Ug936 Vivado Tutorial Programming Debugging
Ug936 Vivado Tutorial Programming Debugging
of this document
Lab 4: Using the Synplify Pro Synthesis Tool and Vivado Design
Suite to Debug a Design....................................................................................... 33
Step 1: Create a Synplify Pro Project.......................................................................................33
Step 2: Synthesize the Synplify Project................................................................................... 39
Step 3: Create DCPs for the Black Box Created in Synplify Pro........................................... 40
Step 4: Create a Post Synthesis Project in Vivado IDE.......................................................... 40
Step 5: Add More Debug Nets to the Project.........................................................................41
Step 6: Implementing the Design and Generating the Bitstream...................................... 44
Lab 10: Using the Vivado Serial Analyzer to Debug PS-GTR Serial
Links...............................................................................................................................150
IBERT PS-GTR Flow.................................................................................................................. 150
Step 1: Generating a Zynq UltraScale+ MPSoC PS Xilinx Support Archive....................... 152
Step 2: Using AMD Software Command-line Tool Flow to Generate a First Stage
Boot Loader.........................................................................................................................162
Step 3: ZCU102 Board Settings.............................................................................................. 163
Using FSBL with Serial I/O Analyzer to Bring Up IBERT PS-GTR........................................164
Troubleshooting...................................................................................................................... 171
Example RTL designs are used to illustrate overall integration flows between the Vivado logic
analyzer, ILA, and the Vivado Integrated Design Environment (IDE). To be successful using this
tutorial, you should have some basic knowledge of the Vivado tool flow.
TRAINING: AMD provides training courses that can help you learn more about the concepts presented in
this document. Use these links to explore related courses:
• Hardware, IP, and Platform Development: Creating the PL IP blocks for the hardware
platform, creating PL kernels, functional simulation, and evaluating the AMD Vivado™ timing,
resource use, and power closure. Also involves developing the hardware platform for system
integration. Topics in this document that apply to this design process include:
• Board System Design: Designing a PCB through schematics and board layout. Also involves
power, thermal, and signal integrity considerations. Topics in this document that apply to this
design process include:
• Lab 6: Using the ECO Flow to Replace Debug Probes Post Implementation
• Lab 7: Debugging Designs Using the Incremental Compile Flow
• Lab 8: Using the Vivado Serial Analyzer to Debug Serial Links
• Lab 10: Using the Vivado Serial Analyzer to Debug PS-GTR Serial Links
Objectives
These tutorials:
• Show you how to take advantage of integrated Vivado logic analyzer features in the Vivado
design environment that makes the debug process faster and simpler.
• Provide specifics on how to use the Vivado IDE and the Vivado logic analyzer to debug
common problems in FPGA logic designs.
• Provide specifics on how to use the Vivado Serial I/O Analyzer to debug high-speed serial
links.
• Validate and debug your design using the Vivado Integrated Design Environment (IDE) and the
Integrated Logic Analyzer (ILA) core.
• Understand how to create an RTL project, probe your design, insert an ILA core, and
implement the design in the Vivado IDE.
• Generate and customize an IP core netlist in the Vivado IDE.
• Debug the design using Vivado logic analyzer in real-time and iterate the design using the
Vivado IDE and a KC705 Evaluation Kit Base Board that incorporates an AMD Kintex™ 7
device.
• Analyze high-speed serial links using the Serial I/O Analyzer.
Getting Started
Setup Requirements
Before you start this tutorial, make sure you have and understand the hardware and software
components needed to perform the labs included in this tutorial.
Software
Hardware
• Pushbutton Switches: Serve as inputs to the de-bounce and control state machine circuits.
Pushing a button generates a high-to-low transition pulse. Each generated output pulse is
used as an input into the state machine.
• De-bounce Circuit: This example, when enabled, provides a clean pulse or transition from high
to low. Eliminates a series of spikes or glitches when a button is pressed and released.
• Sine Wave Sequencer State Machine: Captures and decodes input from the two push buttons.
Provides sine wave selection and indicator circuits, sequencing among 00, 01, 10, and 11 (zero
to three).
• LED Displays: GPIO_LED_0 and GPIO_LED_1 display selection status from the state machine
outputs, each of which represents a different sine wave frequency: high, medium, and low.
Lab 5 includes:
• An IBERT core
• A top-level wrapper that instantiates the IBERT core
Design Files
1. In your C: drive, create a folder called /Vivado_Debug.
2. Download the Reference Design Files from the AMD website.
CAUTION! The tutorial and design files can be updated or modified between software releases. You
can download the latest version of the material from the AMD website.
3. Unzip the tutorial source file to the /Vivado_Debug folder. There are six labs that use
different methodologies for debugging your design. Select the appropriate lab and follow the
steps to complete them.
• Lab 1: This lab walks you through the steps of marking nets for debugging in HDL and the
post-synthesis netlist (netlist insertion method). The following are the required files:
• debounce.vhd
• fsm.vhd
• sinegen.vhd
• sinegen_demo.vhd
• sine_high/sine_high.xci
• sine_low/sine_low.xci
• sine_mid/sine_mid.xci
• sinegen_demo_kc705.xdc
• Lab 2: This lab reviews the details of marking nets for debugging in the source HDL (HDL
instantiation method) and instantiating an ILA core in the HDL. The following are the required
files:
• debounce.vhd
• fsm.vhd
• sinegen.vhd
• sinegen_demo_inst.vhd
• ila_0/ila_0.xci
• sine_high/sine_high.xci
• sine_low/sine_low.xci
• sine_mid/sine_mid.xci
• sinegen_demo_kc705.xdc
• Lab 3: You can test your design even if the hardware is not physically accessible using a VIO
core. This lab walks you through the steps of instantiating and customizing a VIO core that
you hook to the I/Os of the design. The following are the required files:
• debounce.vhd
• fsm.vhd
• sinegen.vhd
• sinegen_demo_inst_vio.vhd
• sine_high/sine_high.xci
• sine_low/sine_low.xci
• sine_mid/sine_mid.xci
• ila_0/ila_0.xci
• sinegen_demo_kc705.xdc
• Lab 4: Nets can also be marked for debugging in a third-party synthesis tool using directives
for the synthesis tool. This lab walks you through the steps of marking nets for debugging in
the Synplify tool and using Vivado to perform the rest of the debugging. The following are the
required files:
• debounce.vhd
• fsm.vhd
• sign_high.dcp
• sign_low.dcp
• sine_mid.dcp
• sine_high.xci
• sine_low.xci
• sine_mid.xci
• sinegen.edn
• sinegen_synplify.vhd
• synplify_1.sdc
• synplify_1.fdc
• sinegen_demo_kc705.xdc
• Lab 5: Take designs created from Lab 1, Lab 2, Lab 3, and Lab 4 and load them onto the
KC705 board.
• Lab 6: Enhance post-implementation debugging by using the ECO flow to replace debug
probes.
• Lab 7: Use the Incremental Compile flow to enable faster debugging flows. Using the results
from a previous implementation run, this flow allows you to make debug modifications and
rerun implementation.
• Lab 8: Debug high-speed serial I/O links using the Vivado Serial I/O Analyzer. This lab uses
the Vivado IP example design.
