Computer Engineering Seminar

GreenDroid: An Architecture for the Dark Silicon Era

Michael TaylorProfessorUCSD
SHARE:

The Dark Silicon Era kicked off with the transition to multicore and will be characterized by a wild chase for seemingly
ever-more insane architectural designs. At the heart of this
transformation is the utilization wall, which states that, with each
new process generation, the percentage of transistors that a chip can
switch at full frequency is dropping exponentially due to power
constraints. This has led to increasingly larger and larger fractions
of a chip's silicon area that must remain passive, or dark.



Our research attacks this dark silicon problem directly through a set
of energy-saving accelerators, called Conservation Cores, or c-cores.
C-cores are a post-multicore approach that constructively uses dark
silicon to reduce the energy consumption of an application by 10x or
more. To examine the utility of c-cores, we are developing GreenDroid,
a multicore chip that targets the Android mobile stack. Our mobile
application processor prototype targets a 32-nm process and is
comprised of hundreds of automatically generated, specialized,
patchable c-cores. These cores target specific Android hotspots,
including the kernel. Our preliminary results suggest that we can
attain an average 7x improvement in energy efficiency using a modest 7
mm^2 of silicon.

Michael B. Taylor is an Assistant Professor of Computer Science and Engineering at UC San Diego. His research interests focus around the design of novel computing artifacts. His recent projects include Kremlin, a tool best described as "like gprof, but for parallelization" GreenDroid, jointly led with Steven Swanson; and SD-VBS, a vision benchmark suite. As a PhD student at MIT, Michael was the lead architect of the 16-core MIT Raw processor, which was later commercialized into the 100-core Tilera chips. Prior to that, he
co-authored the first version of the Connectix VirtualPC x86-to-PowerPC translator, and hacked microkernels at Apple. He received the NSF CAREER Award in 2009, a PhD from MIT in 2007, and an
AB from Dartmouth College in 1996. He has been writing code for 86% of
his life.

Sponsored by

Professor Scott Mahlke