design-compiler.pdf

Synthesis &
Synthesis & Gate
Gate-
-Level
Level
Simulation
Simulation
REF:
• CIC Training Manual – Logic Synthesis with Design Compiler, July, 2006
• TSMC 0.18um Process 1.8-Volt SAGE-XTM Stand Cell Library Databook, September, 2003
• TPZ973G TSMC 0.18um Standard I/O Library Databook, Version 240a, December 10, 2003
9 3G S C 0 8u Sta da d /O b a y ataboo , e s o 0a, ece be 0, 003
• Artisan User Manual
Advanced Reliable
Systems (ARES) Lab.
1

12/8
 課程主題: Synthesis & Gate-level Netlist Simulation
 學習目標
 電路合成觀念
 Edit your synthesis script file
 Familiar with the Design Compiler
g p
 Gate-level netlist simulation
 LAB 2簡介 – 合成simple 8-bit microprocessor
 由於合成的每個步驟，都可透過執行不同的script指令去完成，像是設定的
 由於合成的每個步驟，都可透過執行不同的script指令去完成，像是設定的
各個constraints (ex: timing, area, etc…)。合成時，只需將script丟給
Design Compiler即可完成合成的步驟。因此，學習編輯自己的script file將
可大幅加速合成的步驟。
 步驟一: 本實驗將會給定各個合成參數。只要依照參數去編輯script file並
執行合成即可。
 步驟二: 使用合成後的gate-level netlist跑simulation，並觀察波形。
Advanced Reliable Systems (ARES) Lab. 2

Outline
 Basic Concept of the Synthesis
 Synthesis Using Design Compiler
 Synthesis Using Design Compiler
 Simulation-Based Power Estimation Using PrimePower
 Artisan Memory Compiler
y p
 LAB

Basic Concept of the Synthesis
Basic Concept of the Synthesis

Cell-Based Design Flow
MATLAB/ C/ C++/ System C/
ADS/ Covergen (MaxSim)
Memory Generator
Spec.
System Level
ADS/ Covergen (MaxSim)
NC-Verilog/ ModelSim
Debussy (Verdi)/ VCS
Verilog/ VHDL Syntest
RTL Level
Design/ Power Compiler
DFT Compiler/ TetraMAX
mpiler/
Fusion
Conformal/
Formality
Logic Synthesis
Design for Test
NC-Verilog/ ModelSim
Debussy (Verdi)/ VCS
hysical
Com
gma
Blast
Gate Level
SOC Encounter/ Astro
DRC/ LVS (Calibre)
Ph
Mag
GDS II
Layout Level
Post-Layout
Verification
PVS: Calibre xRC/ NanoSim
(Time/ Power Mill)
Verification
Advanced Reliable Systems (ARES) Lab.
Tape Out
5

What is Synthesis
 Synthesis = translation + optimization + mapping
if(high_bits == 2’b10)begin
residue = state table[i];
_ [ ];
end
else begin
residue = 16’h0000;
end Translate (HDL Compiler)
HDL Source
HDL Source
(RTL)
Optimize + Mapping
(HDL Compiler)
No Timing Info.
(HDL Compiler)
Generic Boolean
(GTECT)
Timing Info.
The synthesis is constraint driven
Target Technology
and technology independent !!
6

Logic Synthesis Overview
RTL Design
HDL Design Ware
DW
Architecture
Compiler
Design Ware
Library
Developer
Optimization
Design
Compiler
Technology
Lib
Compiler
Logic
Optimization Compiler Library
p
Optimized
Optimization
Optimized
Gate-Level Netlist

Compile
RTL code
or netlist
Optimized Design (Gate-Level Netlist)
Compile
Attributes &
Constraints
Schematic
Reports (Timing, Area, Power, …, etc)
Constraints Reports
Flatten
Technology
Lib
Structure
( g, , , , )
Logic Level Optimization
Library
(Can be set by the GUI
interface or user-defined
Script File !!)
Gate Level Optimization
Map
Technology
Library
Library
8

