Very Large Scale Integration

VLSI Design
From Specification to Silicon

VLSI (Very Large Scale Integration) is the process of creating integrated circuits by combining billions of transistors on a single chip. This section covers the complete ASIC design flow — from RTL coding and functional verification through physical design, timing closure, and tape-out — along with the key building blocks every chip designer must master.

ASIC Design Flow
RTL to GDSII
Industry Sign-Off Ready
CMOS Fundamentals Included
🧭 New · Career Guidance
Confused which VLSI domain to choose?
RTL, DV, PD, DFT, STA, analog, CAD — every role explained, with diagrams & a decision tree.
Open the roadmap →
▶ New · Animated
Watch a chip get built — the VLSI flow, animated
Press play and step through RTL → synthesis → floorplan → placement → CTS → routing → GDSII.
Play the flow →
All Topics

VLSI Design Knowledge Base

Topics are organized by design domain. Start with RTL fundamentals and progress through physical design and advanced techniques.

Front-End Design
RTL Design (Verilog / SystemVerilog)
Complete guide to RTL design — FSMs, pipelining, blocking vs non-blocking, synthesis-ready coding, and SystemVerilog constructs for ASIC design.
Explore RTL Design
Clock Domain Crossing (CDC)
Understanding and fixing CDC violations — 2-FF synchronizers, handshake protocols, async FIFO design with Gray code pointers, and CDC verification methodologies.
Explore CDC
Low Power RTL Design
Techniques for minimizing dynamic and static power at the RTL level — clock gating, operand isolation, power domains, multi-voltage design, and UPF-based power intent specification.
Explore Low Power RTL
Design for Testability (DFT)
DFT techniques that make chips manufacturable and testable — scan chain insertion, ATPG, BIST, boundary scan (JTAG), and how DFT constraints affect RTL coding and synthesis.
Explore DFT
RTL Building Blocks
D Flip-Flop Design
The fundamental storage element in synchronous digital design. Covers D flip-flop operation, synchronous vs asynchronous reset, enable signals, scan flops for DFT, and CMOS transistor-level implementation.
Explore D Flip-Flop
Latch Design & When to Use It
Latches vs flip-flops — understanding level-sensitive vs edge-triggered storage. Covers intentional latch design for timing optimization, inferred vs instantiated latches, and how to avoid accidental latch inference in RTL.
Explore Latch Design
Clock Gater (ICG) Design
Integrated Clock Gating (ICG) cells are the primary technique for reducing dynamic power in VLSI designs. Covers latch-based clock gaters, glitch-free clock gating, enable signal timing constraints, and library ICG cell implementation.
Explore Clock Gating
Glitch-Free Clock Mux Design
Switching between two clock sources without generating glitches on the output clock. Covers handshake-based glitch-free clock mux design, synchronization of select signals, and common RTL implementations used in power management controllers.
Explore Clock Mux
Reset Synchronizer Design
How to safely de-assert asynchronous resets in a synchronous manner. Covers the reset synchronizer circuit, why asynchronous assertion but synchronous de-assertion is the industry standard, and multi-domain reset distribution strategies.
Explore Reset Sync
Asynchronous FIFO Design
The canonical solution for passing data between asynchronous clock domains. Complete design guide covering Gray code pointer generation, full/empty flag logic, synchronizer stages, pointer comparison, and FIFO depth calculation methodology.
Explore Async FIFO
Binary to Gray Code Converter
Gray code (reflected binary) is essential for asynchronous FIFO pointers because only one bit changes per count — eliminating multi-bit transition glitches that can corrupt full/empty flag logic when pointers cross clock domains.
Explore Binary/Gray
FIFO Depth Calculation
How to calculate the required FIFO depth for a given burst traffic pattern — covering worst-case burst analysis, producer/consumer rate mismatch, latency budgeting, and the mathematical derivation used in real ASIC architecture planning.
Explore FIFO Depth
SystemVerilog Data Types
A complete reference for SystemVerilog data types used in RTL — logic vs bit vs reg, packed and unpacked arrays, structs, unions, enumerations, and the synthesis implications of each type for targeting real standard cell libraries.
Explore SV Data Types
Physical Design & Back-End
Physical Design Flow
The complete back-end flow from netlist to GDSII — floorplanning, power grid design, standard cell placement, clock tree synthesis (CTS), routing, DRC/LVS sign-off, and timing closure using industry tools like Innovus and ICC2.
Explore Physical Design
How AI Chips Work — Inside an NPU
What a Neural Processing Unit is, how systolic arrays run matrix multiplication, what quantization does, and the real silicon inside Apple, Qualcomm, Google and NVIDIA AI chips — with diagrams.
How TSMC Makes a Chip: From Sand to Silicon
The complete visual guide to semiconductor fabrication — silicon purification, wafer growth, EUV photolithography, FinFET and GAAFET transistors, BEOL metal interconnects, dicing, and advanced packaging. Every stage with diagrams.
CPU vs GPU vs TPU vs NPU — Architecture & Differences
Deep-dive comparison of four processor architectures. Core counts, parallelism models, memory bandwidth, power efficiency, systolic arrays, Tensor Cores, and which chip wins for AI training, gaming, HPC, and edge inference.
Metastability & CDC — Synchronizers & Async FIFO
Why signals go metastable when crossing clock domains, MTBF, the two/three-flop synchronizer, and safe asynchronous FIFO design for reliable clock-domain crossing.
Yosys show -format svg — Export a Schematic as SVG
Visualise your synthesized netlist as an SVG schematic using the Yosys show command — flags, the prefix option, and turning RTL into a clean circuit diagram.
Genus & DC Synthesis Commands Cheat Sheet
Complete Tcl command reference for Cadence Genus and Synopsys Design Compiler (dc_shell) — library setup, read_hdl/analyze, elaborate, compile_ultra, syn_generic/map/opt, report_timing, report_qor, report_area, write_hdl/netlist, and full flow scripts for both tools.
View Cheat Sheet
Floorplanning — Step by Step
A deep dive into IC floorplanning — die sizing, core utilization, IO pad placement, hard macro placement rules, power distribution network (PDN) design, placement blockages, and quality sign-off checks used in real tape-outs.
Explore Floorplanning
CMOS Technology & Device Physics
The transistor-level foundation of all digital chips — NMOS and PMOS operation, threshold voltage, channel length modulation, CMOS inverter analysis, static and dynamic power dissipation, and how process scaling affects chip design constraints.
Explore CMOS
VLSI RTL Design Reference
A curated RTL design reference specifically for VLSI engineers — covering synthesis constraints, area/power/timing tradeoffs, standard cell usage patterns, and practical RTL guidelines used in production ASIC design teams.
Explore VLSI RTL

