Intel’s cores utilise an x86 instruction
set, and running an x86-based smartphone is a problem if there’s no operating
systems you can run on it. However, that’s not a real problem at all. Writing
in a piece on the Anandtech website, Anand Al Shimpi reported Intel had been
“contributing x86 patches to the AOSP [Android Open Source Project] and
Google’s internal developer branch for the past two years.” Such optimisations
could theoretically have been included in new versions of the OS, and according
to the same piece, “All of Intel’s x86 support should be included as of Android
cores utilise an x86 instruction set
Android’s x86 credentials were given a
further boost in December when Android 4.0.1 was ported to work on x86
processors. And of course there’s the ongoing work of the Android x86 project,
dedicated to this specific marriage of originally non-mobile architecture to a
completely mobile-orientated OS. Most telling in regards to the question of an
OS is the fact that as recently as the end of May, Intel released an x86
emulator for Android 4.0.3 (Ice Cream Sandwich - ICS). That gives the rather
strong impression that, like Gingerbread before it, ICS will be hitting x86,
and therefore devices utilizing Intel’s SoC solutions, pretty soon.
This was corroborated at the San Diego
event in London where Intel representatives confirmed that though the device
will launch with Gingerbread, Intel anticipates an update in the not-
too-distant future. Presumably that will be once x86 apps are more available.
Currently around 70% of available Android apps will run on the Medfield
platform. That’s not too impressive, but if Intel can wow on a performance for
price front, any users opting for the San Diego shouldn’t have to wait too long
for Ice Cream Sandwich. Of course, if the aforementioned emulator does its job
in enabling the speedy debugging of x86 apps, the arrival of ICS to the San
Deigo should then also bring a more impressive percentage of compatible apps.
The other issue that the arrival of x86 to
smartphones raises is one of power capabilities - specifically, whether Intel’s
SoC solution can match up to ARM- based ones while still delivering punchy
processing. This question stands to reason, as Intel is scaling down an
architecture most traditionally associated with desktop performance, and a
range of cores primarily developed for laptop/tablet performance. Getting the
balance between x86-style performance with mobile power considerations is a
rather tricky juggling act. Certainly from the brief glimpse offered to us,
performance isn’t much of an issue. The Saltwell core in the Z2460 is single
cored, but has Hyper-Threading to boost its productivity. However, it’s not
quite the same situation for power consumption.
We shouldn’t forget about Smart Idle
Technology (SIT), which will allow the Z2460 to effectively switch off while
the OS remains ‘on’, but we shouldn’t expect too much from it either.
shouldn’t forget about Smart Idle Technology (SIT), which will allow the Z2460
to effectively switch off while the OS remains ‘on’, but we shouldn’t expect
too much from it either.
At the time of writing, no independent
battery tests have been completed, but the information available suggests that
there’s room for improvement (isn’t there always?). Intel says that the San
Diego should last up to eight hours on standby while connected to 2G, and 14
hours while connected to 3G. No figures for talk-time or high-performance use
have been suggested, but it’s fair to say that gaming and other similar
activities will probably deplete the battery pretty quickly. We shouldn’t
forget about Smart Idle Technology (SIT), which will allow the Z2460 to
effectively switch off while the OS remains ‘on’, but we shouldn’t expect too
much from it either.
SIT will put the Saltwell core into what
Intel terms the C6 state, or deep power down. It should help Intel to manage
the power somewhat, but this is a technology developed in the first instance
for laptops and tablets. Even in the technical details provided to us by Intel,
a concession is made: “Leading performance with competitive energy efficiency.”
We shouldn’t really be expecting leading battery life, though. Not yet from
Intel, and not at the price OS of the San Diego. But even then the company is
planning ahead. With 22nm and 14nm fabrication processes planned, deficiencies
in power draw will be certainly addressed going forward.
Intel’s success in smartphones (and any
move towards other mobile devices for which Intel would like to provide SoC
solutions) will depend somewhat on the reception of the San Diego. Not
completely, though. Users will care far less about the technology behind the
Medfield platform than the performance of the device, and if the San Diego can
get across its message of high performance at a good price, both users and
Intel should be satisfied.
success in smartphones will depend somewhat on the reception of the San Diego.
But even if, for some reason, Medfield
doesn’t take off, Intel is now intellectually and financially invested in SoC
solutions for smartphones. Clover Trail and the dual-core Atom Z2580 are just
around the corner, and with all of its financial and technological clout, it’d
be a brave person to think one phone’s arrival in Europe is anything other than
a statement of intent.