Logic Level Optimization
 Operate with Boolean representation of a circuit
 Has a global effect on the overall area/speed
 Has a global effect on the overall area/speed
characteristic of a design
 Strategy
 Strategy
 Structure
 Flatten (default OFF)
( )
 If both are true, the design is “first flattened and then structured”
Ex:
f = acd + bcd +e
g = ae’ + be’
h = cde
f = xy + e
g = xe’
h = ye
f0 = at
f1 = d + t
f2 = t’e
f0 = ab + ac
f1 = b + c + d
f2 = b’c’e
h cde h ye
x = a + b
y = cd
(Structure)
f2 t e
t = b + c
f2 b c e
(Flatten)

Gate Level Optimization - Mapping
 Combinational Mapping
 Mapping rearranges components combining and re-combining
 Mapping rearranges components, combining and re-combining
logic into different components
 May use different algorithms such as cloning, resizing, or
b ff i
buffering
 Try to meet the design rule constraints and the timing/area goals
 Sequential Mapping
 Sequential Mapping
 Optimize the mapping to sequential cells technology library
 Analyze combinational logics surrounding a sequential cell to see
y g g q
if it can absorb the logic attribute with HDL
 Try to save speed and area by using a more complex sequential
cells
cells

Mapping
Combinational Mapping Sequential Mapping
a
a
a
b
a
b
c c
A
B D Q
AND_FF
Q
A
B
a
b
a
b
c c
a a c
c
x1 x1 x2 x4
D Q
A Q
A
B
Critical Path Critical Path
D Q
A
B
Loop_FF
Q
B
a f
g a
f
g
(assume g loading high)
g
11

Boundary Optimization
 Design Compiler can do some optimizations across boundaries
1. Removes logic driving unconnected output ports
2. Removes redundant inverters across boundaries
3. Propagates constants to reduce logic

Static Timing Analysis
 Main steps of STA
 Break the design into sets of timing paths
 Break the design into sets of timing paths
 Calculate the delay of each path
 Check all path delays to see if the given timing constraints are
met
 Four types of paths
( )
 Register - Register (Reg - Reg)
 Primary Input - Register (PI - Reg)
 Register - Primary Output (Reg - PO)
 Register - Primary Output (Reg - PO)
 Primary Input - Primary Output (PI - PO)

Static Timing Analysis (Cont’)
 To meet the setup time requirement:
 Trequire >= Tarrival
- Setup Time
(T l k > 0 denotes
require arrival
 Reg to Reg
 Tarrival = Tclk1 + TDFF1(clk->Q) + TPATH
 T T T
(Tslack > 0 denotes
“no timing violation”)
 Trequire = Tclk2 - TDFF2(setup)
 Tslack = Trequire - Tarrival
Clk_source
clk1
clk1
TDFF1 + Tpath
Tarrival
D Q
DFF1
D Q
DFF2
PATH
Tarrival
data
clk2
Trequire
Tsetup
Tslack
Q
clk1 Q
clk2

 PI to Reg
 Tarrival = TPI(delay) + TPATH
T T T
- Setup Time
 Trequire = Tclk1 - TDFF1(setup)
Clk_source
Tarrival
Tpath
D Q
DFF1
PATH
TPI(delay)
T
PI data
clk1
Trequire
Tslack
D Q
Q
clk1
Tsetup
Tarrival

 Reg to PO
T T T
- Setup Time
 Trequire = Tcycle - TPO(output delay)
Clk_source
clk1
clk1
TDFF1 + Tpath
T
D Q
DFF1
PATH
T
PO data
Tarrival
Trequire Tslack
D Q
Q
clk1
TPO(output delay)
Tarrival
Tslack
16

 PI to PO
 T = T - T
- Setup Time
 Trequire = Tcycle - TPO(output delay)
Clk_source
TPI + Tpath
Tarrival
PATH
T
PO
PI data
Trequire Tslack
TPO(output delay)
Tarrival
Tslack
17

 To meet the hold time requirement:
 Trequire <= Tarrival
- Hold Time
require arrival
 Reg to Reg
 T T + T
 Trequire = Tclk2 + TDFF2(hold)
 Tslack = Tarrival - Trequire
Clk source
Clk_source
clk1
TDFF1 + Tpath
clk2
D Q
DFF1
D Q
DFF2
PATH
Tarrival
data 1
data 0
clk2
T i l
Trequire
Thold
Tslack
Q
clk1 Q
clk2
Tarrival
18

