, at least in terms of tablets, as Intel’s mobile efforts are expected to shift into overdrive with the advent of Clover Trail.
To be sure, Intel’s Mark Miller recently told CNET that Clover Trail is an effective "vehicle for [both] Windows 8 tablets and hybrids."
As such, Intel will likely showcase Clover Trail-powered tablets running the next-gen OS at CES 2012 in January, with Acer and Lenovo slated to launch Windows 8 Clover Trail-based devices during the third quarter of 2012.
However, it remains unclear if Intel will be able to successfully challenge ARM’s current dominance of the lucrative smartphone and tablet market. Although x86 chips are formidable in terms of performance, ARM’s RISC-based architecture offers significant advantages when it comes to low power consumption and longer battery life.
For example, according to preliminary stats obtained by VR-Zone, the prototype "Medfield Tablet Platform" consumes 2.6W in idle mode, despite Intel's internal 2W target.
Meanwhile, video playback at 720p in Adobe Flash significantly ratcheted up power consumption to a whopping 3.6W - 1W more than the 2.6W low Santa Clara had hoped for. The again, Medfield did score some pretty sweet benchmarking stats, weighing in at a very respectable 10,500 in Caffeinemark 3.
In comparison, Nvidia's (last-gen) Tegra 2 scored 7500, Qualcomm's Snapdragon MSM8260 hit 8000, while Samsung's Exynos posted a score of 8500.
"Yes, the performance numbers look very competitive compared to shipping ARM parts. However, it is still high on battery life and will be largely competing with the next generation of ARM which will be more powerful," industry analyst Rob Enderle told TG Daily last week.
"ARM is entrenched which means Intel doesn't just have to be equal or a little better they'll need to be significantly better enough so that users will notice a difference. That is at least 10% but to assure success closer to 20%.
"Remember, this same metric worked for them against Transmeta and DEC Alpha, now they are seeing the ugly side of competing with an entrenched vendor. In short, given their history, being a bit better than ARM should be achievable, but the problem is they need to be a lot better - which will be quite difficult to achieve."