VLSI Digital Circuits

Lecture 15: Timing Issues

[Adapted from Prof. Mary Jane Irwin’s slides, Rabaey’s Digital Integrated
Circuits, Second Edition, ©2003 J. Rabaey, A. Chandrakasan, B. Nikolic]
Timing Classifications
‰ Synchronous systems
z All memory elements in the system are simultaneously updated
using a globally distributed periodic synchronization signal (i.e.,
a global clock signal)
z Functionality is ensure by strict constraints on the clock signal
generation and distribution to minimize
- Clock skew (spatial variations in clock edges)
- Clock jitter (temporal variations in clock edges)

‰ Asynchronous systems
z Self-timed (controlled) systems
z No need for a globally distributed clock, but have asynchronous
circuit overheads (handshaking logic, etc.)

‰ Hybrid systems
z Synchronization between different clock domains
z Interfacing between asynchronous and synchronous domains
Review: Synchronous Timing Basics
R1 R2
In Combinational
D Q D Q
logic

tclk1 tclk2
clk
tc-q, tsu, tplogic, tcdlogic
thold, tcdreg
‰ Under ideal conditions (i.e., when tclk1 = tclk2)
T ≥ tc-q + tplogic + tsu
thold ≤ tcdlogic + tcdreg
‰ Under real conditions, the clock signal can have both
spatial (clock skew) and temporal (clock jitter) variations
z skew is constant from cycle to cycle (by definition); skew can be
positive (clock and data flowing in the same direction) or negative
(clock and data flowing in opposite directions)
z jitter causes T to change on a cycle-by-cycle basis
Sources of Clock Skew and Jitter in Clock Network
4 power supply
3 interconnect
clock 6 capacitive load
1
generation
7 capacitive
PLL coupling
2 clock drivers

5 temperature

‰ Skew ‰ Jitter
z manufacturing device z clock generation
variations in clock drivers z capacitive loading and
z interconnect variations coupling
z environmental variations z environmental variations
(power supply and (power supply and
temperature) temperature)
Positive Clock Skew
‰ Clock and R1 R2
data flow in In D Q
Combinational
D Q
the same logic
direction tclk1 tclk2
clk
delay
T
T+δ
1 3
δ>0

2 4

δ + thold
T: T + δ ≥ tc-q + tplogic + tsu so T ≥ tc-q + tplogic + tsu - δ
thold : thold + δ ≤ tcdlogic + tcdreg so thold ≤ tcdlogic + tcdreg - δ
‰ δ > 0: Improves performance, but makes thold harder to
meet. If thold is not met (race conditions), the circuit
malfunctions independent of the clock period!
Negative Clock Skew
‰ Clock and R1 R2
data flow in In D Q
Combinational
D Q
opposite logic
directions tclk1 tclk2
clk
delay
T
T+δ
1 3

2 4
δ<0

T: T + δ ≥ tc-q + tplogic + tsu so T ≥ tc-q + tplogic + tsu - δ

thold : thold + δ ≤ tcdlogic + tcdreg so thold ≤ tcdlogic + tcdreg - δ

‰ δ < 0: Degrades performance, but thold is easier to meet

(eliminating race conditions)
Clock Jitter
‰ Jitter causes T to
vary on a cycle-by- R1
cycle basis Combinational
In logic
tclk
clk
T

-tjitter +tjitter

T : T - 2tjitter ≥ tc-q + tplogic + tsu so T ≥ tc-q + tplogic + tsu + 2tjitter

‰ Jitter directly reduces the performance of a sequential

circuit
Combined Impact of Skew and Jitter
‰ Constraints R1 R2
on the In D Q
Combinational
D Q
minimum logic
clock period tclk1 tclk2
(δ > 0) T
T+δ
1
δ>0

6 12
-tjitter

T ≥ tc-q + tplogic + tsu - δ + 2tjitter thold ≤ tcdlogic + tcdreg – δ –

2tjitter
‰ δ > 0 with jitter: Degrades performance, and makes thold
even harder to meet. (The acceptable skew is reduced
by jitter.)
Clock Distribution Networks
‰ Clock skew and jitter can ultimately limit the performance
of a digital system, so designing a clock network that
minimizes both is important
z In many high-speed processors, a majority of the dynamic power
is dissipated in the clock network.
z To reduce dynamic power, the clock network must support clock
gating (shutting down (disabling the clock) units)

