My first post and my first tryst with blogging – and the event couldn’t have been better. After all, not everyday does one get to meet so passionate an engineer who could dare to give up the comfortable cushion of an IIT-B degree midway through it , to pursue research in his field of interest.
Someshwara Dixit, processor division senior engineer at ARM was here as part of the ‘Beagle Board Day’ seminar and he gave us wonderful insights into the history of the ARM legacy, salient features of their different cores (if only I had done a computer architecture or related course to be able to grasp all that.Sigh.) I caught up with him during the tea break and was amazed at how the man was patient enough to spend time on basic fundes clearing queries for a total novice like me. I bugged him almost for an hour and got some interesting, simplified and elegant examples about how the ARM guys manage to bring up processor cores that beat all competition in terms of cost and power consumption. He spoke of design considerations in ‘pipelining’ and about ARM guys being the best in branch prediction due to ingenious methods ,about the testing language/platform initiative that his team took up to save the millions that ARM would have to invest otherwise on a proprietary testing language. We also discussed about Qualcomm improving on ARM core architecture and coming up with state of the art 1.0 GHz chipsets by transforming a 600 MHz core. Having worked on the Atmel platform and being still unaware of the exact difference between an MCU and its core, I was keen to know what sort of transformations are made when the ARM core design is taken up by a chipset manufacturer like Atmel, Qualcomm etc.
Scattered through the talk were examples of business wisdom and stragetic decision making they employed – be it dealing with IP issues, partnership strategies (ever wondered why we dont have “ARM inside” written on our cell-phones, even when most Nokia phones and Apple iPods are ARM based,like the way we have “Intel Inside” written on our comps ?) ARM deploying their cores on FPGAs to help out start-ups innovate was probably one of the best parts of the talk.
The examples he gave re-enforced one idea very well – No matter what level you are at, get your technology to the best standards and your vison crystal clear, leadership in business follows irrespective of any shrewdness that others employ.
Talking about the road ahead, ARM has already achieved domination in the mobile space and we got to hear about how ARM plans to compete with other players like Intel not by colliding head-on with them in the Desktops space where Intel is the largest, but by bringing the capability of the Desktop/Laptop devices in their domain by designing cores for chipsets for MIDs and other devices which will change the way we deal with computers. On a rather philosophical note, he said that when they sit working on 32nm technology, they often wonder about where it goes from here. The current 130/100 nm (I am not sure of the exact number) technology used in cell-phones has already brought in mind-boggling possibilities, when 32 nm hits markets, we never know what to expect. He says innovation then will be totally in the hands of CS junta [70 mm smile]. Exciting indeed.