Systemverilog Advanced Verification Using Uvm: Engineer Explorer Series

SystemVerilog® Advanced Verification
Using UVM
Engineer Explorer Series
Version 1.0
Lab Manual May 23, 2011
 1990-2011 Cadence Design Systems, Inc. All rights reserved.
Printed in the United States of America.
Cadence Design Systems, Inc., 2655 Seely Avenue, San Jose, CA 95134, USA
Cadence Trademarks
Trademarks and service marks of Cadence Design Systems, Inc. (Cadence) contained in this document are attributed to Cadence with the appropriate symbol. For
queries regarding Cadence trademarks, contact the corporate legal department at the address above or call 800.862.4522.
Allegro® Incisive® Silicon Express™

Accelerating Mixed Signal Design® InstallScape™ SKILL®
® ™
Assura IP Gallery SoC Encounter™
BuildGates® NanoRoute® SourceLink® online customer support
® ®
Cadence NC-Verilog Specman®
CeltIC® NeoCell® Spectre®
® ®
Conformal NeoCircuit Speed Bridge®
Connections® OpenBook® online documentation library UltraSim®
® ®
Diva OrCAD Verifault-XL®
Dracula® Palladium® Verification Advisor®
® ®
ElectronStorm Pearl Verilog®
Encounter® PowerSuite® Virtuoso®
® ®
EU CAD PSpice VoltageStorm®
Fire & Ice® SignalStorm® Xtreme®
® ™
First Encounter Silicon Design Chain
HDL-ICE® Silicon Ensemble®
Other Trademarks
Open SystemC, Open SystemC Initiative, OSCI, SystemC, and SystemC Initiative are trademarks or registered trademarks of Open SystemC Initiative, Inc. in the
United States and other countries and are used with permission.
All other trademarks are the property of their respective holders.
Confidentiality Notice
No part of this publication may be reproduced in whole or in part by any means (including photocopying or storage in an information storage/retrieval system) or
transmitted in any form or by any means without prior written permission from Cadence Design Systems, Inc. (Cadence).
Information in this document is subject to change without notice and does not represent a commitment on the part of Cadence. The information contained herein is
the proprietary and confidential information of Cadence or its licensors, and is supplied subject to, and may be used only by Cadence customers in accordance
with a written agreement between Cadence and its customers. Except as may be explicitly set forth in such agreement, Cadence does not make, and expressly
disclaims, any representations or warranties as to the completeness, accuracy or usefulness of the information contained in this document. Cadence does not
warrant that use of such information will not infringe any third party rights, nor does Cadence assume any liability for damages or costs of any kind that may
result from use of such information.
RESTRICTED RIGHTS LEGEND Use, duplication, or disclosure by the Government is subject to restrictions as set forth in subparagraph (c)(1)(ii) of the Rights
in Technical Data and Computer Software clause at DFARS 252.227-7013.
UNPUBLISHED This document contains unpublished confidential information and is not to be disclosed or used except as authorized by written contract with
Cadence. Rights reserved under the copyright laws of the United States.
ii Cadence Design Systems, Inc. 5/22/2011

Table of Contents
SystemVerilog Advanced Verification Using UVM
Introduction to the Labs .............................................................................................. 1
Project Overview .......................................................................................................... 2
Before You Begin ......................................................................................................... 8
Lab 1 Creating a Stimulus Model ..................................................................... 10
Lab 2 Creating a Simple UVC ........................................................................... 12
Lab 3 Adding Test Classes and Exploring Phases ........................................ 17
Lab 4 Using Factories ....................................................................................... 20
Lab 5 Generating UVM Sequences .................................................................. 22
Lab 6 Connecting to the DUT Using Virtual Interfaces .................................. 26
Lab 7 Integrating Multiple UVCs ...................................................................... 31
Lab 8 Writing Multichannel Sequences and System-level Tests .................. 34
Lab 9A Creating a Scoreboard Using TLM ........................................................ 36
Lab 9B Router Module UVC (Optional) .............................................................. 38
Lab 9C Further Work with TLM Connections (Optional) .................................. 40
Lab 10 Writing Tests ........................................................................................... 41
Lab B Creating a Simple Functional Coverage Model .................................... 43
Lab C1 Locating Cadence Online Support Solutions ....................................... 45
Lab C2 Customizing Notification and Search Preferences .............................. 47
5/22/2011 Cadence Design Systems, Inc. iii

iv Cadence Design Systems, Inc. 5/22/2011
Introduction
Introduction to the Labs

This section gives a high-level overview of the labs for the SystemVerilog Advanced Verification
Using UVM course.
Software Releases
These exercises and code examples have been tested on the following releases:
 Incisive® Unified Simulator release (9.2 or later)

 Universal Verification Methodology (UVM) release (UVM 1.0)
Lab Database Files
 Classes must be defined in a module scope or package scope.

 In your top-level module, you should import the UVM package and include the UVM
macros, as show below:
module top;
// import the UVM library
import uvm_pkg::*;
// include the UVM macros
ìnclude “uvm_macros.svh”
. . .
endmodule : top
 Your reusable files are located in the <uvc>/sv/ directory.
 Non-reusable files are located in the <uvc>/examples or <uvc>/tb directory.

Simulations are also run in this directory.
5/22/2011 SystemVerilog Advanced Verification Using UVM 1

Introduction
Project Overview
Throughout this training you will be developing a verification environment for a YAPP router
design. These exercises will guide you through building the verification components required to
verify the router design.
The project builds the environment from scratch so you will experience the full process
You will be building one UVC component. The others will be provided for you later in the project.
Packet Router Description
The packet router accepts data packets on a single input port, in_data, and routes the packets to
one of three output channels: channel0, channel1 or channel2. The input and output ports
have slightly different signal protocols. The router also has a host interface for programming
registers that are described in the next section.
High-Level Diagram – YAPP Router (Yet Another Packet Protocol)
data_0 [7:0]
in_data [7:0] data_vld_0

YAPP
in_data_vld Router suspend_0
in_suspend
registers data_1 [7:0]
reset data_vld_1
clock suspend_1
MAXPKT_SIZE
data_2 [7:0]
ROUTER EN
hdata [7:0] data_vld_2
haddr [7:0] suspend_2
Host HBUS I/F
hen
hwr_rd error
2 SystemVerilog Advanced Verification Using UVM 5/22/11

Introduction
Packet Data Specification
A packet is a sequence of bytes with the first byte containing a header, the next variable set of bytes
containing payload, and the last byte containing parity.
The header consists of a 2-bit address field and a 6-bit length field. The address field is used to
determine which output channel the packet should be routed to, with the address 3 being illegal. The
length field specifies the number of data bytes (payload).
A packet can have a minimum payload size of 1 byte and a maximum size of 63 bytes.
The parity should be a byte of even, bitwise parity, calculated over the header and payload bytes of
the packet.
YAPP Packet Structure
7 6 5 4 3 2 1 0
length addr byte 0 Header
payload[0] byte 1
payload[1]
Payload
payload[N-1] byte N
parity byte N+1 Parity

Introduction
Input Port Protocol
All input signals are active high and are to be driven on the falling edge of the clock. The
in_data_vld signal must be asserted on the same clock when the first byte of a packet (the
header byte), is driven onto the in_data bus. As the header byte contains the address, this tells the
router to which output channel the packet needs to be routed. Each subsequent byte of data needs to
be driven on the data bus with each new falling clock.
After the last payload byte has been driven, on the next falling clock, the in_data_vld signal
must be de-asserted, and the packet parity byte needs to be driven. The input data cannot change
while in_suspend signal is active (indicating FIFO full). The error signal asserts when a packet
with bad parity is detected, within 1 to 10 cycles.
Input Protocol
clock
delay
in_data_vld
in_data H D D D P H D D D P
in_suspend
sent packet Packet 1 Packet 2
H = Header, D = Data, P = Parity

