AMD launched its mobile followup to Kabini and Temash yesterday with a bevy of synthetic benchmarks, but absent were hardware wins.

mullins cpu die shot block amd If a chip launches with no hardware wins, is it really a launch at all?

AMD “launched” Beema and Mullins Tuesday, its latest mobile offerings and a followup to Kabini and Temash. In a nutshell these two chips are repackaged versions of Kabini and Temash with slight bumps in clock speed and features added. But there is one thing missing from this launch: a complimentary announcement of hardware that these chips will be featured in.

Granted, most friendly hardware blogs called this a “preview” but the bulk of the correspondence from AMD and a number of other sites called this a launch so launch it is. There’s not much new in this announcement that wasn’t known before from VR-Zone’s initial report last summer, to AMD’s first official announcement in the fall. Nearly everything about the architecture is the same, and these two chips do not support HSA.

Below are the final clock speeds of the chips:

Beema3 1024x252 If a chip launches with no hardware wins, is it really a launch at all? Beema2 1024x294 If a chip launches with no hardware wins, is it really a launch at all?

AMD has also released a series of performance claims based on benchmarks which can be seen below:

Beema5 1024x517 If a chip launches with no hardware wins, is it really a launch at all? amd beema performance chart 100264359 orig 1024x408 If a chip launches with no hardware wins, is it really a launch at all? Beema4 1024x525 If a chip launches with no hardware wins, is it really a launch at all?

As you can see, while AMD has made some sizable performance gains over Kabini and Temash it still falls short of Intel’s mobile offerings — which is why AMD has yet to announce hardware wins for these two chips. An announcement on this matter is promised for AMD’s APU14 conference scheduled for May 15 in Beijing, but considering AMD’s current mobile hardware wins one shouldn’t hold their breath. While Atom is in 229 tablet SKUs — a minuscule amount compared to what ARM commands but an expensive bargain considering Intel’s contra revenue efforts — AMD is in just one.

Manufacturers will be highly hesitant putting an x86 chip — unpopular in the tablet world — into one of their devices without some sort of compelling competitive bargain from the silicon’s owner. Intel can create that bargain with its cash reserves, but AMD can’t do the same. Considering the risk that the unpopular x86 platform presents in the tablet world, the non competitive performance as well as higher-than-Atom power draw of Beema and Mullins, there’s no reason to expect AMD to make any big gains in the tablet world anytime soon. This is a commendable effort on AMD’s part, but not good enough to rack up any hardware wins.