• Lab 9: Use Vivado ILA core to debug JTAG-to-AXI transactions. This lab uses the Vivado IP
example design.
• Lab 10: Use the IBERT UltraScale+ PS-GTR transceiver to evaluate and monitor PS-GTR
transceivers in AMD Zynq™ UltraScale+™ MPSoC devices. This lab is purely software-based,
setting up and testing the processing system (PS) side of the Zynq UltraScale+ MPSoC with no
programmable logic (PL).
The relative locations of SMA cables on the board are shown in Setup Requirements.
Lab 1
10. In the Add Constraints dialog box, click the “+” sign, and click Add Files.
11. Navigate to /src/lab1 directory and select sinegen_demo_kc705.xdc. Click Next.
12. In the Default Part dialog box, specify the xc7k325tffg900-2 part for the KC705 platform.
You can also select Boards and select Kintex 7 KC705 Evaluation Platform. Click Next.
13. Review the New Project Summary page. Verify that the data appears as expected, per the
previous steps, and click Finish.
Note: It could take a moment for the project to initialize.
IMPORTANT! As an optional step, in the Settings dialog box, select Synthesis from the left and
change flatten hierarchy to none. The reason for changing this setting to none is to prevent the
synthesis tool from performing boundary optimizations for this tutorial.
2. In the Vivado Flow Navigator, expand the Synthesis drop-down list, and click Run Synthesis.
In the Launch Runs dialog box, accept all the default settings (Launch runs on local host), and
click OK.
Note: When synthesis runs, a progress indicator shows that synthesis is occurring. This could take a
few minutes.
3. In the Synthesis Completed dialog box, click Cancel, as shown in the following figure. You
implement the design later.
• Verilog:
(* mark_debug = "true" *) wire sine;
(* mark_debug = "true" *) wire sineSel;
This method lets you probe signals at the HDL design level. This can prevent optimization that
might otherwise occur to that signal. It also lets you pick up the signal tagged for post-
synthesis, so you can insert these signals into a debug core and observe the values on this
signal during FPGA operation. This method gives you the highest probability of preserving
HDL signal names after synthesis.
• Right-click and select Mark Debug or Unmark Debug on a synthesized netlist.
This method is flexible because it allows probing the synthesized netlist in the Vivado IDE and
allows you to add/remove MARK_DEBUG attributes at any hierarchy in the design. In
addition, this method does not require HDL source modification. However, there can be
situations where synthesis might not preserve the signals due to netlist optimization involving
absorption or merging of design structures.
• Use a Tcl prompt to set the MARK_DEBUG attribute on a synthesized netlist.
set_property mark_debug true [get_nets -hier [list {sine[*]}]]
In the following steps, you learn how to add debug nets to HDL files and the synthesized design
using Vivado IDE.
TIP: Before proceeding, make sure that the Flow Navigator on the left panel is enabled.
1. In the Flow Navigator under the Synthesis drop-down list, click Open Synthesized Design as
shown in the following figure.
2. In the Window drop-down menu, select Debug. When the Debug window opens, click the
window if it is not already selected.
3. Expand the Unassigned Debug Nets folder. The following figure shows those debug nets that
were tagged with MARK_DEBUG attributes in sinegen_demo.vhd.
4. In the Netlist window, select the Netlist tab and expand Nets. Select the following nets for
debugging as shown in the following figure.
• GPIO_BUTTONS_IBUF[0] and GPIO_BUTTONS_IBUF[1]: Nets folder under the top-level
hierarchy
• sel(2): Nets folder under the U_SINEGEN hierarchy
Note: These signals represent the significant behavior of this design and are used to verify and debug
the design in subsequent steps.
5. Right-click the selected nets and select Mark Debug as shown in the following figure.
6. Mark nets for debug in the Tcl Console. Mark nets “sine(20)” under the U_SINEGEN hierarchy
for debug by executing the following Tcl command.
set_property mark_debug true [get_nets -hier [list {sine[*]}]]
TIP: In the Debug window, you can see the unassigned nets you selected. In the Netlist window, you
can also see the green bug icon next to each scalar or bus, which indicates that a net has the attribute
mark_debug = true as shown in the following two figures.
3. In the Nets to Debug page, shown in the following figure, ensure that all the nets are added
for debug and click Next.
4. In the ILA Core Options page, go to Trigger and Storage Settings section and select both
Capture Control and Advanced Trigger. Click Next.
5. In the Setup Debug Summary page, make sure that all the information is correct and as
expected. Click Finish.
Upon clicking Finish, the relevant XDC commands that insert the ILA core(s) are generated.
2. In the Save Project dialog box, click Save. If a dialog box appears indicating this causes the
Synthesis results to go out of date, click OK. This applies the MARK_DEBUG attributes on
the newly marked nets. You can see those constraints by inspecting the
sinegen_demo_kc705.xdc file.
3. When the No Implementation Results Available dialog box pops up, click Yes. In the Launch
Runs dialog box, accept all of the default settings (Launch runs on local host) and click OK.
4. When the bitstream generation completes, the Bitstream Generation Completed dialog box
pops up. Click OK.
5. In the dialog box, asking you to close the synthesized design before opening the implemented
design. Click Yes.
6. Examine the Timing Summary report to ensure all the specified timing constraints are met.
Proceed to Lab 5: Using the Vivado Logic Analyzer to Debug Hardware to complete the rest
of the steps for debugging the design.
Lab 2
6. In the Add Constraints dialog box, click the “+” sign, and click Add Files.
7. Navigate to /src/lab1 directory and select sinegen_demo_kc705.xdc. Click Next.
8. In the Default Part dialog box, specify the xc7k325tffg900-2 part for the KC705 platform.
You can also select Boards and select Kintex 7 KC705 Evaluation Platform. Click Next.
9. Review the New Project Summary page. Verify that the data appears as expected per the
previous steps. Click Finish.
10. In the Sources window in Vivado IDE, expand sinegen_demo_inst to see the source files for
this lab.
Note: ila_0 core is added to the project.
11. Double-click the sinegen_demo_inst.vhd file, shown in the following figure to open it
and inspect the instantiation and port mapping of the ILA core in the HDL code.
2. The No Implementation Results Available dialog box appears. Click Yes. In the Launch Runs
dialog box, accept all of the default settings (Launch runs on local host) and click OK.
3. After bitstream generation completes, the Bitstream Generation Completed dialog box
appears. Open Implemented Design is selected by default. Click OK.
4. In the Design Timing Summary window, ensure all timing constraints are met.
Lab 3
VIO
CLK
PROBE_IN0[0:0] PROBE_OUT0[255:0]
X27387-111522
This lab walks you through the steps of instantiating and configuring the VIO core. It walks you
through the steps of connecting the I/Os of the design to the VIO core. This way, you can debug
your design when you do not have access to the hardware or the hardware is remotely located.
• One four-bit PROBE_IN0 port. This has two bits to monitor the two-bit Sine Wave selector
outputs from the finite state machine (FSM) and other two bits to mimic the state of the other
two LEDs on the board. You configure these four-bit signals as LEDs during run time to mimic
the LEDs displayed on the KC705 board.
