0% found this document useful (0 votes)
38 views43 pages

PD Flow

The document outlines the process design flow in VLSI, highlighting the evolution of chip design from the Intel 4004 to modern processors like the Itanium 2. It discusses the motivation behind design automation and the importance of partitioning complex design problems into manageable sub-problems. The syllabus includes various lectures on topics such as Verilog, logic synthesis, and design for test.

Uploaded by

coderpro1412
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
38 views43 pages

PD Flow

The document outlines the process design flow in VLSI, highlighting the evolution of chip design from the Intel 4004 to modern processors like the Itanium 2. It discusses the motivation behind design automation and the importance of partitioning complex design problems into manageable sub-problems. The syllabus includes various lectures on topics such as Verilog, logic synthesis, and design for test.

Uploaded by

coderpro1412
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 43

Selected Topics In VLSI

Lecture 1
Process Design Flow
Introduction
1 4
Motivation Chip Design

Motivation and
Introduction
Motivation

1971 -The Intel 4004 2006 - Itanium 2 "Montecito"


2,300 Transistors 1.7B Transistors

© Adam Teman, 2018


Front End r truction
C.he1g Ll Instruction Cache
I Br anch LOP Cache
1gJ
32KB 8-Way Instruction
TLB
Introduced May 2016
Motivation
Predicirtion
" . 1 6 Bytes
Technology 14nm FinFET
Instruction Fetch
I & FreDecode I Memory 320KB L 1-1$
Core i7-6950X Extreme Edition (Broadwell-E ) . 6.ii..de
320KB L 1-D$
Instruction Queue [
4501, 2 2 5 «ntnes1
sir·a.ct.ts 2.56MB L2$

e#lrtr. i=
] w t a o s
M noFudonL cF n o union, y 25MB L3$
c c r Cc d r ~c t r d r

·l.ors Cores 10
Queue, U ncore, I/O orc ac h e ] l W leotier (uu ,6, 228 s
20
(1.5k O Ps; 8-Way )/ 3 z8 y e 1
Threads
ff j
Fe na r e ! All o c a t e ' R e t ir e m e n t Frequency 3.0-3.5 GHz
»] Reorder Butter 4152 ntnesI
2 r ci r q b i c n t i

I I~ I 1a I I & i•· Die Size 246 mm 2


cuter
r i f f fr%rut rt Sb t i n 8.54
[tr cut sewerft]
( t 6 a qr+bars
¢~-4 ururtr%f
]weer trot sawe rr
4t68 saw&rib rs
ft] #Transistors 3.2 B
ML& S itt
vwwwrwwr LJ
LE
LJ u & S itt llLcad A d r s l l Lca d A a r s l l
B rr lln r A u e l[S i r A u e r l
t r
Cata
lHb r
L
I
Auter\l
~
J
%
co A
i r
J
% c

i ·--
Bra r d LEA
5
Shared . "
-
n AM 4 h r4

core 5 4 #
e

M L
e

[ } 4 2
v e

#Aue
let A L

»
"
L3 Cache wet Irt. Mu.l 'w e t r t . ALLJ

- ·-
u m
c t r Lue
weterSitt.
t r Luc

5 t o r e B u f f e r & Fr a r "
(42 entrie s )

32B/Cycle store
-
Data TLB] -
Ll Data Cache

2 32B/Cycle load
lies
l72 entries
efy 32KB 8-Way 64B/Cycle -

7.co
«
Memory

con«a" NEE Execution Engine

http://en.wikichip.org/wiki/intel/core_iZee/i7-6950x Adam Teman, 2018


Motivation
• Houston, we have a problem...
10,000 100,000

1,000 10,000
0
c
O 100 1000
I,....
Q)
O .
en 10 100
en c
I,....
0
.OG- .•.--- 1 10
E
c c
( U .- - 0.1 1

w bl

O
o 0.01 roductivit 0.1
0

" 0.001 0.01


II IIIIIIIIIIIIIIIII IIII III • _ . c [
,-,
0o
O
,-,
oz
~
g
5
~ ~
2
~
5
~ ~
5
~
t
kg
~ ~
S s s = =
i i

7
"Moore's Law of Engi.neers " © Adam Teman, 2018
Motivation
The Solution:

Design Design
Automation Re-use (IP)
Syllabus
• Lecture 1: Introduction
• Lecture 2: Verilog
• Lecture 3: Logic Synthesis
• Lecture 4: Static Timing Analysis
• Lecture 5: Moving to the
Physical Domain
• Lecture 6: Placement
• Lecture 7: Clock Tree Synthesis
• Lecture 8: Routing
• Lecture 9: 1/0 and Packaging
• Lecture 10: Design for Test
1 2 4
Motivation Building a Chip Design
Chip

Building a Chip
General Design Approach
• How do engineers build a bridge?

• Divide and conquer!!!!


• Partition design problem into many sub-problems, Partition
which are manageable
• Define mathematical model for sub-problem Model/Solution
and find an algorithmic solution
• Beware of model limitations and check them !!!!!!! J
• Implement algorithm in individual design tools, define Tools/Interfaces
and implement general interfaces between the tools
• Implement checking tools for boundary conditions Verify/Validate
• Concatenate design tools to general design flows J
which can be managed
• See what doesn't work and start over. Develop Flow
13 © Adam Teman, 2018
VLSI System Design Flow
VLSI Design
GDS (Graphic Data Stream) is a file that was developed by calma company in the year 1971
and the GDS II in the year 1978. It is a binary file format that represents layout data in a
hierarchical format. Data such as labels, shapes, layer information and other 2D and 3D layout
geometric data.
System Level

Gate Level

Transistor Level

Layout Level

Mask Level
1 3 4
Motivation Design Chip Design
Automation

Design Automation
The (really) Olden Days

Schematic of Intel 4004 (1971)

Mainframe CAD System


21 http://www.computerhistory.org/revolution/digital-logic
The (really) Olden Days
• Early chips were prepared entirely by hand:
.-..

. a t

Hand drawn gate layout (Fairchild)

Rubylith Operators ( 1970) The original Tape-Out?


I NTEL BD D - EP n- 0 10 -18 - 6
http://www.computerhistory.org/revolution/digital-logic
8088A Mask Transparent Overlays (1976)
Design Automation Today
EDA tools
Questasim from Mentor Graphics, VCS
from synopsis and IES (ncsim) from
Cadence
1 4
Motivation Chip Design
Flow

Chip Design Flow


How a chip is built

26
Definition & Planning
Definition and Planning

Physical Design - Backend Flow Design and Verification

Logic Synthesis
RTL
Physical Design
SDC
Signoff and Tapeout
Standard Cells and GTL with Scan
Silicon Validation
Macros

Scan Chains Placed Design

Floorplan Design with


Clock Tree
Power Grid,
S ecial Routing
Router Routed Design
Clock Definitions

GDSII

You might also like