Introduction
Output Port Protocol (Channel Ports)
All output signals are active high and are to be sampled on the falling edge of the clock. Each
output port is internally buffered by a FIFO of depth 16 and a width of 1 byte. The router asserts the
data_vld_x signal when valid data appears on the data_x output bus. The suspend_x input
signal must then be de-asserted on the falling clock edge in which data is read from the data_x
bus. As long the suspend_x signal remains inactive, the data_x bus drives a new valid packet
byte on each rising clock edge.
Output Protocol
clock
input in_data_vld
port
in_data H D D D P H D D D P
data_vld_0
channel0
output Response
port suspend_0
delay
data_0 H D D D P
received packet Packet 1 (addr = 0)
H = Header, D = Data, P = Parity

Introduction
Packet Router DUT (Design Under Test) Registers
The packet router contains two internal registers that hold configuration information. These registers
are accessed through the host interface port as follows:
 MAXPKTSIZE – 8 bits – address 0 R/W - Reset value of ‘h3F

 ROUTER_EN - 1 bit - address 1 R/W - Reset value of 1
If the input packet length is greater than the value of the MAXPKTSIZE register, the router drops
the entire packet and the error flag is raised.
The ROUTER_EN register provides control of disabling the routing feature. Enabling or disabling
the router during packet transmission will yield to unpredictable behavior.
7 6 5 4 3 2 1 0
Registers
MAXPKTSIZE 0
length address
ROUTER_EN 1

Introduction
Host Interface Port Protocol (HBUS)
All input signals are active high and are to be driven on the falling edge of the clock. The host port
provides synchronous read/write access to program the router.
A WRITE operation takes one clock cycle as follows:
 hwr_rd and hen must be 1. Data on hdata is then clocked on next rising clock edge in
to the register based on haddr decode.
 hen is driven to 0 in the next cycle.

A READ operation takes two clock cycles as follows:
 hwr_rd must be 0 and hen must be 1. In the first clock cycle, haddr is sampled and
hdata is driven by the design under test (DUT) in the second clock cycle.
 hen is then driver low after cycle 2 ends. This will cause the DUT to tri-state the hdata
bus.
HBUS Protocol
clock
hen
hwr_rd
haddr A A
hdata D D
transaction write read

Introduction
Before You Begin
Finding Your Files
The UVM training lab database includes the following directories and files:
uvm_training_xx/uvm Top-level directory for the training
labxx/ Lab directories where you will be working
hbus HBUS UVC and associated files
channel Output channel UVC and associated files
yapp YAPP input port UVC and associated files
router Router module UVC and associated files
router_rtl RTL Files for the YAPP Router DUT
test_install Simple example to check your UVM installation
uvm1.0xx UVM 1.0 library files
Accessing UVM
UVM 1.0 is supported by IUS versions IES 10.2s11 or later and IES 9.2s35 or later.
UVM 1.0 may not be provided with the simulator, depending on the release version. Your instructor
will tell you whether UVM is provided with your simulator release.
For simulator releases not including UVM, there are two options for accessing the UVM library:
 Use the library provided with the lab files

 Download the library from www.uvmworld.org.
The first option is the best solution for IUS users, as the installation contains extra Cadence additions
for UVM such as transaction recording.
For simulator releases which include UVM, the library can be accessed via the irun command line
option –uvm.
Setting up UVM (Simulator Releases without UVM)
1. Set an environmental variable to the UVM install path which contains the src, docs
and examples directories: e.g. for the UVM library provided with the lab files
setenv UVMHOME <lab path>/uvm_fund_1.0/uvm/uvm10-cdns
.
Introduction
Testing Your UVM Installation
2. Change directory to test_install.
3. There are two options for compiling the test, depending if your simulator includes or does
not include a compiled copy of UVM 1.0. Make sure you use the right script for your
version of IUS. Ask your instructor if you are unsure.
a. For IUS versions with UVM, execute the following command:

irun –f emblib.f
b. For IUS versions without UVM, execute the following command:

irun –f lablib.f
4. Check the version number of the UVM library. You should see:
-------------------------------------------------------
UVM-1.0p2...
5. Check the installation test passed. You should see:

UVM_INFO ... UVM TEST INSTALL PASSED!
Accessing Help Files
If you are using the UVM library provided with the lab files, you may wish to bookmark the
following file in a web browser to give convenient access to the UVM HTML documentation
$UVMHOME/docs/html/index.html
Additional Notes
File names, component names, and instance names are suggested for many of the labs. You are not
required to use these names but if you do not, you may need to edit code provided later in the week
to match your path and instance names.
These labs do not include step-by-step instructions, and do not tell you exactly what you need to
type.

Creating a Stimulus Model Lab 1
Lab 1 Creating a Stimulus Model
Objective: To use the UVM class library to create the YAPP packet data item and
explore the automation provided.
For this lab, start in the lab01/sv directory.
Creating a Data Item
1. Review the YAPP packet data specification on pages 1-4 and create your packet data
class (yapp_packet) in a file yapp_packet.sv. Name your class properties to
match the specification.
a. Use uvm_sequence_item as the base class.
b. Use the ùvm_object_utils and the ùvm_field_* macros for built-in

automation of the class properties.
c. Add an UVM data constructor new() in the class definition.
2. Add support for randomization of the packet:
a. Declare the length, address and payload properties as rand. Do not use a
rand property for parity.
b. Create a method calc_parity() method to calculate packet parity:

function bit [7:0] calc_parity();
c. Declare an enumeration type, parity_t, outside the yapp_packet class, with the
values GOOD_PARITY and BAD_PARITY. Create a property parity_type as an
abstract control “knob” for controlling parity and declare this property as rand.
d. Add a post_randomize()method to assign the parity property after the other

packet properties have been randomized. If parity_type has the value
GOOD_PARITY, assign parity using the calc_parity() method. Otherwise
write an incorrect parity value.
e. Add constraints for valid address and payload length (see packet description).
f. Add a constraint for parity_type with a distribution of 5:1 in favor of good parity
g. Constrain the size of the payload to be equal to the packet length.
h. Create another rand control “knob” (packet_delay) of type int, to insert a delay
when transmitting a packet. Constrain the delay to be inside the range 0 to 20.

Lab 1 Creating a Stimulus Model
Creating a Simple Test
3. Move to the lab01/tb directory.
4. Modify the top-level test module (top.sv):
a. Import the UVM library and include the UVM macros.

// import the UVM library
import uvm_pkg::*;
// include the UVM macros

ìnclude “uvm_macros.svh”
b. Include the yapp_packet.sv definition and create an instance of the class.
c. Using a loop, generate five random packets and use the UVM print() method to
display the results. Try printing using both the table printer and the tree printer
options.
d. Edit the run file run.f, in the tb directory, to use either the embedded or supplied
UVM library depending on your simulator version.
e. Add the following lines to your run.f file:

-incdir ../sv
top.sv
f. Compile and simulate your code using the following command (IUS9.2).
% irun –f run.f
Ask your trainer for the compilation command if using other tools or versions.
5. (Optional) Edit the top.sv file to explore the UVM built-in automation: copy(),
clone() and compare().