• One two-bit PROBE_OUT0 port to drive the input buttons on the FSM. You configure it so
one bit can be used as a toggle switch during run time to mimic PUSH_BUTTON switch SW3,
and the second bit is used as PUSH_BUTTON switch SW6.
12. Review the New Project Summary page. Verify that the data appears as expected by the
previous steps. Click Finish.
Note: It might take a moment for the project to initialize.
13. In the Sources window in Vivado IDE, expand sinegen_demo_inst_vio to see the source
files for this lab.
Note: The ila_0 core is added to the project. However, vio_0 (the VIO core) is missing.
14. Instantiate and configure this VIO core as follows. From the Flow Navigator, click IP Catalog,
expand Debug & Verification, expand Debug, and double-click VIO. The Customize IP dialog
box opens.
15. On the General Options tab, leave the Component Name as its default value of vio_0, set
Input Probe Count to 1, Output Probe Count to 1, and select the Enable Input Probe Activity
Detectors check box.
17. On the PROBE_OUT Ports tab, set Probe Width to 2 and Initial Value to 0x0.
18. Click OK to generate the IP. The Generate Output Products dialog box appears. Click
Generate. An additional dialog box can appear, indicating an out-of-context module run is
launched. If so, click OK.
Output product generation should take less than a minute. At this point, you have finished
customizing the VIO. This core is already instantiated in the top-level design.
At this point, the Sources window should look as shown in the following figure.
19. Double-click sinegen_demo_inst.vhd in the Sources window to open it, and inspect the
instantiation and port mapping of the ILA core in the HDL code.
5. Proceed to Lab 5: Using the Vivado Logic Analyzer to Debug Hardware to complete the rest
of the steps for debugging the design. Proceed to the Verifying the VIO Core Activity (Only
Applicable to Lab 3) section in Lab 5 Step 2 to complete the rest of this lab.
Lab 4
5. If you get a dialog box asking you to create a non-existing directory, click OK.
6. In the left panel of the Synplify Pro window, click Add File as shown in the following figure.
7. In the Add Files to Project dialog box, change the Files of Type to HDL File. Navigate to
C:\Vivado_Debug\src\lab4, which shows all the VHDL source files needed for this lab.
Select the following three files by pressing the Ctrl key and clicking on them.
• debounce.vhd
• fsm.vhd
• sinegen_demo.vhd
8. Click Add.
9. In the same dialog box, set Files of type to Constraints Files. This shows the
synplify_1.sdc file. Select the file and click Add as shown in the following figure.
10. In the same dialog box, set Files of type to FPGA Constraint Files. This shows the
synplify_1.fdc file. Select the file and click Add as shown in the following figure. Click OK.
13. This includes the Implementation Options dialog box, as shown in the following figure. In the
Device tab, set Technology to Xilinx Kintex7, Part to XC7K325T, Package to FFG900, and
Speed to -2. Leave all the other options at their default values. Click OK.
14. You need to preserve the net names that you want to debug by putting attributes in the HDL
files. These attributes are already placed in the sinegen_demo.vhd, file of this tutorial.
Open the sinegen_demo.vhd file and inspect the lines shown.
15. You also can specify the MARK_DEBUG attributes in the source HDL files to mark the signals
for debugging, as shown in the code snippet from singen_demo.vhd file.
16. The synplify_1.sdc file contains various kinds of constraints, such as pin location, I/O
standard, and clock definition. The synplify_1.fdc file contains directives for the
compiler. Here is where the nets of interest marked for debugging are located. The attribute
and the nets selected for debugging are shown in the following figure.
In the previous constraints, sinegen is defined as a black box by using the syn_black_box
attribute. Second, the syn_no_prune attribute is used so that the I/Os of this block is not
optimized away. Finally, two nets, sine[20:0] and sel[1:0], are assigned the
MARK_DEBUG attribute such that these two nets should show up in the synthesized design
in AMD Vivado™ IDE for further debugging. For further information on these attributes, refer
to the Synplify Pro User Manual and Synplify Pro Reference Manual.
3. During synthesis, status messages appear in the Tcl Script tab. Warning messages are
expected, but there should not be any Error messages. To see detailed messages, click the
Messages tab in the bottom left-hand corner of the Synplify Pro console.
4. When synthesis completes, the output netlist is written to the file: rev_1/
sinegen_demo.edf
[Optional] To view the netlist select View → View Result File.
5. Click File → Save All to save the project, and click File → Exit.
5. Under Add Netlist Sources, click Add Files, navigate to the Vivado_Debug/synopsys/
rev_1 folder, and select sinegen_demo.edf. Click OK.
6. Add the netlist file created in the previous section. Click Add Files again, navigate to the
proj_synplify_netlist/proj_synplify_netlist.runs/synth1 folder and select
sinegen.dcp.
Add the DCP files created for the sub-module IPs in the previous section. Click Add
Directories again, navigate to the proj_synplify_netlist/
proj_synplify_netlist.srcs/sources_1/ip folder and select the following:
• sine_high
• sine_mid
• sine_low
Click OK in the Add Source Files dialog box. In the Add Netlist Sources dialog box ensure that
Copy Sources into Project is selected. Click Next.
7. Click Add Files, navigate to the Vivado_Debug/src folder, and select the
sinegen_demo_kc705.xdc file. This file has the appropriate constraints needed for this
Vivado project. Click OK in the Add Constraints File dialog box. In the Add Constraints
(optional) dialog box ensure that Copy Constraints into Project is selected. Click Next.
8. Under Default Part, select Boards and select Kintex 7 KC705 Evaluation Platform and the
right version number for your hardware. Click Next.
9. Under New Project Summary, ensure that all the settings are correct and click Finish.
10. In the Sources window, ensure sinegen_demo.edf is selected as the top module.
3. You should be able to see all the nets that are marked for debug, as shown in the following
figure.
2. Click through the wizard to create Vivado logic analyzer debug cores, keeping the default
settings.
Note: In the Specify Nets to Debug dialog box, ensure that all the nets marked for debug have the
same clock domain.
Lab 5
• How to debug the design using the AMD Vivado™ logic analyzer.
• How to use the currently supported Tcl commands to communicate with your target board
(KC705).
• How to discover and correct a circuit problem by identifying unintended behaviors of the
push-button switch.
• Useful techniques for triggering and capturing design data.
1. Connect the Digilent USB JTAG cable of your KC705 board to a USB port on a Windows
system.
Leave this cmd prompt open while the hw_server is running. Note the machine name you
are using. You use this later when opening a connection to this instance of the hw_server
application.
• Connecting to the Target Board Locally: If you plan to connect locally, ensure that the KC705
board is plugged into a Windows machine and perform the following steps:
1. Connect the Digilent USB JTAG cable of your KC705 board to a USB port on a Windows
system.
2. Ensure that the board is plugged in and powered on.
3. Power cycle the board to clear the device.
4. Turn the DIP switch positions (pin 1 on SW11, De-bounce Enable) to the OFF position.
2. The Hardware Manager window opens. Click Open Target → Open New Target.
Note: Depending on your connection speed, this can take about 10 to 15 seconds.