 PI to Reg
- Hold Time
 Trequire = Tclk1 + TDFF(hold)
 Reg to PO
g
 Tarrival = Tclk1 + TDFF(clk->Q) + TPATH
 Trequire = - TPO(output delay)
 T l k = T i l - T i
 Tslack Tarrival Trequire
 PI to PO
 T T
 Trequire = - TPO(output delay)

Synthesizable Verilog
 Verilog Basis
 parameter declarations
p
 wire, wand, wor declarations
 reg declarations
 input output inout
 input, output, inout
 continuous assignment
 module instructions
 gate instructions
 always blocks
 task statement
 task statement
 function definitions
 for, while loop
 Synthesizable Verilog primitives cells
 and, or, not, nand, nor, xor, xnor
 bufif0, bufif1, notif0, notif1
, , ,
20

Synthesizable Verilog (Cont’)
 Operators
 Binary bit-wise ( ~, &, |, ^, ~^ )
y ( | )
 Unary reduction ( &, ~&, |, ~|, ^, ~^ )
 Logical ( !, &&, || )
 2’s complement arithmetic ( + * / % )
 2 s complement arithmetic ( +, -, , /, % )
 Relational ( >, <, >=, <= )
 Equality ( ==, != )
 Logic shift ( >>, << )
 Conditional ( ?: )
 Concatenation ( { } )
 Concatenation ( { } )

Notice Before Synthesis
 Your RTL design
Area
Cycle
Better
 Functional verification by some high-level language
 Also, the code coverage of your test benches should be verified (i.e. VN)
 Coding style checking (i.e. n-Lint)
Time
 Coding style checking (i.e. n Lint)
 Good coding style will reduce most hazards while synthesis
 Better optimization process results in better circuit performance
 E d b i f h i
 Easy debugging after synthesis
 Constraints
 The area and timing of your circuit are mainly determined by your
 The area and timing of your circuit are mainly determined by your
circuit architecture and coding style
 There is always a trade-off between the circuit timing and area
 In fact, a super tight timing constraint may be worked while synthesis,
but failed in the Place & Route (P&R) procedure

Synthesis Using Design Compiler
Synthesis Using Design Compiler

<.synopsys_dc.setup> File
 Create individual synopsys setup file for each folder
 link_library : the library used for interpreting input description
 Any cells instantiated in your HDL code
y y
 Wire load or operating condition modules used during synthesis
 target_library : the ASIC technology which the design is mapped
 f
 symbol_library : used for schematic generation
 search_path : the path for unsolved reference library
 synthetic path : designware library
 synthetic_path : designware library

<.synopsys_dc.setup> File (Cont’)
 MEMs libraries are also included in this file
Ex:
MEM Libraries (.db file)
Note that the MEM DB files are converted from
the LIB files which are generated from the Artisan !!
(.synopsys_dc.setup File)
the LIB files which are generated from the Artisan !!

Hard Macro
 Memory block
 Memory library files (synopsys model) are generated by memory
il
compiler
 Translate library files (.lib) to db files (.db) for synthesis
 Four corners: fast@-40C fast@0C typical and slow
 Four corners: fast@ 40C, fast@0C, typical, and slow
Memory block

Settings for Using Memory
 Convert *.lib to *.db
 > dc shell
any memory LIB file
 > dc_shell
 dc_shell-t> read_lib t13spsram512x32_slow_syn.lib
 dc_shell-t> write_lib t13spsram512x32 -output
t13spsram512x32_slow_syn.db
 Modify <.synopsys_dc.setup> File:
“*
user library name, which should
be the same as the library name
in the Artisan
 set link_library “* slow.db t13spsram512x32_slow.db
dw_foundation.sldb”
 set target library “slow db t13spsram512x32 slow db”
memory DB file add to the file
 set target_library slow.db t13spsram512x32_slow.db
 Before the synthesis, the memory HDL model should be
blocked in your netlist
y