Creating a Simple UVC Lab 2
Lab 2 Creating a Simple UVC
Objective: To create a simple UVM Verification Component (UVC, aka env) and
print the topology.
You will be creating a simple driver, sequencer, monitor, agent and env for the YAPP input port of
the router. You will focus on the transmit (TX) agent for this lab.
Note: Remember to add the UVM component constructor new() to each class definition.
Work in the lab02/sv directory. Use the training slides for suggestions on implementing these
components.
Creating the UVC
1. First - copy your yapp_packet.sv definition from lab01/sv into the lab02/sv
directory.
Also copy your run.f file from lab01/tb into the lab02/tb directory.
2. Create the yapp_tx_driver in the file yapp_tx_driver.sv.
a. Use uvm_driver as the base class for the driver and use the
ùvm_component_utils macro. Using type parameters will make your UVC
easier to develop.
b. Add a run_phase() task. Use a forever loop to get and send packets, using the
seq_item_port prefix to access the communication methods
(get_next_item(), item_done() )..
c. Add a send_to_dut() task. For the moment, this task should just print the packet.
Use an ùvm_info macro with a verbosity of UVM_LOW. Use the following code in
the string portion of the macro (where pkt is the handle name for your YAPP
packet):
$sformatf("Packet is \n", pkt.sprint())
Note: sprint() is a sub-method of print(), which creates a print string, but does not
write it to the output. We use sprint() in a report macro, rather than print()
directly, to allow us to control the printing of the packet with verbosity settings.
3. Create the yapp_tx_sequencer in the file, yapp_tx_sequencer.sv.
a. Use uvm_sequencer as the base class for the sequencer. Remember to use a type
parameter if you used one for the driver.
b. Use the ùvm_component_utils macro.

4. Create the yapp_tx_monitor in the file yapp_tx_monitor.sv:
a. Use uvm_monitor as the base class for the monitor and add the
ùvm_component_utils macro.
b. Add a simple run_phase() task which prints an uvm_info message that you are
in the monitor.
5. Create the yapp_tx_agent in the file yapp_tx_agent.sv.
a. Use uvm_agent as the base class for the agent and use the
ùvm_component_utils macro.
b. The agent will contain instances of the yapp_tx_monitor, yapp_tx_driver

and yapp_tx_sequencer components. Declare handles for these and name them
monitor, driver, and sequencer, respectively
c. The agent will contain an is_active flag to control whether the agent is active or
passive. Initialize this to active:
uvm_active_passive_enum is_active = UVM_ACTIVE;
d. Add a build_phase() method to the agent to construct the component hierarchy.

Call super.build_phase(phase), then construct the monitor which is
always present. You will then conditionally construct the driver and sequencer,
only if the is_active flag is set to UVM_ACTIVE.
e. Add a connect_phase() method to the agent. Conditionally connect the

seq_item_export of the sequencer and the seq_item_port of the driver,
based on the is_active flag.
6. Create and implement the Environment (yapp_env) in the file yapp_env.sv.
a. This component will contain an instance of the yapp_tx_agent. Declare a handle

named agent for this instance.
b. Construct the agent in a build_phase() method for the env. Remember to call
super.build_phase(phase) first.
c. Implement a run_phase() task which prints the env (just call print()) ..

7. Create an UVC include file, yapp.svh.

This file needs to contain all of the reusable component definitions for this environment.
Include all of the files you created for this lab, together with the yapp_packet.sv
from Lab01, and the supplied file yapp_tx_seqs.sv, as follows:
ìnclude "yapp_packet.sv"
ìnclude "yapp_tx_monitor.sv"
ìnclude "yapp_tx_sequencer.sv"
ìnclude "yapp_tx_seqs.sv"
ìnclude "yapp_tx_driver.sv"
ìnclude "yapp_tx_agent.sv"
ìnclude "yapp_env.sv"
This file will be included in the top-level test (top.sv).
Don't forget to include the supplied file yapp_tx_seqs.sv. This file contains a
sequence to help you verify your UVC. We will discuss sequences in detail later in the
class.
Checking the UVC Hierarchy
8. Work in lab02/tb directory. Create a simple test to check the UVC hierarchy as
follows:.
a. Edit the top-level module file: top.sv.
b. Include the yapp.svh file you created in step 7.
c. For this simple test, create an instance of yapp_env.
d. In an initial block, construct the env using the correct syntax.
e. In a separate initial block, call a task run_test() without any arguments. This
built-in method will run a default test (details described later).
9. Run a simulation using the following command:

% irun –f run.f
a. Look for the printed hierarchy. Does the hierarchy match your expectations?

b. Use the printed hierarchy to work out the full hierarchical pathname to your
sequencer (e.g. env.agent.sequencer) and write it below.
Sequencer pathname: _________________________________________________
c. Use your printed hierarchy to find the value of the yapp_tx_agent is_active
property.
What is the value of the is_active variable when you printed the hierarchy?
Running a Simple Test
10. Your top-level module, top.sv, contains the following commented code:
uvm_config_wrapper::set(null, "<path>.run_phase",
"default_sequence",
yapp_5_packets::type_id::get());
This code sets the default sequence of the run_phase to the sequence
yapp_5_packets defined in the yapp_tx_seq_lib.sv file.
a. Un-comment this code and move it into your initial block containing
run_test(). Place the code in front of the run_test() call.
b. Edit the code to replace <path> with the hierarchical pathname to your sequencer
as recorded above.
c. Check your initial block looks like this (although your pathname may be different):
initial begin
uvm_config_wrapper::set(null, "env.agent.sequencer.run_phase",
"default_sequence",
yapp_5_packets::type_id::get());
run_test()
end
Note: In UVM 1.0, a UVC will not create any stimulus without an explicit sequence. In this
Lab you will use a pre-supplied sequence to allow your UVC to generate YAPP
packets. In UVM 1.0 we must use a configuration database set call to define the
sequence to be executed.
Both sequences and configurations are covered in later modules of this course.

11. Run a simulation using the following command:

% irun –f run.f
Your UVC should now generate and print YAPP packets.

How many packets were generated?
Is every field of the packet printed?
Do you see good and bad parity packets?
12. Explore the verbosity options. Verbosity can be changed by adding the following option
to the irun command line or compile file:
+UVM_VERBOSITY=UVM_LOW
If you have use the correct syntax and verbosity for ùvm_info, you will see different
amounts of data printed when using different verbosity options. Try UVM_NONE and
UVM_FULL. You should also note that the testbench is not re-compiled when you change
verbosity.
13. Edit your run.f file to add the following compilation option:
+SVSEED=random
This sets a random value for the initial randomization seed of the simulation. Run the
simulation again and you should see different packet data created for each run. The
simulation reports the actual seed used for each simulation in the irun.log file.

Lab 3 Adding Test Classes and Exploring Phases
Objective: To use a test class to verify the simple YAPP UVC and to explore the
built-in phases of uvm_component.
Create and work in lab03/sv and lab03/tb directory. For this lab, you will construct the YAPP
UVC in a test class instead of in the top module.
1. First – copy your files from lab02/ into lab03/ , e.g. from the uvm directory, type:
% cp –R lab02/* lab03/
2. In the lab03/sv/ directory, edit the yapp_env.sv file and delete the
run_phase() task containing the print() method call.
3. In the lab03/sv directory, add a start_of_simulation_phase()method to

your sequencer, driver, monitor, agent and environment components.
The method should simply display a message indicating in which component the method
is being called. (use ùvm_info with a verbosity of UVM_HIGH).
4. In the lab03/tb directory, create a test file, yapp_test_lib.sv as follows:
a. Name the test class base_test and inherit from uvm_test.
b. Add the ùvm_component_utils macro.
c. Add a component constructor new().
d. Declare a handle for yapp_env inside the test.
e. Add a build_phase() method containing super.build_phase(phase).
f. Construct the yapp_env instance in build_phase(). Add the constructor call

after super.build_phase(phase).
g. From the lab03/tb/top.sv file, copy the uvm_config_wrapper::set

code and paste it into the build_phase() method of base_test before
super.build_phase().
h. Edit the uvm_config_wrapper::set code in base_test to replace null

with this.
When set is called from the top module, it must use null. When you move set
into the test class, you must change null to this, which allows the simulator to
resolve the sequencer hierarchical pathname from the test class scope.