5. If there is more than one target connected, you see multiple entries in the Select Hardware
Target page. In this tutorial, there is only one target, as shown in the following figure. Click
Next.
6. In the Open Hardware Target Summary page, click Finish as shown in the following figure.
7. Wait for the connection to the hardware to complete. The dialog in the following figure
appears while hardware is connecting.
After the connection to the hardware target is made, the Hardware window appears as in the
following figure.
Note: The Hardware tab in the Debug view shows the hardware target and XC7K325T device detected
in the JTAG chain.
8. Next, program the XC7K325T device using the previously created .bit bitstream by right-
clicking the XC7K325T device and selecting Program Device as shown in the following
figure.
9. In the Program Device dialog box verify that the .bit and .ltx files are correct for the lab
that you are working on and click Program to program the device as shown in the following
figure.
CAUTION! The file paths of the bitstream and debug probes to be programmed are different for
different labs. Ensure that the relative paths are correct.
Note: Wait for the program device operation to complete. This can take a few minutes.
10. Ensure that an ILA core was detected in the Hardware panel of the Debug view.
11. The Integrated Logic Analyzer dashboard opens, as shown in the following figure.
2. In the Waveform window, verify that there is activity on the 20-bit sine signal as shown in
the following figure.
CAUTION! The waveform does not look like a sine wave. This is because you must change the radix
setting from Hex to Signed Decimal, as described in the following subsection.
1. Cycle the sine wave sequential circuit by pressing the GPIO_SW_E push button as shown in
the following figure.
2. Click Run Trigger Immediately again to see the new sine selected sine wave. You should see
the mid frequency as shown in the following figure. The sel signal also changed from 0 to 1
as expected.
Note: As you sequence through the sine wave selections, you can see that the LEDs do not light up in
the expected order. You debug this in the next section of this tutorial. For now, verify for each LED
selection that the correct sine wave displays.
Note: The signals in the Waveform window are re-arranged in the previous three figures.
Before starting the actual debug process, it is important to understand more about the sine wave
sequencer state machine.
• The input is a scalar signal called “button”. When the button input equals “1”, the state
machine advances from one state to the next.
• The output is a 2-bit signal vector called “Y”, and it indicates which of the four sine wave
generators is selected.
The input signal button connects to the top-level signal GPIO_BUTTONS_re[1], which is a low-
to-high transition indicator on the Sine Wave Sequencer button. The output signal Y connects to
the top-level signal, sineSel, which selects the sine wave.
As described earlier, the glitch reveals itself as multiple low-to-high transitions on the
GPIO_BUTTONS_IBUF_1 signal, but it occurs intermittently. Because it could take several
button presses to detect it, you now set up the Vivado logic analyzer tool to Repetitive Trigger
Run Mode. This setting makes it easier to repeat the button presses and look for the event in the
Waveform viewer.
CAUTION! For different labs the GPIO_BUTTONS_IBUF might show up differently or have a
different name such as button_in4_in. This might also show up as two individual bits or two bits
lumped together in a bus. Ensure that you are using bit 1 of this bus to set up your trigger
condition. For example, in case of a two-bit bus, you set the Value field in the Compare Value dialog
box to RX.
2. Select Enable Auto Re-trigger mode on the ILA debug core as shown as follows.
CAUTION! The ILA properties window can look slightly different for different labs.
When you issue a Run Trigger or a Run Trigger Immediate command after setting the Auto
Retrigger mode, the ILA core does the following repetitively until you disable the Auto
Retrigger mode option:
• Arms the trigger.
• Waits for the trigger.
• Uploads and displays waveforms.
3. On the KC705 board, press the Sine Wave Sequencer button until you see multiple
transitions on the GPIO_BUTTONS_IBUF_1 signal (this could take 10 or more tries). This is a
visualization of the glitch that occurs on the input. An example of the glitch is shown in the
following two figures.
CAUTION! You have to repeat the previous two steps repeatedly to see the glitch. After you see the
glitch, you can observe that the signal glitches are not exactly in the same location as shown in the
following figure.
1. Enable the de-bouncer circuit by setting DIP switch position on the KC705 board (labeled
De-bounce Enable in Figure 1) to the ON or UP position.
2. Enable the Auto-Retrigger mode on the ILA debug core and click RunTrigger on the ILA core,
and
• Ensure that you no longer see multiple transitions on the GPIO_BUTTON_re[1] signal on a
single press of the Sine Wave Sequencer button.
• Verify that the state machine is working correctly by ensuring that the sineSel signal
transitions from 00 to 01 to 10 to 11 and back to 00 with each successive button press.
5. Ensure that you are connected to the right target by selecting the target from the Hardware
Targets page. If there is only one target, that target is selected by default. Click Next.
6. In the Set Hardware Target Properties page, click Next.
7. In the Open Hardware Target Summary page, verify that all the information is correct, and
click Finish.
8. Program the device by selecting and right-clicking the device in the Sources window and
selecting Program Device.
9. In the Program Device dialog box, ensure that the bit file to be programmed is correct. Click
OK.
10. After the FPGA is programmed, you see the VIO and the ILA core in the Hardware window.
You now have a debug dashboard for the ILA core as shown in the following figure.
12. Make sure that there is activity on the sine [19:0] signal.
13. Select the sine signal in the Waveform window, right-click and select Waveform Style →
Analog.
14. Select the sine signal in the Waveform window again, right-click and select Radix → Signed
Decimal. You should be able to see the sine wave in the Waveform window.
15. Instead of using the GPIO_SW push button to cycle through each different sine wave output
frequency, you are going to use the virtual “push_button_vio” toggle switch from the VIO
core.
16. You can now customize the ILA dashboard options to include the VIO window. This allows
you to toggle the VIO output drivers and observe the impact on the ILA waveform window all
in one dashboard. Slide out the Dashboard Options window.
17. Add the VIO window to the ILA dashboard by selecting hw_vio_1.
Note: The ILA dashboard now contains the VIO window as well.
18. Adjust the Trigger Setup – hw_ila_1 window and the hw_vio_1 window so that they are side
by side, as shown in the following figure.
19. In the hw_vio_1 window, select the “+” button, and select all the probes under hw_vio_1.
20. Click OK.
Note: The initial values of all the probes.
22. Set the push_button_reset output probe by right-clicking push_button_reset and select
Toggle Button.
This toggles the output driver from logic from 0 to 1 to 0 as you click. It is similar to the
actual push button behavior, though there is no bouncing mechanical effect as with a real
push button switch.
24. Follow the previous step to change the push_button_vio to Toggle button as well.
25. Set these two bits of the “sineSel” input probe by right-clicking PROBE_IN0[0] and
PROBE_IN0[1] and selecting LED.
26. In the Select LED Colors dialog box, pick the Low Value Color and the High Value Color of
the LEDs as you desire and click OK.
27. When finished, your VIO Probes window in the Hardware Manager should look similar to the
following figure.
28. To cycle through each different sine wave output frequency using the virtual
“push_button_vio” from the VIO core, perform the following simple steps:
a. Toggle the value of the “push_button_vio” output driver from 0 to 1 to 0 by clicking on
the logic displayed under the Value column. You notice the sineSel LEDs changed
accordingly – 0, 1, 2, 3, 0, etc.
b. Click Run Trigger for hw_ila_1 to capture and display the selected sine wave signal from
the previous step.
