UMPC: Future Wars...

Now that WL's UMPC woes are over, WL is taking time to reflect on the current situation of UMPCs/MIDs and where they are going.

In recent months, with the release of Intel Atom Centrino, it is obvious that WL's idea of UMPC is becoming more and more distorted. First, WL learnt that the Intel Atom CPUs were designed more for low battery consumption, and in doing so sacrifice CPU power (or at the very least not improve CPU power). In 2009, Intel plans in their architecture aims for more power conservation, and we are unlikely to see stellar improvements in CPU performance. This means that the Dothan-class 800Mhz++ CPU with 512kB L2 cache is probably going to stay for the next 2-3 years. This also means that no UMPC or rather MID as they prefer to term it now will outclass MY P161o Yonah 1.2GHz ULV CPU. Futuremore, Microsoft recent announcement to move Windows Mobile into Intel Atom and the MID category further reinforces my prediction that this is not all about CPU power game.

Now, I know as a secondary electronic device CPU is not all that important, but a decent CPU is needed to achieve some productivity. This is why the PPC platform has failed. It is too fucking slow. What's the point of carrying all my files in a PPC when it takes like 5sec to 30sec+ to open a PDF file depending on its size? The "instant"-feature promised by the PPC gets drowned when one really tries to use it for proper work. Admittedly, I have not used a Eee PC or Kohjinsha before, but based on my current ULV challenge (a test to see whether I can live with a ULV processor), in which I purposely downclock my CPU to 1.0Ghz to 1.2Ghz, I can tell you that it is sometimes unbearable. Ok... so I am a demanding person and I like everything fast-fast... Sue me...

This is good news for me actually. It means that my P1610 will have a longer lifespan since being a unique class of item, it will remain unique for the foreseeable future and will not be challenged by upcoming MID and UMPC. Even Intel's definition of a Intel Atom device (or was it Intel Atom centrino) was to have a screen size from 5" to 7"... which automatically excludes a device like P1610 from receiving the brand name. The focus of upcoming devices I predict, is more of MID-class device with few UMPC-class device in the future. And those UMPC-class device are likely to remain at a high price premium, while those MID-class device will remain too underpowered and small for my use.

At the end state, in 3-4 years from now the entire Intel Atom line and its successor may prove to be useful. It is possible that in the future we shall see a Pocket PC Phone armed with a Intel Atom CPU (probably the 2009 or even 2010 version) that consumes sufficiently low power to the point that it can be used to power a phone. This would bring much rejuvenation to the PPC market, who has been stagnating at 400-500Mhz XScale CPU power for like the past 4-5 fucking years. In fact, this would probably tie in to the reason why Intel abandoned their XScale business... to start from mobile CPU and tune their power requirements down all the way to the phone requirement (this in fact has been reported before). Finally then, the PPC phone might actually live up to its intended function. Web browsing will be smoother, PDF documents won't take ages to load... hey it's not going to be a blazing speed demon capable of encoding videos, but I believe that is a good CPU baseline for a maximum mobility device like a phone.

This is also a message to me, to fuck care the PPC arena until it changes. I believe a 2010 timeframe would be a good time to outlook for such uber-PPC devices. This means WL, no more new phones for you, unless your current one breaks down... and even if it does, no need to get the new flashy ones... cause ultimately they are all castrated until we get a new technical revolution in the PPC world.