Test Pins Reservation
 You can add the floating test pins to your design before synthesis
 se: scan enable
 si: scan input
 so: scan output
 scantest: control signal for memory shadow wrapper (i e memory is used)
 scantest: control signal for memory shadow wrapper (i.e. memory is used)
 Ex:
Normal IO Declaration
Test IO Declaration
 The pins will be connected to scans after the scan chain insertion
p
28

CHIP-Level Netlist
 The CHIP-level netlist consists of your Core-level netlist and the
PADs
E CHIPv
CORE
CORNER2
CORNER1
Ex: CHIP.v
CORE.v
(CHIP-Level Declaration)
CORE
I_CLK
CLK
CORNER3 CORNER4
O_CSO
CSO
CORNER3 CORNER4
(CORE-Level Design)
(Input PAD) I_CLK CLK
PAD C
(Output PAD) CSO O_CSO
I PAD
(IO PAD D l ti )
(Output PAD) (IO PAD Declaration)
29

How To Choose the IO PAD
 You can reference the Databook of the IO PAD in CIC Design Kit
 Generally, the “PDIDGZ” is used as the input PAD
 Trade-off when considering
the output PAD
 High driving  SSN
Ex:
 High driving  SSN
 Low driving  Delay
 Note that the loading of the
 Note that the loading of the
CIC tester is 40pf
[REF: TPZ973G TSMC 0.18um Standard
I/O Library Databook, Version 240a,
December 10, 2003]
December 10, 2003]
30

Synthesis Flow
DFT Insertion
Design Import
Setting Design
Environment
Setting Clock
Compile After
DFT
Assign Violation
Setting Clock
Constraints
Setting Design
Assign Violation
Avoidance
Naming Rule
Rule Constraints
Compile the
Design
Changing
Save Design
Design

Getting Started
 Prepare Files:
 *.v files
 *.db files (i.e. memory is used)
S h i i fil (i d ib d l )
 Synthesis script file (i.e. described later)

linux %> dv
Tool Bar
Logic Hierarchy
Vi
View
Log Window
(GUI view of the Design Vision)
Command Line

Read File Design Import
 Read netlists or other design descriptions into Design Compiler
 File/Read
 Supported formats
 Verilog: .v
 VHDL: .vhd
 System Verilog: .sv
 EDIF
 PLA (Berkeley Espresso): .pla
 Synopsys internal formats:
 DB (binary): db
 DB (binary): .db
 Enhance db file: .ddc
 Equation: .eqn
 State table: st
 State table: .st
read_file -format verilog file name.v (“/usr/LAB_DV/syn/cpu.v”)
read file -format ddc file name.ddc
{ Command Line }
read_file format ddc file name.ddc
33

PAD Parameters Extraction
 Input PAD
 Input delay
p y
 Input driving
 Output PAD
 O t t d l
 Output delay
 Output loading
CHIPv
CORE.v
(delay, loading)
(delay, driving) (chip_const.tcl)
{ Command Line }
CHIP.v
current_design CHIP
characterize [get_cells CORE]
current_design CORE
write script -format dctcl -o chip const.tcl
write_script format dctcl o chip_const.tcl
34

Uniquify
 Select the most top design of the hierarchy
 Hierarchy/Uniquify/Hierarchy
uniquify
{ Command Line }
(Design View) (Log Window)
uniquify
35

Design Environment Setting Design
Environment
 Setting Operating Environment
 Setting Don’t Used Cells
 Setting Don t Used Cells
 Setting Input Driving Strength
 Setting Output Loading
 Setting Output Loading
 Setting Input/Output Delay
 Setting Wire Load Model
 Setting Wire Load Model

Setting Operating Condition
 Attributes/Operating Environment/Operating Conditions
Setup/Hold time is evaluated
set_operating_conditions -max “slow” -max_library “slow” -min “fast”
-min library “fast”
{ Command Line }
min_library fast
37