Adding Test Classes and Exploring Phase Lab 3
i. Add an end_of_elaboration_phase() method to the test and use the

uvm_top.print_topology() command to print the testbench hierarchy.
5. In the lab03/tb directory, modify the top-level module, top.sv as follows:
a. Delete the yapp_env handle and constructor, as the environment is now created in
the test class component.
b. Delete the uvm_config_wrapper::set code, as the set call is now made in

the test class.
Executing the Test
6. Run the test suite.
a. Edit your run.f file to add the following lines:

+UVM_TESTNAME=base_test
+UVM_VERBOSITY=UVM_MEDIUM
b. Run a simulation using the following command:

irun –f run.f
7. Check the output from simulation as follows:
a. Check the testbench hierarchy is correct. Does it match your expectations?
b. Check which start_of_simulation_phase() method was called first. Which

is called last? Why? You will need to set the proper +UVM_VERBOSITY option to
see the phase method messages.
8. (Optional) In the test library file yapp_test_lib.sv, create a second test named
test2.
Extend your test2 class from base_test. What is the minimum amount of code for
test2, given that we are inheriting from base_test?
9. (Optional) Compile with the option +UVM_TESTNAME=test2.

You should edit the run file to change the UVM_TESTNAME option. Simply adding
another UVM_TESTNAME option at the end of the command line will not over-ride the
run file option.
Note that you can switch between base_test and test2 via the command-line option
+UVM_TESTNAME, without re-compiling your test environment.

Use your printed hierarchy to make sure the correct test is being executed.

Using Factories Lab 4
Lab 4 Using Factories
Objective: To create verification components and data using factory methods, and
to implement test classes using configurations.
Create and work in lab04/sv and lab04/tb directories. For this lab, you will modify your
existing files to use factory methods, and explore the benefits of configurations.
Note: Remember to include the UVM component constructor new() in each new class
component definition.
1. First – copy your YAPP files from lab03/ into lab04/, e.g. from the uvm directory,
type:
% cp –R lab03/* lab04/sv
Note: You will be making many changes to files in this lab – so PLEASE copy the sv and
tb directories from lab03 to lab04 to keep a working copy available!
Using the Factory
The first step is to use the factory methods to allow configuration and test control from above
without changing the sub-components.
2. Replace the new() constructor calls in the build_phase() methods by calls to the
factory method create(). You will need to modify the following files:
tb/test_lib.sv
sv/yapp_env.sv
sv/yapp_tx_agent.sv
3. Run your original test (base_test) to make sure the changes are working.

Lab 4 Using Factories
Using Configurations and Overrides
4. Create a new short packet test as follows:
a. Define a new packet type, short_yapp_packet, which extends from

yapp_packet. Add this subclass definition to the end of your
sv/yapp_packet.sv file. Set a constraint to limit the packet length to less than
15. Set another constraint to exclude an address value of 2.
b. Define a new test, short_packet_test in the file tb/yapp_test_lib.sv.

Extend this from base_test.
c. In the build_phase() method of short_packet_test, use a

set_type_override method to change the packet type to
short_yapp_packet.
d. Run a simulation using the new test, (+UVM_TESTNAME=short_packet_test)

and check the correct packet type is created.
5. Create a new configuration test in the file yapp_test_lib.sv.
a. Define a new test, set_config_test which extends from base_test.
b. In the build_phase() method, use the set_config_int() method to set the

is_active bit to UVM_PASSIVE in the YAPP TX agent before building the
yapp_env instance.
c. Make sure you are calling uvm_top.print_topology() from an

end_of_elaboration_phase() method, either explicitly in the
set_config_test class, or via inheritance from base_test.
d. Run a simulation using the set_config_test test class

(UVM_TESTNAME=set_config_test) and check the topology print to ensure
your design is correctly configured.

Generating UVM Sequences Lab 5
Lab 5 Generating UVM Sequences
Objective: To use the uvm_sequence mechanism to define a sequence library and

to control execution of sequences.
Create and work in lab05/sv and lab05/tb directories. For this lab, you will explore sequence
writing and the objection mechanism for coordinating simulation time.
Creating Sequences
1. First – copy your yapp files from lab04/ into lab05/, e.g. from the uvm directory,
type:
% cp –R lab04/* lab05/
tb directories from lab04 to lab05 to keep a working copy around!
2. In the lab05/sv director, edit the sequences file, yapp_tx_seqs.sv, to add the
following sequences.
For every sequence:
a. Inherit the sequence from yapp_base_seq to use the objection mechanism.
b. At the start of every sequence body(), add an ùvm_info call to print the
sequence name. Use a verbosity of UVM_LOW.
c. Remember to add a constructor and object utilities macro.
3. Create two new packet sequences with different constraints in the body.
yapp_012_seq – three packets with incrementing addresses
(do_with addr ==0; do_with addr==1; do_with addr==2)
yapp_1_seq – single packet to address 1
(do_with addr==1)
4. Create a simple nested sequence.

yapp_111_seq – three packets to address 1
(do yapp_1_seq three times.)
5. Create a repeating address sequence.

yapp_repeat_addr_seq – two packets to the same (random) address
(do_with addr==prev_addr: remember packet address cannot be 3)

6. Create an incrementing payload sequence

yapp_incr_payload_seq –
Create a packet to send
Randomize the packet.
Set the payload values to increment from 0 to length -1.
Update parity
Send it
(Hint: Use ùvm_create and ùvm_send macros).
7. (Optional) Create a random number of packets

yapp_rnd_seq –
Declare a rand int property count in the sequence and generate a number of
random packets according to the count value.
Set a constraint inside the sequence to limit count to the range 1 to 10.
Include the value of count in the sequence ùvm_info message.
8. (Optional) Create a nested sequence with a constraint

six_yapp_seq – do yapp_rnd_seq with count constrained to six.
Running a Test Using a New Sequence
9. Create a new test in the file yapp_test_lib.sv from the lab05/tb directory:
a. Call the test short_incr_payload and extend from base_test.
b. Move the uvm_config_wrapper::set code from base_test to the

build_phase() method of short_incr_payload.
c. Edit the uvm_config_wrapper::set to set the run_phase default sequence

to yapp_incr_payload_seq.
d. Add a set_type_override() method to use the short_packet data type

defined in Lab04.
e. Run the test and verify the results. Setting verbosity to UVM_FULL will allow you to
see which sequence is executed in the run_phase().
Using a Sequence Library

You could check your new sequences by creating a separate test for each sequence, but
this is impractical. Instead we can collect all the sequences into a sequence library and
execute every sequence from the library.

Generating UVM Sequences Lab 5
10. Create a new file in the lab05/tb directory called yapp_seq_lib.sv. Edit the file
as follows:
a. Create the library class with the name yapp_seq_lib and inherit from
uvm_sequence_library. Type parameterize the class for yapp_packet,
a. Add an ùvm_object_utils macro and a component constructor.
b. In the constructor, use separate add_sequence calls to add your sequences to the
library. For example:
add_sequence(yapp_repeat_addr_seq::get_type());
c. Add an include for yapp_seq_lib.sv in the top.sv module.
11. Create a new test in the file yapp_test_lib.sv:
a. Call the test exhaustive_seq_test and extend from base_test.
b. Add a handle for the yapp_seq_lib library and create the instance in the test
constructor.
c. In the build_phase() method, change the selection_mode property of the

library instance to UVM_SEQ_LIB_RANDC.
This sets the sequence library to use cyclic randomization when randomly choosing
sequences, i.e. each sequence will be executed once before any repetition.
d. Also in the build_phase(), change both the min_random_count and

max_random_count properties of the library instance to the number of sequences
you added in the sequence library constructor.
This sets the sequence library to randomly execute the specified number of
sequences. Combined with the cyclic randomization selection mode, this will execute
each sequence once only, but in a random order.
e. Use a uvm_config_db#( uvm_sequence_base)::set in the

build_phase() to set the default sequence of the run_phase of the YAPP
sequencer to your library instance.
f. Finally add a set_type_override() method to use the short_packet

data type defined in Lab04.
g. Add an end_of_elaboration_phase() method which prints the sequence

library (<instance>.print()). This will help in debugging. Make sure you still
have a print_topology(), either explicitly or via a super call to base_test.