Lab 6
Note: To learn more about using the ECO flow, refer to the Debugging Designs Post Implementation chapter
in the Vivado Design Suite User Guide: Programming and Debugging (UG908).
1. Open the Vivado Design Suite, and select File → Open Checkpoint.
2. Open the routed checkpoint that you created in Lab 2: Using the HDL Instantiation Method
to Debug a Design.
Change the layout in the Vivado Design Suite toolbar drop-down to ECO.
Note: The Flow Navigator window now changes to ECO Navigator with a different set of options.
3. In the ECO Navigator window, click Replace Debug Probes to bring up the Replace Debug
Probes dialog box. Note the Debug Hub and ILA cores in the design.
IMPORTANT! AMD strongly recommends that you do not replace the clock nets associated with ILA
and Debug Hub cores.
4. In the Replace Debug Probes dialog box, highlight the probes whose nets you want to
change. In this lab you replace the GPIO_BUTTONS_dly[0] net that is probed.
5. Click the Edit Probes button to the right of the GPIO_BUTTONS_dly[0] probe net to bring up
the Choose Nets dialog box.
6. In the Choose Nets dialog box, choose the U_DEBOUNCE_0/clear net to replace the existing
GPIO_BUTTONS_dly[0] probe net. Click OK.
7. Type for “*clear net” in the Name field and Click Find. Notice the U_DEBOUNCE_0 net in the
Found nets area. Select U_DEBOUNCE_0/clear net using the “->” arrow and click OK. The
U_DEBOUNCE_0/clear net to replaces the existing GPIO_BUTTONS_dly[0] probe net.
8. Now click OK in the Replace Debug Probes dialog. An additional dialog box can appear if the
nets were marked with DONT_TOUCH indicating that it must be removed to proceed. If so,
click Unset Property and Continue.
IMPORTANT! Check the Tcl Console to ensure that there are no Warnings/Errors.
9. Save your modifications to a new checkpoint. Use the Save Checkpoint As option in the ECO
Navigator to bring up the Save Checkpoint As dialog box. Specify a file name for the .dcp file
and click OK.
10. Click Write Debug Probes in the ECO Navigator. When the Write Debug Probes dialog
appears, click OK to generate a new .ltx file for the debug probes.
11. Click Generate Bitstream in the ECO navigator. When the Generate Bitstream dialog appears,
change the bit file name to project_sinegen_demo_routed_debug_changes.bit in
the Bit File field and click OK to generate a new .bit file that reflects the debug probe
changes.
12. Connect to the Vivado Hardware Manager by selecting Open Hardware Manager in the ECO
Navigator.
13. Connect to the local hardware server by following the steps in the Target Board and Server
Set Up section in Lab 5: Using the Vivado Logic Analyzer to Debug Hardware.
Program the device using the .bit file and.ltx files that you created in the previous steps.
14. Select Window → Debug Probes from the Vivado Design Suite toolbar. Ensure that the
probes that were replaced in the previous steps 8 and 9, are reflected in the probes
associated with hw_ila_1.
15. Run the Trigger on the ILA. Ensure the probes that were replaced in the previous steps 8 and
9, are reflected in the Waveform window as well.
Lab 7
Procedure
This lab consists of five generalized steps followed by general instructions and supplementary
detailed steps that allow you to make choices based on your skill level as you progress through
the lab.
If you need help completing a general instruction, go to the detailed steps it, or if you are ready,
simply skip the step-by-step directions and move on to the next general instruction.
Alternatively, you can use the following Tcl command to set the MARK_DEBUG property on
the signals specified.
set_property mark_debug true [get_nets [list {cpuEngine/
dcqmem_dat_qmem[*]}
{cpuEngine/dcpu_dat_qmem[*]} {cpuEngine/dcqmem_adr_qmem[*]}
{cpuEngine/du_dsr[*]} {cpuEngine/dvr0__0[*]} {cpuEngine/du_dsr[*]}
{cpuEngine/dcqmem_sel_qmem[*]}]]
9. In the Flow Navigator, click Set Up Debug to invoke the Set Up Debug wizard.
11. When the ILA Core Options screen appears, click Next again.
12. When the Set Up Debug Summary screen appears, ensure that one debug core is created and
click Finish.
13. Check the Debug window to ensure the u_ila_0 core is inserted into the design.
14. Save the new debug XDC commands by selecting File → Constraints → Save or clicking the
Save Constraints button.
TIP: When you re-run the implementation, the previous results are deleted. Save the intermediate
implementation results to a new directory or create a new implementation run for your incremental
compile to preserve the reference implementation run directory.
4. In the Design Runs window, right-click impl_1 and select Open Run Directory from the
popup menu. This opens the run directory in a file browser, as seen in the following figure.
The run directory contains the routed checkpoint (top_routed.dcp) to be used later for
the incremental compile flow. The location of the implementation run directory is a property
of the run.
5. Get the location of the current run directory in the Tcl Console by typing:
get_property DIRECTORY [current_run]
This returns the path to the current run directory that contains the design checkpoint. You
can use this Tcl command and the DIRECTORY property to locate the DCP files needed for
the incremental compile flow.
1. From the Vivado toolbar, select Flow → Create Runs to invoke the Create New Runs wizard.
2. In the Create New Runs screen, click Next.
3. The Configure Implementation Runs screen opens, as shown in the following figure. Select
the Make Active check box, and click Next.
4. From the Launch Options window, select Do not launch now and click Next.
5. In the Create New Runs Summary screen, click Finish to create the new runs.
The Design Runs window displays the new active runs in bold.
1. If you have closed the synthesized netlist, go back to the synthesized design using the Flow
Navigator.
2. For this tutorial, assume that you now need to debug some other nets beside the ones
already being debugged. However, you want to reuse the previous place and route results.
Now, you debug the nets fftEngine/fifo_out[*].
3. Apply the MARK_DEBUG property to this bus in the netlist window.
4. Click Set Up Debug to invoke the Set Up Debug wizard in the Flow Navigator.
5. In the Existing Debug Nets tab, select Continue debugging 110 nets connected to the
existing debug core.
7. Click Next and ensure the new nets are in the list of Nets to Debug.
8. Click Next and ensure that two debug cores are created, and click Finish.
9. Save the new debug XDC commands by clicking the Save Constraints button or selecting
File → Constraints → Save from the main Vivado toolbar.
1. Start by defining the design checkpoint (DCP) file to use as the reference design for the
incremental compile flow. This is the design from which the Vivado Design Suite draws
placement and routing data.
2. In the Design Runs window, right-click the impl_2 run and select Set Incremental
Implementation from the popup menu. The Set Incremental Implemenation dialog box opens.
3. Select Automatically use the checkpoint from the previous run.
4. Click OK. This information is stored in the INCREMENTAL_CHECKPOINT property of the
selected run. Setting this property tells the Vivado Design Suite to run the incremental
compile flow during implementation.
