Some answers:
power. Multiprocessor (DSP+ARM), more address space, IO, ISA, power again..
here is one set:
1) speed/MIPS.2) co-processor support3) No. of I/O s4)
the peripherals
supported by both the processors5) amount of addressable
memory6) security
options
now, my view:(note, I have been a ARM guy through out my professional life).
Have ya all heard of Intel's SOC with Pentium? if u guys been following news of intel selling off its Xscale processor family to Marvin, u guys might be aware of this.. dont yet sign off Intel from the match.. with it comes a BIG set of apps... http://www.linuxdevices.com/news/NS2460545479.html and ofcourse.. two things to keep in mind: the next gen 35nm phreus or what ever new hafnium based chips will boost intel and IBM.. all ARM IP users need to scale to that tech.. but with guarenteed lower power.. will the reign of ARM still remain?? imagine the entire power of win32 and other already huge development base coming on intel based cellphones.. it'd be fun to watch how ARM responds.. look at it this way: GB ram - with IBM-DRAM on 45nm technology today?- is going to be dirt cheap - so already is the in non-volatile memories 4GB nands and what not.. 3D cores already available.. all u need is a processor - the missing link.. u'd have windows vista on cellphones - kill 2 birds in one shot -develop for a "PC" and it works on a cell too.. the only killer will be I/O ->LCD replace with tvout(already here - not too useful on travels), and input keypad.. that would be only differentiator i think.. more of cross breeding to see... life is going to be pretty exciting once more in the technology front now.. :)
For those with short memories: remember the success factors:
1. technology/sales(read cost/performance)
2. user base (again read cost/feature set)
3. developer base (again read cost of development)
No comments:
Post a Comment