Testing the Sequence Library and Fixing Randomization Errors
12. Run the test (+UVM_TESTNAME=exhaustive_seq_test) and check the results.

Note: You should get randomization failures for particular sequences due to constraint
violations. Examine the simulation output carefully to see how the constraint
violations are reported. Note that in batch mode, simulation is NOT stopped.
Why do you get violations?
Answer:
What happens to the packet when a constraint violation is found?

Answer:
How could you fix these violations?

Answer:
13. Run a simulation using the GUI. The simulation will now be stopped on a constraint
violation and the constraints manager tool opened. This tool allows complex constraint
violations to be more easily debugged.
14. Fix your code to make the simulation run without constraint violations.
It is important that you fix the violation before moving on to the next lab.

Connecting to the DUT Using Virtual Interfaces Lab 6
Lab 6 Connecting to the DUT Using Virtual Interfaces
Objective: To connect the YAPP UVC to the input port of the DUT.
Work in the lab06/sv and lab06/tb directories. For this lab, you will connect your YAPP
UVC to the RTL router Design Under Test (DUT) using interfaces, virtual interfaces, and
(optionally) a DUT wrapper.
The RTL model of the router DUT is located in the router_rtl directory.
With a few lines of extra code, the router DUT will operate correctly without connecting anything to
the HBUS interface signals or the channel outputs. This will allow you to test your YAPP UVC
(yapp_env) connection to the DUT without having to use the HBUS or channel UVCs.
Modifying the Yapp UVC
1. First – copy your yapp files from lab05/ into lab06/.

tb directories from lab05 to lab06 to keep a working copy around!
2. Check the YAPP interface to the DUT, yapp_if.sv, which is supplied in the
lab6/sv directory.
Note that the interface has two input ports (clock, reset) and three DUT signals,
in_data, in_data_vld and in_suspend.
3. Connecting the YAPP interface via the configuration database will be easier if you
declare a typedef for the uvm_config_db with a yapp_if type parameter. This
declaration has to be visible to your monitor, driver, and top-level module.
Add the following declaration to yapp.svh, before the include statements:
typedef uvm_config_db#(virtual yapp_if) yapp_vif_config;
4. Update your Monitor, yapp_tx_monitor.sv.
a. You will to need to use a collect_packets() method to capture the packet data.
The methods and some of the declarations for the monitor are provided in the file
monitor_example.sv. Check you understand the code.
Use the supplied code to update your monitor.
b. Add a declaration for the virtual interface:

virtual interface yapp_if vif;

c. Add a build_phase() method containing a yapp_vif_config::get call to

assign vif from the configuration database. Remember get returns bit 1 if it was
successful. Passing this return value to an assert or if statement will help debug
the virtual interface connection. For example:
if (!yapp_vif_config::get(this,"","vif", vif))
ùvm_fatal("NOVIF",{"vif not set for: ",get_full_name(),".vif"})
5. Update your Driver protocol, yapp_tx_driver.sv.
a. You will need to use a send_to_dut() method to transmit packet data.

You will also need to reset the input DUT signals when the reset is active using a
reset_signals() method, and call this in the driver run_phase().
These methods and some of the declarations for the monitor are provided in the file
driver_example.sv. Check you understand the code.
Use the supplied code to update your driver.
b. Make sure you update the get_and_drive() method to pull down packets from
the sequencer (see comments in code).
c. Add a declaration for the virtual interface :

virtual interface yapp_if vif;
d. Add a build_phase() method containing a yapp_vif_config::get call to

assign vif from the configuration database, as for the monitor.
Adding a Testbench
6. In the lab06/tb directory, create a top-level testbench, router_tb.sv.
a. The router_tb extends from uvm_component. It contains a single instance of

the yapp_env.
b. The testbench requires an ùvm_component_utils macro, a constructor and

build_phase()method. Create the yapp_env instance in build_phase().
c. Add the following configuration line to build_phase()method, before

super.build_phase(), to enable transaction recording.
set_config_int( "*", "recording_detail", 1);
7. Modify your tests to include the testbench layer, by editing yapp_test_lib.sv.
a. Replace the yapp_env instance in base_test with a router_tb instance.

b. Any hierarchical references to yapp_env need to include an additional

router_tb handle reference, for example in configuration set pathnames.
c. Add a run_phase() method to base_test which sets a drain time for the
objection mechanism as follows:
phase.phase_done.set_drain_time(this, 200ns);
The drain time will allow enough time for the packets to pass through the router
design before the simulation ends.
d. Rename the test library file to router_test_lib.sv, as it will contain test

classes for the entire router from this point on.
Initial Simulation Without the DUT
8. A top-level module, top_no_dut.sv, is provided for you in the tb directory. The

module supplies the following functionality:
 Declares the clock, reset and in_suspend signals; initializes them and
generates the waveforms required.
 Instantiates the yapp_if interface:
yapp_if in0 (clock, reset); // interface instance
 Calls run_test() in an initial block to initiate the simulation.
a. Use set to write the YAPP interface instance into the configuration database (Hint:
Use your typedef from step 3). Use wildcards in the pathname to affect both
monitor and driver with a single statement. Remember for a top module set, the
context is null.
b. Add a ìnclude for your YAPP UVC, YAPP sequence library, router testbench,
and renamed router test library files.
9. Test your changes before adding the DUT.
a. Run a test to verify that things are working correctly before connecting to the DUT.
Remember the interface file yapp_if.sv must be compiled on the command line
alongside top_no_dut.sv.
b. Examine the simulation output carefully to check that the monitor is capturing correct
packets according to the sequence being used.
A test class which uses a configuration set to define the sequence will make the
output easier to interpret.

Testing with the DUT
10. Edit your top_no_dut.sv module to add the DUT.
a. Instantiate the yapp_router in the module and connect the ports to the yapp_if
interface signals via named mapping. For example:
.in_data(in0.indata),
b. Remove the following assignment of the in_suspend interface signal:

in0.in_suspend <= 0;
c. Set the suspend_0, suspend_1 and suspend_2 router input signals to 1’b0
via assign statements or port mapping to allow packets to pass through the device.
d. Save your top module as top_dut.sv.
11. Run a test to make sure the router is correctly connected.. Remember to compile the
yapp_router.v file from the router_rtl directory. You may need the following
irun timescale option to avoid timescale errors:
-timescale 1ns/100ps
12. Use SimVision to display signals from the yapp_router and the yapp_if. If you
have transaction recording in the monitor or driver, you can also view the transactions.
Ask your trainer for guidance.
Optional: Implementing a SystemVerilog DUT Wrapper
A wrapper is a module which instantiates the router DUT. The wrapper has interface ports for the
YAPP, HBUS, and channel connections and maps the interface signals to the individual ports of the
router DUT. This makes instantiation and connection of the router DUT in the top module much
easier.
13. Create a SystemVerilog module in the file lab06/tb/yapp_wrapper.sv and add

the following:
a. Instantiate the YAPP router DUT (copy from top_dut.sv).
b. Add ports for clock, reset and error.
c. Add the yapp_if, the HBUS and three channel_if interface ports. The
channel_if.sv file can be found in the channel/sv directory.
The hbus_if.sv file can be found in the hbus/sv directory
d. Connect the interface signals to the router instantiation.

e. Modify your top_dut.sv file to instantiate all five interfaces and the wrapper
instead of the router. Connect up the wrapper and simulate your design to check for
correct connection. Remember to drive the channel interface suspend signals to
1'b0.