ASIC Design Flow — RTL to GDSII

Every chip goes through a structured flow from specification to the final layout file sent to the foundry.

01

Specification

Architecture and micro-architecture decisions: interfaces, clock frequencies, power budget, and feature set.

02

RTL Coding

Hardware described in Verilog/SystemVerilog at register transfer level. Functional simulation verifies behavior.

03

Synthesis

RTL is mapped to technology-specific standard cells. Timing constraints guide optimization for PPA targets.

04

Physical Design

Floorplan, placement, clock tree synthesis, and routing. The netlist becomes a physical layout on silicon.

05

Sign-Off

STA, DRC, LVS, power analysis, and IR drop checks confirm the chip is ready for fabrication.

06

Tape-Out

GDSII file is sent to the foundry. Masks are fabricated and wafers are processed into finished ICs.

Design Flow

The VLSI Design Flow — RTL to GDSII Explained

A chip does not go from idea to silicon in one step. Understanding each stage of the ASIC design flow — and what can go wrong at each stage — is fundamental knowledge for every VLSI engineer.

RTL Design & Functional Verification

The design begins as RTL code — Verilog or SystemVerilog — that describes how data moves between registers and through combinational logic. RTL is the entry point for every chip: it defines the architecture, the protocol interfaces, the data path, and the control logic. RTL code must be not just functionally correct but synthesis-ready — written using coding styles that map cleanly to standard cell libraries without unintentional latches, incomplete sensitivity lists, or reset strategy inconsistencies. Functional verification (simulation with testbenches and UVM environments) validates that the RTL does what the specification says before it moves forward.