Setting Don’t Used Cells
 Some standard cells are for special purposes
 Buffering clock
 Buffering clock
 CLKBUF*
 Inverting clock
 CLKINV*
 Creating delay (delay cell)
 Tie cells
 Tie cells
 TIE*
 DFT cells
 SDF* and SEDF*
{ Command Line }
set_dont_use slow/CLKBUF*
set_dont_use typical/CLKBUF*
set_dont_use fast/CLKBUF*
{ Command Line }

Setting Drive Strength/Input Delay for PADs
 Assume that we use the input PAD “PDIDGZ”
my design
PAD C
FF
Q
D
b
Input PAD
(PDIDGZ)
(PDIDGZ)
set_drive [expr 0.288001] [all_inputs]
set input delay [expr 0.34] [all inputs]
{ Command Line }
_ p _ y [ p ] [ _ p ]
39

Setting Load/Output Delay for PADs
 Assume that we use the output PAD “PDO24CDG”
my design
FF
Q
D d
clk
Output PAD PAD
I
(PDO24CDG)
clk
OEN
(PDO24CDG)
{ Command Line }
set_load [expr 0.06132] [all_outputs]
set_output_delay [expr 2] [all_outputs]

Setting Wire Load Model
 Attributes/Operating Environment/Wire Load
Recommend
(Worst Case)
set_wire_load_model -name “tsmc18_wl10” -library “slow”
set wire lode mode “top”
{ Command Line }
set_wire_lode_mode top
41

Clock Constraints Setting Clock
Constraints
 Period
 Waveform
 Waveform
 Uncertainty
 Skew
 Skew
 Latency
 Source latency
 Source latency
 Network latency
 Transition
 Input transition
 Clock transition
 Combination Circuit – Maximum Delay Constraints

Sequential Circuit  Specify Clock
 Select the “clk” pin on the symbol
 Attributes/Specify Clock
 set fix hold: respect the hold time
 set_fix_hold: respect the hold time
requirement of all clocked flip-flops
 set_dont_touch_network: do not re-buffer
th l k t k
{ Command Line }
the clock network
set sys_clk clk1
creat_clock -name $sys_clk -period 10 $sys_clk
set_dont_touch_network $sys_clk
t fi h ld $ lk
{ Command Line }
set_fix_hold $sys_clk
43

Setting Clock Skew
 Different clock arrival time
Ex:
FF
clk
FF
FF
FF
 experience
 Small circuit: 0.3 ns
 Large circuit: 0 5 ns (Ti i R t)
set_clock_uncertainty -setup 0.6 $sys_clk
set clock uncertainty -hold 0.5 $sys clk
{ Command Line }
 Large circuit: 0.5 ns (Timing Report)
set_clock_uncertainty hold 0.5 $sys_clk
44

Setting Clock Latency
 Source latency is the propagation time from the actual clock origin to
the clock definition point in the design
 This setting can be avoid if the design is without the clock generator
Ex:
Your Design
Origin of Clock
Source Latency
3ns
 experience
 Small circuit: 1 ns
 Large circuit: 3 ns
Source Latency
set clock latency 1 [get clocks clk]
{ Command Line }
 Large circuit: 3 ns
set_clock_latency 1 [get_clocks clk]
45

Setting Ideal Clock
 Since we usually let the clock tree synthesis (CTS)
procedure performed in the P&R (i e
procedure performed in the P&R (i.e.
set_dont_touch_network), the clock source driving
capability is poor
 Thus, we can set the clock tree as an ideal network
without driving issues
 Avoid the hazard in the timing evaluation
set ideal network [get clocks clk]
{ Command Line }
set_ideal_network [get_clocks clk]
46

Setting Don’t Touch Macro
 Modules have been synthesized/optimized
 S t d t t h t id ti i th ith
 Set dont_touch to avoid optimize the macro with
other modules
Synthesized
core
core
set dont touch module name
{ Command Line }
set_dont_touch module_name

Setting Clock Transition
set_clock_transition
create_clock
CLK
 experience
 < 0.5ns
 CIC tester: 0 5 ns
set input transition -max 0.1 $sys clk
{ Command Line }
 CIC tester: 0.5 ns
set_input_transition max 0.1 $sys_clk
48

