On 9/20/2013 1:22 PM, Chris de Morsella wrote:
>> A computation always takes a nonzero amount of energy to perform, theoretically you
can make the energy used be as close to zero as you like, but the less energy you use
the slower the calculation.
How does that square with the increased (well measured) energy efficiency per
fundamental unit of logic (single machine operation) -- it takes far less energy to
perform an elementary logic operation on a modern CPU than it did on say a CPU from ten
years ago (even if the modern CPU may suck down more total power -- it is performing far
more work)
Modern CPUs clearly are also operating at much higher speeds. I think you are not
factoring in the dimension of scale or the physical size of the logic
container/state-machine. As the size of a logic gate is scaled down it takes less energy
and can operate at a higher clock speed.
http://en.wikipedia.org/wiki/Performance_per_watt
"For example, the early UNIVAC I <http://en.wikipedia.org/wiki/UNIVAC_I> computer
performed approximately 0.015 operations per watt-second (performing 1,905 operations
per second (OPS), while consuming 125 kW). The Fujitsu
<http://en.wikipedia.org/wiki/Fujitsu> FR-V <http://en.wikipedia.org/wiki/FR-V> VLIW
<http://en.wikipedia.org/wiki/VLIW>/vector processor
<http://en.wikipedia.org/wiki/Vector_processor> system on a chip
<http://en.wikipedia.org/wiki/System_on_a_chip> in the 4 FR550 core variant released
2005 performs 51 Giga-OPS with 3 watts of power consumption resulting in 17 billion
operations per watt-second.^[1]
<http://en.wikipedia.org/wiki/Performance_per_watt#cite_note-1> ^[2]
<http://en.wikipedia.org/wiki/Performance_per_watt#cite_note-2> This is an improvement
by*over a trillion times in 54 years*."
Size (or rather the lack of it) matters in this equation.
But there still a limit because entropy has to be dumped into the environment, which is
not at 0deg, if a register is to be erased.
http://en.wikipedia.org/wiki/Landauer%27s_principle
But we are many orders of magnitude from the Landauer limit now - lots of room for
improvement.
Brent
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/groups/opt_out.