5. You can check this property on the current run using the following Tcl command:
get_property INCREMENTAL_CHECKPOINT [current_run]
TIP: To disable Incremental Compile for the current run, clear the INCREMENTAL_CHECKPOINT
property. This can be done using the Set Incremental Compile dialog box, or by editing the property
directly through the Properties window of the design run, or through the reset_property
command.
Note: This is an extremely small design. The advantages of the incremental compile flow are greater
and significant with larger, more complex designs.
8. Select the Reports tab in the Results window area and under Place Design, double-click
Incremental Reuse Report as shown in the following figure.
The Incremental Reuse Report opens in the Vivado IDE text editor. This report shows the
percentage of reused Cells, Ports, and Nets. A higher percentage indicates more effective
reuse of placement and routing from the incremental checkpoint.
In the report, fully reused nets indicate that the entire routing of the nets is reused from the
reference design. Partially reused nets indicate that some of the routing of the nets reuses
routing from the reference design. Some segments re-route due to changed cells, changed
cell placements, or both. Non-reused nets indicate that the net in the current design is not
matched in the reference design.
Conclusion
This concludes the lab. You can close the current project and exit the Vivado IDE.
In this lab, you learned how to run the Incremental Compile Debug flow, using a checkpoint from
a previously implemented design. You inserted a new debug core using the Set Up Debug wizard
on the synthesized netlist. You examined the similarity between a reference design checkpoint
and the current design by examining the Incremental Reuse Report.
Lab 8
The customizable AMD LogiCORE™ IP Integrated Bit Error Ratio Tester (IBERT) core for 7 series
FPGA GTX transceivers is designed for evaluating and monitoring the GTX transceivers. This core
includes pattern generators and checkers that are implemented in FPGA logic, and provides
access to ports and the dynamic reconfiguration port attributes of the GTX transceivers.
Communication logic is also included to allow the design to be run time accessible through JTAG.
• Create, customize, and generate an Integrated Bit Error Ratio Tester (IBERT) core design using
the Vivado tool.
• Interact with the design using Serial I/O Analyzer. This includes connecting to the target
KC705 board, configuring the device, and interacting with the IBERT/Transceiver IP cores.
• Perform a sweep test to optimize your transceiver channel and to plot data using the IBERT
sweep plot GUI feature.
Design Description
You can customize the IBERT core and use it to evaluate and monitor the functionality of
transceivers for a variety of AMD devices. The focus of this tutorial is on AMD Kintex™ 7 GTX
transceivers. Accordingly, the KC705 target board is used for this tutorial.
The following figure shows a block diagram of the interface between the IBERT Kintex 7 GTX
core interfaces with Kintex 7 transceivers.
• DRP Interface and GTX Port Registers: IBERT provides you with the flexibility to change GTX
transceiver ports and attributes. Dynamic reconfiguration port (DRP) logic is included, which
allows the runtime software to monitor and change any attribute in any of the GTX
transceivers included in the IBERT core. When applicable, readable, and writable registers are
also included. These are connected to the ports of the GTX transceiver. All are accessible at
runtime using the Vivado logic analyzer.
• Pattern Generator: Each GTX transceiver enabled in the IBERT design has a pattern generator
and a pattern checker. The pattern generator sends data out through the transmitter.
• Error Detector: Each GTX transceiver enabled in the IBERT design has a pattern generator and
a pattern checker. The pattern checker takes the data through the receiver and checks it
against an internally generated pattern.
DRP
DRP
Interface
TxN/TxP
External Serial
JTAG Pattern
Tx Data Loopback via
BSCAN Generator
SMA Cables
Kintex 7 GTX
Transceiver
Error
Rx Data
Detector
TxN/RxP
GTX Port
Detector Ports
X27386-111522
2. In the Quick Start screen, click Create Project to start the New Project wizard, and click Next.
3. In the Project Name page, name the new project ibert_tutorial and provide the project
location (C:/ibert_tutorial). Ensure that Create Project Subdirectory is selected. Click
Next.
4. In the Project Type page, specify the Type of Project to create as RTL Project. Click Next.
5. In the Add Sources page, click Next.
6. In the Add Existing IP page, click Next.
7. In the Add Constraints page, click Next.
8. In the Default Part page, select Boards and select Kintex 7 KC705 Evaluation Platform. Click
Next.
9. Review the New Project Summary page. Verify that the data appears as expected, per the
previous steps. Click Finish.
Note: It might take a moment for the project to initialize.
2. In the search field of the IP catalog type IBERT, to display the IBERT 7 series GTX IP.
3. Double-click IBERT 7 series GTX IP. This brings up the customization GUI for the IBERT.
4. In the Customize IP dialog box, choose the following options in the Protocol Definition tab:
a. Type the name of the component in the Component Name field. In this case, leave the
name as the default name, ibert_7series_gtx_0.
b. Ensure that the Silicon Version is selected as General ES/Production.
c. Ensure that the Number of Protocols option is set to 1.
d. Change the LineRate (Gb/s) to 8.
e. Change DataWidth to 40.
f. Change Refclk (MHz) to 125.
g. Ensure that the Quad Count is set to 2.
h. Ensure Quad PLL box is selected.
6. Click the Clock Settings tab and make the following changes for both QUAD_117 and
QUAD_118:
a. Leave the Source column at its default value of External.
b. Change the I/O Standard column to DIFF SSTL15.
c. Change the P Package Pin to AD12.
d. Change the N Package Pin to AD11.
e. Leave the Frequency (MHz) at its default value of 200.00.
7. Click the Summary tab and ensure that the content matches the following figure, click OK.
8. When the Generate Output Products dialog box opens, click Generate.
9. In the Sources window, right-click the IP, and select Open IP Example Design.
10. In the Open IP Example Design dialog box, and specify the location of your project directory.
Ensure that the Overwrite existing example project is selected and click OK.
Note: This opens a new instance of Vivado IDE with the new example design opened.
When the bitstream generation is complete, the Bitstream Generation Completed dialog box
opens.
2. Select Open Hardware Manager, and click OK.
1. Click Open New Target. When the Open Hardware Target wizard opens, click Next.
4. In the Open Hardware Target Summary page, review the options that you selected. Click
Finish.
5. The Hardware window in Vivado IDE should show the status of the target FPGA on the
KC705 board.
6. Select XC7K325T_0(0) in the Hardware window, right-click, and select Program Device.
7. The Program Device dialog box opens. Make sure that the correct .bit file is selected, and
click Program.
8. The Hardware window now shows the IBERT IP that you customized and implemented from
the previous steps. It contains two QUADS, each of which has four GTX transceivers. These
components of the IBERT are detected while scanning the device after downloading the
bitstream. If you do not see the QUADS select the XC7K325 device, right-click and select
Refresh Device.
9. Next, create links for all eight transceivers. Vivado Serial I/O analyzer is a link-based analyzer
that allows you to link between transmitter and receiver GTs within the IBERT design. For this
tutorial, link the TX and RX of the same channel. To create a link, right-click the IBERT Core in
the Hardware window and click Create Links.
11. Click the “+” button to add a new link. In the Link group description field, type Link Group
SMA. Select the Internal Loopback check box.
For the first link group, call this Link Group SMA, as this is the only transceiver channel linked
through the SMA cables. The new link shows up in the Links window.
