mbed Technical Lead at ARM
Cambridge, United Kingdom
mbed Technical Lead at ARM
Cambridge, United Kingdom
I work as an applied researcher in the microelectronics industry.
My approach is to be a generalist among many highly specialised experts, understanding a wide range of fields and hence the implications of a decision in one narrow area on the many others.
The goal is to enable me to spot problems, opportunities and solutions earlier, and create revolutionary advances by simultaneous evolutionary development in multiple fields.
My current interests are simplicity, productivity, conventions, the paradox of choice and my new kick-ass project which aims to apply all of these!
(Public Company; 1-10 employees; Semiconductors industry)
April 2007 — Present (1 year 7 months)
Give us a few months...
(Public Company; 1001-5000 employees; ARM.L; Semiconductors industry)
January 2006 — April 2007 (1 year 4 months)
Various research focused on system and multi-processing hardware, software and tool challenges.
(Public Company; 1001-5000 employees; ARM.L; Semiconductors industry)
June 2003 — December 2005 (2 years 7 months)
Technical Lead for the ARM NEON/v7 Architecture, first featured in the Cortex-A8 Applications processor. NEON provides easy-to-target media acceleration within the ARM programming model, and includes a number of innovations which compilers love.
Involved in SIMD research, developing the instruction set architecture, and vectorising compiler research. We built the prototype compiler at the same time, and by defining SIMD architectural concepts that a compiler can consistently follow and some novel architectural features to enable it, NEON is the best SIMD compiler target in the world! Welcome to SIMD 2.0
I announced NEON to the world at Fall Processor Forum '04, and have been involved with productisation and partner engagements during and after development.
NEON was actually the code name, but we thought it was a just such a cool name we campaigned to keep it!
(Public Company; 1001-5000 employees; ARM.L; Semiconductors industry)
June 2002 — June 2003 (1 year 1 month)
Next-generation computer architecture modeling and translation research.
Doing some interesting experiments with new architecture ideas under our internal rapid prototyping framework. Some of the findings fed in to both Thumb-2 and NEON.
The framework was built using Haskell. If you haven't ever learnt a functional programming language (a sensible one like Haskell or ML), you really should. It'll make you look at the world in a whole different way!
MEng, Electronics, 1998 — 2002
Awarded a MEng 1st in Elecronics from Southampton University (still #1 for electronics and computer science!).
Work included optical audio transmission, bluetooth stacks, a new approach to asynchronous logic, and a stressful phone-call analysis project.
A highlight has to be the littleRASCAL 16-bit RISC processor design I did with Rob and Alex, which was our own architecture, layed out by hand using our own custom designed cell library, programmed using our own assembler. It got fabbed too, but our rip-off of the intel inside logo we did in a metal layer gave some DRC headaches!
The fab exploded (see http://www.ecs.soton.ac.uk/regenesis/pictures/), but they'll have some fab new facilities to replace them.