Lab 7 Integrating Multiple UVCs
Objective: To connect and configure the HBUS UVC and three output channel
UVCs with the input port UVC and the YAPP_router design.
For this lab, you will connect the HBUS and Channel UVCs to the router DUT.
The HBUS UVC is provided. This will need to be connected to the HBUS port.
The Channel UVC is also provided. This is almost identical to the YAPP input UVC, except the
Channel UVC must properly set/reset the in_suspend signal.
These are the directories we will be using for this and subsequent labs:
hbus/sv HBUS UVC files
channel/sv Channel UVC files
yapp/sv YAPP input UVC (your files from lab06)
router_rtl Router DUT
lab07/tb Your working directory for this lab
Setting Up the Directory Structure
1. First – Your YAPP UVC is now complete enough to stand by itself. Copy your YAPP
files from lab06/sv into yapp/sv.
Note: PLEASE copy the sv directory from lab06 to yapp to keep a working copy
around!
2. We will still be working on the testbench, testclass, and top files. Copy these files from
lab06/tb into lab07/tb.
Note: PLEASE copy the tb directory from lab06 to lab07 to keep a working copy
around!
Work in the lab07/tb directory.
Testbench: Channel UVC
3. Update your testbench, router_tb.sv, to add the Channel UVCs.
a. Add three handles of the Channel UVC (channel_env) and create the instances in
the build_phase() method using factory calls.

Integrating Multiple UVCs Lab 7
b. The Channel UVC has both transmit and receive agents. For the router testing, we
only need the Receive agent. Use a set_config_int method to set the has_tx
property of each Channel instance to 0.
Testbench: HBUS UVC
4. Update your testbench, router_tb.sv, to add the HBUS UVC.
a. Add a handle of the HBUS UVC (hbus_env) and create the instance in the
build_phase() method using a factory call.
b. The HBUS UVC has both master and slave agents. For the router testing, we only
need the Master agent. Use set_config_int methods to set the num_masters
property of the HBUS env to 1, and the num_slaves property to 0.
Test Library
5. Add a new test class, simple_test, in router_test_lib.sv as follows:
a. Extend from base_test class.
b. Use set_type_override and uvm_config_wrapper::set to create the

following (copy from existing tests):
 Set the YAPP UVC to create short YAPP packets.
 Set the run_phase default sequence of the YAPP UVC to the
yapp_012_seq sequence.
 Set the run_phase default sequence of each Channel UVC to
channel_rx_resp_seq.
 Do not define a default sequence for the HBUS UVC.
c. (Optional) Now might be a good time to clean up the test library and remove the older
tests. Delete all the other test classes besides base_test and simple_test.
Top Module
6. Update your top-level module to add interface instantiations for the Channel and HBUS
interfaces if you do not already have these:
a. Add the HBUS and three channel_if interface ports.

The channel_if.sv file can be found in the channel/sv directory. The
hbus_if.sv file can be found in the hbus/sv directory.

b. Connect the interface signals to the router instantiation.
c. Add the include files for the Channel and HBUS UVCs. These can be found in the sv
directory of each UVC.
d. Set the HBUS and Channel UVC virtual interfaces to the correct interface (Hint: Use
typedef where possible – check the UVC header files for existing typedefs.).
Use wildcards in the pathname to update all UVC components with a single
statement. Remember for a top module set, the context is null.
Running a Test
7. Run a simulation. Check the following:
a. Sequencer settings are correct for all three UVCs.
b. Packets are passed correctly through the router and collected at the right channel.
8. (Optional) Write a new YAPP sequence in the yapp/sv/yapp_tx_seqs.sv file to

generate packets for all four channels. The packets should have incrementing payload
sizes from 1 to 22 and parity distribution of 20% bad parity (88 packets in total).
Hint: You could create packets without randomization, using nested loops (for address
and payload) and uvm_create and uvm_send macros.
9. (Optional) Create a new test, test_uvc_integration, in the

lab07/tb/router_test_lib.sv file to perform the following:
a. Set the run_phase default sequence of the HBUS UVC to set-up the router with
MAXPKTSIZE = 20 and ROUTER enabled. (Hint: There is a sequence defined for this
in the HBUS master sequences hbus_master_seqs.sv.)
b. Set the YAPP UVC to run your new YAPP sequence defined above.
10. (Optional) Run a simulation and check the results to see that the three channels are
properly addressed, that there is an error signal when parity is wrong, and that packets are
dropped if bigger than MAXPKTSIZE.
How many packets are dropped per channel?
Are there any packets dropped due to bad address (=3)?

Writing Multichannel Sequences and System-level Tests Lab 8
Lab 8 Writing Multichannel Sequences and System-level Tests
Objective: To build and connect virtual sequences to your testbench.
For this lab, you will build and connect a virtual sequencer for the router, and create virtual
sequences to co-ordinate the activity of the three router UVCs.
1. We will be working on the testbench, testclass and top files from lab07. Copy these files
from lab07/tb into lab08/tb.
Note: You will be making many changes to files in this lab, so PLEASE copy the tb
directory from lab07 to lab08 to keep a working copy around!
Work in the lab08/tb directory.
2. Create the virtual sequencer router_virtual_sequencer.sv.
a. Add a component macro and constructor.
b. Add the references for the HBUS and YAPP UVC sequencer classes. You could also
add a reference to the Channel sequencer class, but since there is only one sequence
for that sequencer, the handle can be omitted.
3. Create a virtual sequence file, router_virtual_seqs.sv and define a simple

virtual sequence, router_simple_vseq, as follows:
a. Add an object macro and constructor.
b. Add a ùvm_declare_p_sequencer macro to access the virtual sequencer

references.
c. Using the sequences defined in the YAPP and HBUS UVC sequence libraries (and
optionally also the Channel), create a virtual sequence to:
 Raise an objection on starting_phase.
 Set the router to accept small packets (payload length < 21) and enable it.
 Read the router MAXPKTSIZE register to make sure it has been correctly set.
 Send six consecutive packets to address 0, 1, 2, cycling the address.
 Set the router to accept large packets (payload length < 64).
 Read the router MAXPKTSIZE register to make sure it has been correctly set.
 Send a random sequence of six packets.
 Drop the objection on starting_phase.

Lab 8 Writing Multichannel Sequences and System-level Tests
4. Modify the router_tb.sv testbench to instantiate, build, and connect the virtual
sequencer.
5. Create a router_vtest_lib.sv test file to instantiate and build the virtual

sequencer testbench.
Tip: Copy or extend your test from the router_test_lib.sv created in lab07.
Use configuration set methods, type overrides and the following checklist to correctly
configure your testbench:
a. Set a type override for short packets only.
b. Set the default sequence of all output channel sequencers to

channel_rx_resp_seq.
c. Set the default sequence of the virtual sequencer to the router_simple_vseq

sequence declared above.
d. Do not set a default sequence for the YAPP or HBUS sequencer. Control is now
solely from the virtual sequencer.
6. Add include statements to your top module to reference the new files.
7. Run a test and check your results. If you open the irun.log file in an editor, you
should be able to track packets through the router and see the HBUS read and write
transactions.
If necessary, you can insert extra delays between the YAPP and HBUS sequences in the
virtual sequence to clearly separate transactions on the different interfaces.

