SLIDE 1
1
Neuromorphic Analog VLSI
David W. Graham
West Virginia University
Lane Department of Computer Science and Electrical Engineering
2
Neuromorphic Analog VLSI David W. Graham West Virginia University - - PowerPoint PPT Presentation
Neuromorphic Analog VLSI David W. Graham West Virginia University - - PowerPoint PPT Presentation
Neuromorphic Analog VLSI David W. Graham West Virginia University Lane Department of Computer Science and Electrical Engineering 1
SLIDE 2
SLIDE 3
3
- !
- "#
$% $ !
&'
SLIDE 4
4
!&()
!&() * +' ,(-' .
/)0
- .
- '
' 1 +'
'
2 ' , /
- 3224
/ 522 &
'
- 6'
SLIDE 5
5
Interest in exploring neuroscience Interest in building neurally inspired systems Key Advantages
- The dynamics is the system
- What if our primitive gates were a neuron computation?
a synapse computation? a piece of dendritic cable?
- Efficient implementations compute in their memory elements
– more efficient than directly reading all the coefficients
- Precise systems out of imprecise parts
' 7
SLIDE 6
6
Similar physics of biological channels and p-n junctions
- Exponential distribution of particles
(Ions in biology and electrons/holes in silicon)
- Drift and Diffusion equations form a built-in Barrier
(Vbi versus Nernst Potential) Both biological channels and transistors have a gating mechanism that modulates a channel.
&'56
SLIDE 7
7
/2
0.1nm 1m 1cm 0.1mm 1µm 10nm CNS Neurons Synapses Channels Molecules Silicon Transistors Logic Gates Multipliers PIII Parallel Processors
SLIDE 8
8
8
6''29,6: '6' 9 ;1 &9909(&. 622 9
SLIDE 9
9
/ <7
&(' 8'
862
- 8'
' .2' 34 62
82)=('
- ,
/
(8
/ 56 5
SLIDE 10
10
'7
) /2'2 22' >' '
SLIDE 11
11
86
1980 1990 2000 2010 2020 2030
Year Programmable Analog Power Savings >20 Year Leap in Technology
0.1mW 10mW 100W 1W
Gene's Law DSP Power CADSP Power Gene's Law DSP Power CADSP Power
10nW 1µ µ µ µW
Power Dissipated / MMAC
<?
82 ='6 >)? ' 6-$' '
>>;69/)
@$ AA91,B >>;# C(58D E#F 62
SLIDE 12
12
'7
- Cheaper (and easier to mass produce)
- Smaller
- Reduces power
- Keeps everything contained
– Reduces noise – Reduces coupling from the environment
- Need a large number of transistors to perform
real-world computations/tasks
- Allows a high density or circuit elements
(therefore, VLSI reduces costs)
SLIDE 13
13
522&55
Use more power (Large currents >mA) Efficient (Small currents pA-mA) Power Concern Can more easily match/replace Difficult to deal with Major concern Stuck with whatever was fabricated
- ex. 50% mismatch is not uncommon
Matching Exist, but rarely affect performance (Large size of devices and currents) Very big concern Seriously alter system performance Parasitics Use when needed Only feasible for very high frequencies Extremely expensive Inductors Easy to Use Cheap Mostly bad Very expensive (large real estate) Resistors Large
- ex. Capacitors 100pF-100F
Relatively Small
- ex. Capacitors 10fF-10pF
Device Size and Values
Discrete Analog Analog VLSI
SLIDE 14
14
;BG
Good Things about Analog VLSI
- Inexpensive
- Compact
- Power Efficient
Not So Good Things about Analog VLSI (not necessarily bad)
- Limited to transistors and capacitors (and sometimes resistors
if a very good reason)
- Parasitics and device mismatch are big concerns
- You are stuck with what you built/fabricated (no swapping parts
- ut)
However, Neuromorphic Analog VLSI is all about how to cope with these “problems,” how to get around them, and how to use them as an advantage
SLIDE 15
15
/
We will limit our discussion to CMOS technologies
- No BJTs
- Only MOSFETs
Therefore, we will discuss only silicon processes
SLIDE 16
16
6''G
- Begin at MOS device physics
- Look at circuits using the device properties
- Building small systems from circuits