12. Click Create Link again to create link groups for the rest of the transceiver pairs. To do this,
ensure that the transceiver pairs are selected, and click the + sign icon (add new link)
repeatedly until all the links are added to the link group Internal Loopback. Click OK.
13. After creating the links, they are added to the Links window as shown.
The Create Scan dialog box opens. In this dialog box, you can change the various scan
properties. In this case, leave everything to its default value and click OK. For more
information on the scan properties, see Vivado Design Suite User Guide: Programming and
Debugging (UG908).
17. The Create Sweep dialog box opens as follows. Various properties for the Sweep test can be
changed in this dialog box. Leave all the values to its default state and click OK.
Because there are four different Sweep Properties, each of these properties has three
different values (as seen in the Values to Sweep column), a total number of 81 sweep tests
are carried out. The Scans window shows the results of all the scans that are done for the
selected link.
CAUTION! Because there are 81 scans to be done, it could be a few minutes before all the scans are
complete.
To see the results of any scans performed, highlight the scan, right-click, and select Display
Scan Plots.
The Scan Plots window opens, showing the details of the scan performed.
Lab 9
The AMD LogiCORE™ IP JTAG-AXI core is a customizable core that can generate AXI
transactions and drive AXI signals internal to the FPGA at run-time. This supports all memory-
mapped AXI interfaces (except AXI4-Stream) and Lite protocol and can be selected using a
parameter. The width of the AXI data bus is customizable. This IP can drive any AXI4-Lite or
Memory-Mapped Slave directly. It can also be connected as master to the interconnect. Run-time
interaction with this core requires the use of the AMD Vivado™ logic analyzer feature.
Key Features
Additional Documentation
JTAG to AXI Master LogiCORE IP Product Guide (PG174) contains additional information
Design Description
This section has three steps as follows:
1. Creating a simple design in IP integrator that includes a System ILA and JTAG-to-AXI master.
2. Programming the AMD Kintex™ 7 FPGA KC705 Evaluation Kit Base Board and interacting
with the JTAG to AXI Master IP core.
3. Using the ILA Advanced Trigger Feature to Trigger on an AXI Read Transaction.
7. In the leftmost panel of the Flow Navigator, under Project Manager, click Create Block
Diagram. A dialog box appears that allows you to specify a block diagram name. You can
choose to specify a custom name or take the default. Click OK.
8. In the far right of the window is an empty block diagram design window labeled Diagram.
Click the + sign in the middle of the pane or the + toolbar button to bring up a search
window. In the Search field, type “JTAG to AXI” and double-click it to add the JTAG to AXI
Master to the block diagram.
9. The JTAG to AXI Master core appears on the IP integrator canvas. Double-click the core to
view the Customization dialog. Review the available settings and click OK to accept the
default core settings.
10. Following the same process from the previous step, add the additional IP to the block
diagram: AXI BRAM controller and Block Memory Generator. This creates a design using a
simple AXI infrastructure to create AXI transactions that demonstrate the debugging
capabilities of the System ILA core.
11. Before continuing. You need to customize AXI BRAM Controller and Block Memory
Generator. Begin by locating the AXI BRAM Controller in the block diagram canvas and
double-clicking on it. This invokes the Customization Dialog for the IP. Locate the Number of
BRAM interfaces and set the value to 1. Click OK.
12. Next, locate the Block Memory Generator in the block diagram and double-click as in the
previous step to invoke the Customization dialog. Clear Enable Safety Circuit check box. Click
OK.
13. At this point, the design should look like the following figure.
14. Notice the green banner indicating that Designer Assistance is available at the top of the
block diagram canvas. Click the Run Connection Automation button on this banner. When
the Connection Automation window appears, click the radio button for All Automation, and
click OK.
15. The Clocking Wizard and Processor System Reset and an AXI SmartConnect, are auto-
inserted into the design.
Note: The Clocking Wizard clock and reset inputs are disconnected, and the Run Connection
Automation banner persists. These inputs are connected to physical input ports on the FPGA and
wired to buttons on the KC705 board through the customization of the Clocking Wizard.
16. Invoke the Customization Dialog for the Clocking Wizard by double-clicking the IP in the
block diagram canvas. When the dialog appears, set CLKIN_1 to sys_diff_clk and
EXT_RESET¬_IN to reset. Click OK.
Note: Adding constraints for these ports is unnecessary because the project is generated using an
evaluation board as the target, and the IP allows the constraint information to be selected with the
sys_diff_clk.
17. Locate the green banner indicating that Designer Assistance is Available and click Run
Connection Automation. When the Run Connection Automation dialog appears, select the
button for All Automation. Click OK.
18. Now, sys_diff_clk and reset are connected to external ports. Examine the connectivity of
the design and it might be necessary to monitor AXI transactions between the JTAG to AXI
master and the AXI BRAM Controller slave. This is possible if a System ILA is added to probe
the AXI bus between the AXI BRAM Controller and the JTAG to the AXI master.
19. To add a System ILA to the design, click the Add IP (+) button as in the previous steps. Search
for System ILA and double-click to add it to the block diagram. When it appears in the block
diagram canvas, double-click on it to invoke the Customization Dialog. Ensure that both
Capture Control and Advanced Trigger are selected. Also, set the Number of Comparators to
the value 3. Click OK.
20. Connect the System ILA SLOT_0_AXI port and the S_AXI port on the AXI BRAM Controller.
Do this by clicking on the SLOT_0_AXI port and clicking again on the S_AXI port on the AXI
BRAM Controller.
21. When the Run Connection Automation banner appears, click it and select All Automation.
Click OK.
Note: The clk and resetn ports on the System ILA are connected to the AXI clock and the AXI reset.
22. In the upper left side of the Vivado IDE, click File → Save Block Design. Select File → Close
Block Design in the same menu to close the block design.
23. In the sources window, right-click on design_1 block design and select Create HDL Wrapper.
Allow Vivado IDE to manage the wrapper, and click OK.
24. In the Flow Navigator on the left side of the Vivado IDE, click Generate Bitstream.
25. Click OK to implement the design.
26. Wait until the Vivado Status window shows write_bitstream complete.
27. In the Bitstream Generation Completed dialog, select Open Hardware Manager, and click
OK.
Note: Depending on your connection speed, this can take about 10 to 15 seconds.
4. If more than one target is connected to the hardware server, you see multiple entries on the
Select Hardware Target page. In this tutorial, only one target is shown in the following figure.
Leave these settings at their default values, and click Next.
Wait for the connection to the hardware to complete. After the hardware target is connected,
the Hardware dialog shown in the following figure opens.
Note: The Hardware tab in the Debug view shows the hardware target and XC7K325T device detected
in the JTAG chain.
7. Next, program the previously created XC7K325T device using the .bit bitstream file by
right-clicking the XC7K325T device, and selecting Program Device as shown in the following
figure.
8. In the Program Device dialog, verify that the .bit file is correct for the lab on which you are
working. Click Program to program the device.
Note: Wait for the program device operation to complete. This can take a few minutes.
9. Verify that the JTAG to AXI Master and ILA cores are detected by locating the hw_axi_1 and
hw_ila_1 instances in the Hardware Manager window.