Creating a Scoreboard Using TLM Lab 9A
Lab 9A Creating a Scoreboard Using TLM
Objective: To build a scoreboard using TLM.
For this lab, you will build and connect a scoreboard for the router, and create TLM analysis port
connections to hook up the scoreboard to the UVCs.
The router Module UVC is a complex design, so this lab has been deliberately broken down into
separate steps to build the UVC progressively.
First step is to implement the scoreboard component itself and connect it up to the YAPP and
Channel UVCs. For this part of the lab we assume all packets are sent to legal addresses with legal
payload length, i.e. the router does not drop any packets.
testbench
Channel0 port
UVC
Chan0 imp
Scoreboard Mon
YAPP
UVC
Chan1
Mon (also for

Chan2 other
channels)
1. We will be working with the testbench, testclass and top files from lab08. Copy these
files from lab08/tb into lab09a/tb. Work in the lab09a/tb directory.
2. A TLM analysis port has already been implemented in the Channel UVC monitors.
Check this to make sure you understand how it is written.
Note that the Channel UVC has a common monitor, channel_monitor.sv, for both
the TX and RX agents. The Channel monitor analysis port for collected YAPP packets is
named item_collected_port.
3. Modify yapp/sv/yapp_tx_monitor.sv to create an analysis port instance.
a. Declare an analysis port object, parameterized to the correct type.
b. Construct the analysis port in the monitor constructor.
c. Call the port write() at the appropriate point.

Lab 9A Creating a Scoreboard Using TLM
4. In the lab9a/sv directory, create the scoreboard, router_scoreboard.sv.
a. Define four analysis imp objects (for the YAPP and three Channels) using
ùvm_analysis_imp_decl macros and uvm_analysis_imp_* objects.
b. Add a constructor, and create analysis imp instances in the constructor.
c. Use the YAPP write()implementation to clone the packet and then push the
packet to a queue. Hint: Use a queue for each address.
d. Use Channel write()implementations to pop packets from the appropriate queue

and compare them to the channel packets.
e. Add counters for the number of packets received, wrong packets (compare failed) and
matched packets (compare passed).
f. Add a report_phase() method to print the number of packets received, wrong

packets, matched packets and number of packets left in the queues at the end of
simulation.
5. In the lab09a/tb directory, modify the router_tb.sv as follows:
a. Declare and build the scoreboard.
b. Add the TLM connections between YAPP, Channel, and scoreboard.
c. Use the same multichannel sequences as in lab08 to test your scoreboard

implementation, i.e. short packets with legal addresses, so that no packets are dropped
by the router.
d. Check that the simulation results are correct.

Router Module UVC (Optional) Lab 9B
Lab 9B Router Module UVC (Optional)
Objective: To create a module UVC for the router using the scoreboard
In reality, the scoreboard will only be one part of a larger router module UVC. For example the
router UVC may also contain reference models and coverage. All these components will be enclosed
in an env class.
In our example, we need to know the MAXPKTSIZE and ENABLE router settings so we know which
packets are dropped. We can implement this in a separate router reference model component.
The router reference connects to the YAPP and HBUS UVC analysis ports and selectively passes on
YAPP input packets to the scoreboard depending on the HBUS settings. An env wrapper will
instantiate both reference and scoreboard into a single router module UVC, as shown.
router module UVC
port
Chan0
imp
YAPP YAPP Scoreboard
Reference lid Chan1
HBUS
Chan2
1. We will be working with the files from lab09a. Copy these files from lab09a into
lab09b. Work in the lab09b directory
2. A TLM analysis port has already been implemented in the HBUS UVC monitor.
Note that the HBUS UVC has a common monitor, hbus_monitor.sv, for both the
master and slave agents. The HBUS monitor analysis port for collected
hbus_transaction's is named item_collected_port.
Check this to make sure you understand how it is written.
3. Create the router reference, router_reference.sv, in the lab09b/sv directory.
a. Extend from uvm_component.
b. Define two analysis imp objects for the YAPP and HBUS monitor analysis ports,
using ùvm_analysis_imp_decl macros and uvm_analysis_imp_*
objects. (Copy declarations from your scoreboard.) These are for input data to the
reference.

Lab 9B Router Module UVC (Optional)
c. Define one analysis port object for the valid YAPP packets. This is for output data to
the scoreboard.
d. Define variables to mirror the MAXPKTSIZE and ENABLE registers of the router and
update these in the HBUS write() implementation.
e. In your YAPP write() implementation, forward the YAPP packets onto the
scoreboard only if the packet is valid (router enabled; MAXPKTSIZE not exceeded;
address valid). Keep a separate count of invalid packets dropped due to size, enable
and address violations.
Note in a real-world example, we would have to explore race conditions when the
HBUS changes packet size during the processing of a packet. However there is a
REGMEM UVM feature specifically for handling and verifying registers in a design.
REGMEM is covered in the advanced UVM training.
4. Create the router module environment, router_module_env.sv, in the

lab09b/sv directory.
a. Declare and build the scoreboard and router reference components.
b. Connect the “valid YAPP” analysis port of the reference model to the YAPP analysis
imp of the scoreboard model.
5. In the tb directory, modify the router_tb.sv.
a. Replace the scoreboard declaration and build with the router module.
b. Modify the TLM connections for the YAPP and Channel analysis ports to allow for
the router_env layer.
c. Add a connection for the HBUS analysis port.
6. Create a router.svh file containing includes for all the router module UVC files, and
reference this in your top module.
a. Use the same multichannel sequences to test your scoreboard and system monitor
implementation.
b. Check the simulation results are correct and the scoreboard and monitor report the
right number of packets.
7. The router module can now be used as a standalone UVC. Copy your router module UVC
files to the router directory.

Further Work with TLM Connections (Optional) Lab 9C
Lab 9C Further Work with TLM Connections (Optional)
Objective: To use export TLM connections with the router module UVC
You must complete labs 9A and 9B before starting this lab.
A module UVC does not have such a clearly defined architecture as an interface UVC. With an
interface UVC, TLM analysis port objects are always defined in the monitor components. We
know where to look in order to find these. With a module UVC, it can be more difficult to find
declarations for the TLM analysis imp objects.
One technique is to extract all the TLM objects used in the module UVC to the top level
environment. These top-level objects are then connected internally to the correct sub-component.
router module UVC port

imp
Chan0
export
YAPP YAPP Scoreboard
Reference lid Chan1
HBUS
Chan2
1. Modify your router module UVC to present the external TLM connections as top-level
objects of the router module env. Note the following:
 You will need to declare export objects in the Router Module environment for the
imp objects of the monitor and scoreboard.
 You must connect all the export objects to imp objects in the Router Module
connect method.
 The internal YAPP valid connection between monitor and scoreboard does not need
to be routed to the Router Module environment.
 The testbench connections to the Router Module UVC will need to be modified.
2. Test your changes in simulation.

Lab 10 Writing Tests
Lab 10 Writing Tests
Objective: To explore structuring tests and create a generation scheme with an

UVM SystemVerilog verification environment.
To achieve this objective, use the yapp environment to create directed random tests.
Creating Virtual Sequences
1. Create a set of sequences that configures the device once, and sends packets with good
and bad parity and bad size.
Bad size is related to max_packet_size.
2. Change the max_packet_size on the fly and have the bad size sequence adjust
dynamically.
Making a Virtual Sequences and Scoreboard
3. Send 20-30 packets to the device and disable the DUT by writing to the enable register.
4. Wait for a few cycles and enable the device.
5. Verify using the waveform viewer that the device is operating properly.
Writing a Test
6. Write a test in which the distribution of packet size is as follows:

20: < (max_packet_size - 2)
30: (max_packet_size -1)
30: max_packet_size
20: >max_packet_size
7. Generate packets of this distribution with all generated sequences.

Writing Tests Lab 10
Writing a Test (Another Method)
8. Write an API for writing simple tests, directed or random.

For example, create a write and a read task for the host interface
write_hif(addr, data)
read_hif(addr, output data)
Tip: Make this task invoke ùvm_do_with(...) inside a sequence.
9. From the run_phase() method of a test, call a series of write_hif and read_hif
to test this simplified interface.
Would this mechanism work for engineers who want to do directed testing or for
designers that want a simple test writer interface?