Combination Circuit – Maximum Delay
Constraints
 For combinational circuits primarily (i.e. design with no clock)
 Select the start & end points of the timing path
 Attributes/Optimization Constraints/Timing Constraints
 Attributes/Optimization Constraints/Timing Constraints
Ex:
Maximum
Maximum
Delay
Constraint
(5ns = 200 MHz)
Minimum
Delay
Delay
Constraint

Design Rule Constraints Setting Design
Rule Constraints
 Area Constraint
 Fanout Constraint
 Fanout Constraint

Setting Area/Fanout Constraint
 Attributes/Optimization
Constraints/Design
Constraints
 If you only concern the circuit
area but don’t care about the
area, but don t care about the
timing
 You can set the max area
constraints to 0
constraints to 0
set_max_area 0
set max fanout 32 [get designs CORE]
{ Command Line }
set_max_fanout 32 [get_designs CORE]
51

Compile the Design Compile the
Design
 Design/Compile Design
compile -map_effort high -boundary_optimization
compile ultra -no autoungroup (time consuming)
{ Command Line }
compile_ultra -no_autoungroup (time consuming)
52

Example for DFT Insertion DFT Insertion
####DB mode####
set_dft_configuration -autofix
set dft configuration -shadow wrapper
DB
Mode
{ Command Line }
set_dft_configuration shadow_wrapper
set_scan_configuration -style multiplexed_flip_flop
set_scan_configuration -clock_mixing no_mix
set_scan_configuration -methodology full_scan
set scan signal test scan in -port si
Mode
set_scan_signal test_scan_in port si
set_scan_signal test_scan_out -port so
set_scan_signal test_scan_enable -port se
set_dft_signal test_mode -port scantest
set test hold 0 rst
set_test_hold 0 rst
set_test_hold 1 scantest
set_test_hold 1 se
create_test_clock -period 100 -waveform [list 40 60] [find port "clk"]
set port configuration -cell RA1SHD256x8 -clock clk
set_port_configuration cell RA1SHD256x8 clock clk
set_port_configuration -cell RA1SHD256x8 -port "Q" -tristate -read {"OEN" 0} -clock clk
set_port_configuration -cell RA1SHD256x8 -port "A" -write {"WEN" 0} -clock clk
set_port_configuration -cell RA1SHD256x8 -port "D" -write {"WEN" 0} -clock clk
set wrapper element RA1SHD256x8 -type shadow
set_ appe _e e e t S 56 8 type s ado
set_wrapper_element FJU_MEM -type shadow
set_fix_multiple_port_nets -all -constants -buffer_constants [get_designs *]
insert_dft

Example for DFT Insertion (Cont’)
####XG mode####
create_port –dir in scan_in XG
Mode
{ Command Line }
create_port –dir out scan_out
create_port –dir in scan_en
compile –scan –boundary_optimization
set scan configuration –internal clocks single –chain count 1 –clock mixing no mix
Mode
set_scan_configuration –internal_clocks single –chain_count 1 –clock_mixing no_mix
set_dft_signal –view exist –type TestClock –timing {45 55} –port {clk}
set_dft_signal –view exist –type Reset –active 1 –port reset
set_dft_signal –view spec –type ScanEnable –port scan_en –active 1
g y
set_dft_signal –view spec –type ScanDataIn –port scan_in
set_dft_signal –view spec –type ScanDataOut –port scan_out
set_scan_path chain1 –view spec –scan_data_in scan_in –scan_data_out scan_out
create_test_protocol
dft_drc
preview_dft –show scan_clocks
set false path –from [get ports scan en]
set_false_path –from [get_ports scan_en]
insert_dft

Compile After DFT Compile After
DFT
compile -scan
check_scan
{ Command Line }
report_test -scan_path
estimate_test_coverage
 The fault coverage will be shown as below:
 The fault coverage will be shown as below:

Assign Problem Assign Violation
Avoidance
 The syntax of “assign” may cause problems in the LVS
assign A[19] = A[19];
assign A[18] = A[18];
BUFX1 X37X( .I(A[19]), .Z(ABSVAL[19]) );
assign A[17] = A[17];
assign A[16] = A[16];
assign A[15] = A[15];
assign ABSVAL[19] = A[19];
g [ ] [ ]
set_fix_multiple_port_nets -all -constants -buffer_constants [get_designs *]
{ Command Line }

