• Just looking around
    Posts: 15 from 2016/1/22
    >What makes you think it is going to blossom over the next ten years?
    The change of owner has brought some growth. Moreover, the market needs a server processor which is power efficient and the ARM alternative is no better than x86. Finally, MIPS is also being used in cheap chinese smartphones and tablets due to the low cost, high energy efficiency and availability of the Android operating system. In other words, it has potential both in the datacenter (due to high energy efficiency) and the mobile world (due to both energy efficiency and availability of Android).

    > Convergence box? You mean those little Android TV boxes you can get off eBay for £25 ($35 US)?
    No, these only do basic streaming, please read my posts again. Moreover, the Roku is much more popular than GoogleTV/AndroidTV.

    >>In other words there will be another popular architecture along the x86 and the ARM since ARM
    >> has an adequate power efficiency in the mobility context but not the datacenter context.
    > Why do you keep saying this?
    > The current ARM implementations targeted for the datacentre have all been based on older
    > silicon processes thus giving Intel an advantage
    Well, Intel is not good, even if they get to the level of Intel, it would still be terribly bad. MIPS and PPC do have a higher power efficiency and many datacenter operators want something with a power efficiency better than what Intel (and soon: ARM) can offer.

    > http://www.adapteva.com/andreas-blog/semiconductor-economics-101/
    > I estimate development of a SoC as described by you would cost at least 150 million USD.
    Well, the link which you supply states that the cost won't necessarly be that high. According to your link much of the cost is in paying the hardware engineers and the IP licencing cost. Sometimes it may be possible to pay a cost per unit produced for the IP licencing, this allows much lower upfront costs (but much higher total costs). For the cost of the hardware engineers, the key is to have only 3 engineers to keep costs reasonnable. This cost can be brought even lower if we find some who are former Amiga users and do it as a labour of love, similar to the way the developers of the MorphOS team are not paid the "true price".
  • »16.03.16 - 23:52
    Profile