Lab B Creating a Simple Functional Coverage Model
Lab B Creating a Simple Functional Coverage Model
Objective: To understand where to implement cover groups in UVM architecture.
To achieve this objective, you need to create a coverage model for the input packet traffic to collect
the following info:
 REQ1: Ensure all lengths of packets are sent into dut. Create buckets to detect MIN,
MAX, BABY, TEENY, GROWNUP packets.
 REQ2: Ensure all addresses received a packet, including illegal address.

 REQ3: Ensure all size packets were sent to all addresses with parity errors, except for
addr3.
 REQ4 (optional): Ensure that packets of different lengths are transmitted to all addresses
with zero delay between transmissions.
1. Create a covergroup in the yapp_tx_monitor.sv.
Remember the syntax for a covergroup instantiated inside a class is different than one
instantiated in a module. In a class, a covergroup instance is created by calling new() on
the covergroup name. A separate covergroup variable is not required:
function new (...);
super.new(...);
covergroup_name = new();
endfunction: new;
2. Decide how to sample coverage for packets sent to the DUT.
3. Create a coverpoint for REQ1 to sample the length field and create bins to reflect the
following ranges:
MIN = 1
MAX = 63
BABY in [2..10]
TEENY in [11..40]
GROWNUP in [41..62]
4. Run a simulation.
You’ll have to modify your irun file to enable functional coverage collection inside the
simulator. Use the coverage options as described in the appendix and online tool
documents.

Creating a Simple Functional Coverage Model Lab B
5. Analyze coverage results in ICCR.

By default, Incisive® Unified Simulator puts the coverage file icc.fcov in the
directory cov_work/design/test.
a. You can load this file in iccr by launching iccr -gui and loading the functional
coverage file from the file menu.
b. Select the Functional tab and see the distribution of packets observed.
6. Create a coverpoint for REQ2 and REQ3. Create a coverpoint inside the covergroup
created in step 1 for sampling address for REQ2 as follows:
a. Create a legal address bin to verify that all addresses were sampled.
If address 2 wasn’t generated, then it needs to be reflected in this coverpoint.
b. Create an illegal address bin that reflects how many packets were sent to address 3.
7. Create a cross inside the covergroup created in step 1 for coding REQ3 by creating a
cross with the appropriate fields.
How can you work around the limitation that we cannot have ignores in the cross?
8. Run a simulation and analyze coverage results in ICCR.
9. Optional: Create a coverpoint for REQ4.
a. Use a cross to model a transition, because the simulator currently doesn’t support
transitions on coverpoint.
Tip: You will have to keep track of the previous packet.
b. To implement back-to-back tracking you’ll have to make sure that the delay is
captured in the monitor.
Tip: You can use guard expression to control sampling of a cross.
10. Run multiple simulations with random svseed and analyze coverage results in ICCR.

Lab C1 Locating Cadence Online Support (COS) Solutions
Lab C1 Locating Cadence Online Support Solutions
Objective: Log in to Cadence Online Support (COS) and search for information
about a specific issue.
You can only complete this lab if you have access to the internet and a Cadence Online Support
account. If you do not, your instructor might be able to perform a demo of this lab for the class.
1. In a web browser enter

http://support.cadence.com
2. Log in to Cadence Online Support with your email and password.
The Support Home page appears.

Locating Cadence Online Support (COS) Solutions Lab C1
3. In the Support Home page, make sure the following options are selected:
 All Document Types
 All Products
4. In the Search field enter

UVM
and click the SEARCH button. A window containing the search results opens.
5. From the search results, select any of the matches. This will open a window providing the
information in the match title..
6. Close the solution window.

Note: You can filter the search results by selecting specific document types or products
which are listed to the left of the results.

Lab C2 Customizing Notification and Search Preferences
Lab C2 Customizing Notification and Search Preferences
Objective: Set preferences so you can improve search results and receive email
notification.
You can set product and other preferences for improved search results and email notification.
1. Click on the My Account link.
2. Click on the Notification Preferences tab.
3. On the Set Email Notification Preferences page:
a. If you are interested in receiving email notifications, check Send me email

notifactions about new product releases.
b. Click on Edit Product List and select your products of interest.
c. Specify your preferred email format.
d. Select the document types your are interested in and the frequency of delivery.
e. Click Save.
4. Click on the Search Preferences tab.
5. On the Set Search Preferences page:
a. Click on Use same product and document type preferences as my Notification

Preferences.
b. Click Save.

Systemverilog Advanced Verification Using Uvm: Engineer Explorer Series

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Systemverilog Advanced Verification Using Uvm: Engineer Explorer Series

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Systemverilog Advanced Verification Using Uvm: Engineer Explorer Series

Uploaded by

Copyright:

Available Formats

SystemVerilog® Advanced Verification

Allegro® Incisive® Silicon Express™

ii Cadence Design Systems, Inc. 5/22/2011

Introduction to the Labs .............................................................................................. 1

Project Overview .......................................................................................................... 2

Before You Begin ......................................................................................................... 8

Lab 1 Creating a Stimulus Model ..................................................................... 10

Lab 2 Creating a Simple UVC ........................................................................... 12

Lab 3 Adding Test Classes and Exploring Phases ........................................ 17

Lab 4 Using Factories ....................................................................................... 20

Lab 5 Generating UVM Sequences .................................................................. 22

Lab 6 Connecting to the DUT Using Virtual Interfaces .................................. 26

Lab 7 Integrating Multiple UVCs ...................................................................... 31

Lab 8 Writing Multichannel Sequences and System-level Tests .................. 34

Lab 9A Creating a Scoreboard Using TLM ........................................................ 36

Lab 9B Router Module UVC (Optional) .............................................................. 38

Lab 10 Writing Tests ........................................................................................... 41

Lab B Creating a Simple Functional Coverage Model .................................... 43

Lab C1 Locating Cadence Online Support Solutions ....................................... 45

Lab C2 Customizing Notification and Search Preferences .............................. 47

5/22/2011 Cadence Design Systems, Inc. iii

Introduction to the Labs

 Incisive® Unified Simulator release (9.2 or later)

Lab Database Files

 Classes must be defined in a module scope or package scope.

 Your reusable files are located in the <uvc>/sv/ directory.

 Non-reusable files are located in the <uvc>/examples or <uvc>/tb directory.

5/22/2011 SystemVerilog Advanced Verification Using UVM 1

Packet Router Description

High-Level Diagram – YAPP Router (Yet Another Packet Protocol)

in_data [7:0] data_vld_0

2 SystemVerilog Advanced Verification Using UVM 5/22/11

Packet Data Specification

YAPP Packet Structure

length addr byte 0 Header

parity byte N+1 Parity

5/22/2011 SystemVerilog Advanced Verification Using UVM 3

Input Port Protocol

sent packet Packet 1 Packet 2

H = Header, D = Data, P = Parity

4 SystemVerilog Advanced Verification Using UVM 5/22/11

Output Port Protocol (Channel Ports)

received packet Packet 1 (addr = 0)

H = Header, D = Data, P = Parity

5/22/2011 SystemVerilog Advanced Verification Using UVM 5

Packet Router DUT (Design Under Test) Registers

 MAXPKTSIZE – 8 bits – address 0 R/W - Reset value of ‘h3F

6 SystemVerilog Advanced Verification Using UVM 5/22/11

Host Interface Port Protocol (HBUS)

A WRITE operation takes one clock cycle as follows:

 hen is driven to 0 in the next cycle.

transaction write read

5/22/2011 SystemVerilog Advanced Verification Using UVM 7

Before You Begin

Finding Your Files

 Use the library provided with the lab files

Setting up UVM (Simulator Releases without UVM)