1. NVIDIA Confidential
Exascale
An Innovator’s Dilemma
Jen-Hsun Huang, CEO
SC11, Seattle, Washington | Nov. 15, 2011
2. "Generally, disruptive innovations were
technologically straightforward,
consisting of off-the-shelf components
put together in a product architecture
that was often simpler than prior
approaches. They offered less of what
customers in established markets
wanted and so could rarely be initially
employed there. They offered a
different package of attributes valued
only in emerging markets remote from,
and unimportant to, the mainstream.”
8. CPUs Fast But Complex
Optimized for single-threaded
performance
~50X energy to schedule
instruction than the operation
~20X energy to move data across
chip than the calculation
9. Super Efficient Processors Needed
Many simple processors with
minimal overhead
Locality reduces data movement
energy
Poor single-threaded performance
Innovator’s Dilemma!
10. PRINCIPLE #1
Companies depend on customers
and investors for resources.
Clayton M. Christensen
(1997) The Innovator’s Dilemma:
When New Technologies Cause Great Firms to Fail
11. PRINCIPLE #2
Small markets don’t solve growth
needs of large companies.
Clayton M. Christensen
(1997) The Innovator’s Dilemma:
When New Technologies Cause Great Firms to Fail
14. Disruptive technologies underperform
established products in mainstream markets.
Cheaper, smaller, and frequently more
convenient.
Clayton M. Christensen
(1997) The Innovator’s Dilemma:
When New Technologies Cause Great Firms to Fail
18. Lifecycles of fish
in Australia
University of Melbourne
Stars and galaxies
12.5B years ago
University of Groningen
Neural networks in a
self-learning robot
The University of Plymouth
Directives
65x in 2 Days 5.6x in 5 Days 4.7x in 4 Hours