Synthesis — RTL to Gate-Level Netlist

Synthesis converts RTL into a gate-level netlist — a structural description of the design in terms of actual standard cells (AND gates, flip-flops, multiplexers, buffers) from a foundry-specific technology library. Tools like Synopsys Design Compiler or Cadence Genus read the RTL along with SDC timing constraints, optimize for area/power/timing, and output a netlist and timing reports. A poorly constrained synthesis — missing clocks, wrong input/output delays, absent clock uncertainty — produces a netlist that looks correct but will fail STA signoff and potentially fail in silicon.

Static Timing Analysis (STA) & Timing Closure

STA verifies that every timing path in the design meets setup and hold requirements across all PVT corners — without applying test vectors. It is exhaustive: every path is checked simultaneously. A chip cannot be taped out with negative slack anywhere. Timing closure is the iterative process of fixing violations — setup violations by shortening data paths (pipelining, gate sizing, logic restructuring) and hold violations by inserting delay buffers. Post-CTS (after clock tree synthesis) is the definitive timing check because the actual clock tree delays are known, clock skew is measured, and uncertainty values are updated from the pre-CTS estimates.

Physical Design — Floorplan, Place, Route, and Signoff

Physical design converts the gate-level netlist into a GDSII layout — the file sent to the foundry for fabrication. The flow includes floorplanning (arranging blocks and macros on the die), power grid planning (VDD/VSS rail routing for IR drop and electromigration), placement (positioning every standard cell), clock tree synthesis (building a balanced clock network to minimize skew), global and detailed routing (connecting all nets through metal layers), and physical verification (DRC for manufacturing rule compliance and LVS to confirm the layout matches the schematic). Every step can introduce timing, power, or reliability problems that feed back into earlier stages.

Why Most Chip Failures Happen at the Boundaries

The most dangerous bugs in VLSI design are those that pass all verification at one stage but fail in the next. An RTL design that simulates correctly may fail synthesis if the reset strategy is wrong. A synthesized netlist that passes pre-layout STA may fail post-layout STA because the physical wire lengths add routing delays the tool did not anticipate. A chip that passes STA signoff at nominal conditions may fail in the field at the slow-slow PVT corner. This is why the VLSI flow is not a straight line — it is an iterative feedback loop where each stage's results drive changes in earlier stages.

Clock domain crossing (CDC) is particularly dangerous because CDC bugs are invisible to both simulation and STA. A signal that crosses asynchronous clock domains without a 2-FF synchronizer will pass all functional tests, pass all timing checks, and then fail randomly in silicon whenever the two clocks happen to be in the wrong phase relationship. This is why CDC verification with dedicated static analysis tools (Synopsys SpyGlass CDC, Cadence JasperGold) is a mandatory step in every serious ASIC flow — not optional post-processing.

The Complete VLSI Engineering Reference

VLSI design is one of the most technically demanding engineering disciplines — combining semiconductor physics, computer architecture, digital design, signal integrity, and manufacturing constraints into a single chip. This section of EcrioniX brings together the key concepts, design techniques, and reference material that every VLSI engineer needs.

Topics are organized from front-end design (RTL, CDC, low power) through fundamental building blocks (flip-flops, latches, clock gaters, FIFOs) to physical design and CMOS fundamentals. Each topic is written with the depth and technical accuracy expected in the semiconductor industry, backed by the same methodologies used in companies like Intel, Qualcomm, ARM, and TSMC.

Whether you are a student learning VLSI for the first time, an RTL engineer wanting to understand physical design, or a physical design engineer brushing up on front-end techniques, EcrioniX gives you the structured, reliable reference you need.

Topics Covered in This Section

  • CMOS Transistor Physics & Inverter Analysis
  • RTL Design in Verilog & SystemVerilog
  • FSM Design — Mealy, Moore, One-Hot
  • Pipelining & Micro-Architecture
  • Clock Domain Crossing (CDC) Techniques
  • Asynchronous FIFO Design & Gray Code Pointers
  • FIFO Depth Calculation Methodology
  • D Flip-Flop & Latch Design
  • Integrated Clock Gating (ICG) Cells
  • Glitch-Free Clock Mux Design
  • Reset Synchronizer — Async Assert, Sync De-assert
  • Low Power RTL — Clock Gating, Power Domains
  • Design for Testability (DFT) & Scan Insertion
  • Physical Design — Floorplan to GDSII
  • Static Timing Analysis & Timing Closure
  • DRC, LVS & Tape-Out Sign-Off