Floating Port Removing
 Due to some ports in the standard cells are not used in
your design
your design
remove_unconnected_ports -blast_buses [get_cells -hierarchical *]
{ Command Line }

Chang Naming Rule Script Naming Rule
Changing
 Purpose: Let the naming-rule definitions in the gate-level netlist are
the same as in the timing file (e.g. *.sdf file)
 Also, the wrong naming rules may cause problems in the LVS
{ Command Line }
set bus_inference_style {%s[%d]}
set bus_naming_style {%s[%d]}
set hdlout_internal_busses true
{ Command Line }
change_names -hierarchy -rule verilog
define_name_rules name_rule -allowed "A-Z a-z 0-9 _" -max_length 255 -type cell
define_name_rules name_rule -allowed "A-Z a-z 0-9 _[]" -max_length 255 -type net
define name rules name rule -map {{"*cell*""cell"}}
define_name_rules name_rule -map {{ cell cell }}
define_name_rules name_rule -case_insensitive
change_names -hierarchy -rules name_rule

Save Design Save Design
 Five design files:
 *.spf: test protocol file for ATPG tools (i.e. TetraMax)
 *.sdc: timing constraint file for P&R
 *.vg: gate-level netlist for P&R
 * sdf: timing file for Verilog simulation
 .sdf: timing file for Verilog simulation
 *.ddc: binary file (i.e. all the constraints and synthesis results are
recorded)
{ C d Li }
REPORT WORK syn tbench script
write -f ddc -o [format "%s%s" [format "%s%s" "../WORK/" $TOP_BLOCK] ".ddc"] -
no_implicit -hier
write -f verilog -o [format "%s%s" [format "%s%s" "../WORK/" $TOP_BLOCK] “_syn.v"] -
{ Command Line }
g [ [ _ ] _ y ]
no_implicit -hier
write_sdc [format "%s%s" [format "%s%s" "../WORK/" $TOP_BLOCK] ".sdc"]
write_sdf -version 2.1 -context verilog [format "%s%s" [format "%s%s" "../WORK/"
$TOP BLOCK] “ sdf“]
$TOP_BLOCK] .sdf ]
59

Synthesis Report
 Report Design Hierarchy
 Report Area
 Report Area
 Design View
 Report Timing
 Report Timing
 Critical Path Highlighting
 Timing Slack Histogram
 Timing Slack Histogram

Report Design Hierarchy
 Hierarchy report shows the component used in your each block & its
hierarchy
 Design/Report Design Hierarchy
Ex:

Report Area
 Design/Report Area
2
Ex: (0.18um Cell-Library: 1 gate ≈ 10 um2)
(0.13um Cell-Library: 1 gate ≈ 5 um2)
Ex:
(um2)

Design View
 List/Design View
All the block area are listed !!
Ex:
Ex:

Report Timing
 Timing/Report Timing
setup time
Critical Path
Ex:
max: setup time
min: hold time
Slack = Data Require Time – Data Arrival Time

Critical Path Highlighting
 View/Highlight/Critical Path
Ex:
Ex:

Timing Slack Histogram
 Timing/Endpoint Slack Totally 190 paths are in the slack range between
0 to 1.78
Ex:
Resolution

Edit Your Own Script File
 For convenient, you should edit your own synthesis script file.
Whenever you want to synthesis a new design, you just only change
some parameters in this file.
 Execute Script File
 Setup/Execute Script
Ex:
 Setup/Execute Script
 Or use “source your_
script.dc” in dc shell
p _
command line

Gate-Level Simulation
 Include the Verilog model of standard cell and gate-level netlist to
your test bench Standard Cell Library
 Add the following Synopsys directives to the test bench
Gate- Level Netlist
 Add the following Synopsys directives to the test bench
*.sdf File Instance Name
Delay

LAB
LAB

design-compiler.pdf

More Related Content

design-compiler.pdf