‰ Clock distribution techniques

z Balanced paths (H-tree network, matched RC trees)
- In the ideal case, can eliminate skew
- Could take multiple cycles for the clock signal to propagate to the
leaves of the tree
z Clock grids
- Typically used in the final stage of the clock distribution network
- Minimizes absolute delay (not relative delay)
 H-Tree Clock Network
‰ If the paths are perfectly balanced, clock skew is zero

Can insert clock gating at

multiple levels in clock tree
Can shut off entire subtree
if all gating conditions are
Clock satisfied

Idle
condition
Gated
Clock clock
 Clock Grid Network
‰ Distributed buffering reduces absolute delay and makes
clock gating easier, but is sensitive to variations in the
buffer delay
‰ The secondary buffers
isolate the local clock
local logic nets from the upstream
area load and amplify the
clock signals degraded
by the RC network
Clock
z decreases absolute skew
main clock z gives steeper clocks
buffer
‰ Only have to bound the
skew within the local
secondary clock buffers logic area
DEC Alpha 21164 (EV5) Example
‰ 300 MHz clock (9.3 million transistors on a 16.5x18.1
mm die in 0.5 micron CMOS technology)
z single phase clock

‰ 3.75 nF total clock load

z Extensive use of dynamic logic

‰ 20 W (out of 50) in clock distribution network

‰ Two level clock distribution
z Single 6 inverter stage main clock buffer at the center of the
chip
z Secondary clock buffers drive the left and right sides of the
clock grid in m3 and m4

‰ Total equivalent driver size of 58 cm !!

Secondary Clock Buffers
Clock Skew in Alpha Processor
‰ Absolute skew smaller than 90 ps

‰ The critical
instruction and
execution units all
see the clock within
65 ps
Dealing with Clock Skew and Jitter
‰ To minimize skew, balance clock paths using H-tree or
matched-tree clock distribution structures.
‰ If possible, route data and clock in opposite directions;
eliminates races at the cost of performance.
‰ The use of gated clocks to help with dynamic power
consumption make jitter worse.
‰ Shield clock wires (route power lines – VDD or GND – next to
clock lines) to minimize/eliminate coupling with neighboring
signal nets.
‰ Use dummy fills to reduce skew by reducing variations in
interconnect capacitances due to interlayer dielectric
thickness variations.
‰ Beware of temperature and supply rail variations and their
effects on skew and jitter. Power supply noise fundamentally
limits the performance of clock networks.
Major Components of a Computer

Processor Devices

Control Input
Memory

Datapath Output

‰ Modern processor architecture styles (CSE 431)

z Pipelined, single issue (e.g., ARM)
z Pipelined, hardware controlled multiple issue – superscalar
z Pipelined, software controlled multiple issue – VLIW
z Pipelined, multiple issue from multiple process threads -
multithreaded
Basic Building Blocks
‰ Datapath
z Execution units
- Adder, multiplier, divider, shifter, etc.
z Register file and pipeline registers
z Multiplexers, decoders

‰ Control
z Finite state machines (PLA, ROM, random logic)

‰ Interconnect
z Switches, arbiters, buses

‰ Memory
z Caches, TLBs, DRAM, buffers
MIPS 5-Stage Pipelined (Single Issue) Datapath
Fetch Decode Execute Memory WriteBack

0
pipeline
stage
isolation
Add register
Add
4 Shift
left 2
Read Addr 1
I$ D$
Register Read Address
Read Addr 2Data 1

Dec/Exec
IF/Dec

Read
PC

File Read

Exec/Mem

Mem/WB
Address Write Addr ALU 1
Read Data
0
Data 2 Write Data 0
Write Data
1

Sign
16 Extend 32
Icache Dcache
precharge precharge

RegWrite
clk
Datapath Bit-Sliced Organization
Control Flow

Pipeline Register
Bit 3

Pipeline Register

Pipeline Register

Pipeline Register
Register File

Multiplexer
Multiplexer
Bit 2

Shifter
Adder
From
I$ Bit 1

Bit 0

Data Flow To/From D$

Tile identical bit-slice elements