10. You can communicate with the JTAG to AXI Master core via Tcl commands only. You can
issue AXI read and write transactions using the run_hw_axi command. However, before
issuing these transactions, it is important to reset the JTAG to the AXI Master core. Because
the aresetn input port of the jtag_axi_0 core instance is not connected to anything, you need
to use the following Tcl commands to reset the core:
reset_hw_axi [get_hw_axis hw_axi_1]
11. The next step is to create a 4-word AXI burst transaction to write to the first four locations of
the BRAM:
set wt [create_hw_axi_txn write_txn [get_hw_axis hw_axi_1] -type WRITE -
address C0000000 -len 128 -data {44444444_33333333_22222222_11111111}]
where:
• write_txn is the name of the transaction.
• [get_hw_axis hw_axi_1] returns the hw_axi_1 object.
• -address C0000000 is the start address.
• -len 128 sets the AXI burst length to 128 words
• -data {44444444_33333333_22222222_11111111} is the data to be written.
Note: The data direction is MSB to the left (that is, address 3) and LSB to the right (that is, address 0).
The data is repeated from the LSB to the MSB to fill the entire burst.
12. The next step is to set up a 128-word AXI burst transaction to read the contents of the first
four locations of the AXI-BRAM core:
set rt [create_hw_axi_txn read_txn [get_hw_axis hw_axi_1] -type READ -
address C0000000 -len 128]
where:
• read_txn is the name of the transaction.
• [get_hw_axis hw_axi_1] returns the hw_axi_1 object.
• -address C0000000 is the start address.
• -len 128 sets the AXI burst length to 4 words.
13. After creating the transaction, you can run it as a write transaction using the run_hw_axi
command:
run_hw_axi $wt
14. After creating the transaction, you can run it as a read transaction using the run_hw_axi
command:
run_hw_axi $rt
4. In the New Trigger State Machine File dialog box, set the name of the state machine script to
txns.tsm.
5. A basic template of the trigger state machine script is displayed in the Trigger State Machine
gadget. Expand the trigger state machine gadget in the ILA dashboard. Copy the following
script after line 17 of the state machine script and save the file.
# The "wait_for_arvalid" state is used to detect the start
# of the read address phase of the AXI transaction which
# is indicated by the axi_arvalid signal equal to '1'
#
state wait_for_arvalid:
if (design_1_i/system_ila_0/U0/net_slot_0_axi_arvalid == 1'b1) then
goto wait_for_rready;
else
goto wait_for_arvalid;
endif
#
# The "wait_for_rready" state is used to detect the start
# of the read data phase of the AXI transaction which
# is indicated by the axi_rready signal equal to '1'
#
state wait_for_rready:
if (design_1_i/system_ila_0/U0/net_slot_0_axi_rready == 1'b1) then
goto wait_for_rlast;
else
goto wait_for_rready;
endif
#
# The "wait_for_rlast" state is used to detect the end
# of the read data phase of the AXI transaction which
# is indicated by the axi_rlast signal equal to '1'.
# Once the end of the data phase is detected, the ILA core
# will trigger.
#
state wait_for_rlast:
if (design_1_i/system_ila_0/U0/net_slot_0_axi_rlast == 1'b1) then
trigger;
else
goto wait_for_rlast;
endif
Note: Use the state machine to detect the various phases of an AXI read transaction:
7. In the Trigger Capture Status window, the ILA core waits for the trigger to occur and that the
trigger state machine is in the wait_for_a_valid state.
Note: The pre-trigger capture of 512 samples are completed successfully.
8. In the Tcl Console, run the read transaction you set up in the previous section of this tutorial.
run_hw_axi $rt
Note: The ILA core has triggered, and the trigger mark is on the sample where the axi_rlast signal is
equal to '1', as the trigger state machine program intended.
Lab 10
• Perform eye scans with raw pseudo-random binary sequence (PRBS) data patterns
• Measure Bit Error Ratio (no bit or error counters)
TIP: This is a supported feature in AMD Vivado™ Design Suite 2017.2 and above.
1. Generating Zynq UltraScale+ MPSoC PS Xilinx Support Archive (XSA) file from the Vivado
tool after configuring the PS-GTR.
2. Using the AMD Vitis™ AMD Software Command-line Tool (XSCT) flow to generate an FSBL
file using the XSA file.
3. Using the FSBL file with Vivado Serial I/O Analyzer to bring up IBERT PS-GTR.
Tools Required
• Vivado
• Vitis
• XSCT (Part of the Vitis tool)
Board/Part/Components Required
○ PCI Express 4x Male to PCIe 16x Female Riser Cable if PCIe card is larger than x4
• SATA
○ SanDisk 128 GB SATA SSD Drive
• USB
○ SanDisk Ultra 32 GB USB 3.0 Flash Drive
Required Files
• FSBL executable and linkable format file (ELF File) (Created using the following instructions),
which configures the PS-GTR
• Configuration Bitstream File (Optional file that can be needed to custom configure the FPGA
depending on the board setup)
• Tcl script to generate the FSBL and modify C-source for USB Support (when available)
Assumptions
3. Set your project name, and specify the project directory. Click Next.
4. Select Project type as RTL project.
5. Select do not specify sources at this time checked, and click Next.
6. To choose the board, click the board icon, and select Zynq UltraScale+ ZCU102 Evaluation
board, with Board Rev 1.0. Click Next.
9. An empty design diagram is displayed. Click the Add IP button to add IP. Select the IP based
on the selected board (for the ZCU102 evaluation board, search for Zynq UltraScale+
MPSoC) and double-click the selected IP.
10. In the design diagram window, select Run Block Automation. Click OK to continue creating
the ZCU102 design.
11. When the design diagram appears, use the following steps to validate the design:
a. Connect maxihpm0_fpd_aclk and maxihpm1_fpd_aclk together to pl_clk0, as
shown in the following figure.
i. Select maxihpm0_fpd_aclk and drag it to maxihpm1_fpd_aclk.
ii. Select maxihpm1_fpd_aclk and drag it to pl_clk0.
b. Right-click the Zynq UltraScale+ MPSoC block and select Validate Design to validate the
design. It says validation is successful. Click OK.
12. Customize the design by double-clicking the Zynq UltraScale+ MPSoC block and configuring
the parameters. There are four valid GT configurations for the ZCU102 board, as shown in
the following table.
13. Select the settings based on your requirements by double-clicking the Zynq UltraScale+
MPSoC block to customize GT Lane configuration.
14. Select I/O Configuration → High Speed. Select one of the four combinations using the
settings in the following screenshots.
a. PCIe – Display Port - USB - SATA (Default Vivado preset)
18. Leave the option Let Vivado manage wrapper and auto-update selected. Click OK in the
dialog to create the HDL wrapper.
19. Right-click design_1_i in the IP Sources tab, and click Generate Output Products.
20. Click Generate to generate with the default options in the panel.
21. After the generation is complete, click OK.
22. Select File → Export → Export Hardware.
23. In the Export Hardware Platform wizard, select a Fixed platform type. Click Next.
24. On the next page, select Pre-Synthesis for the platform output type.
25. Leave the XSA name as design_1_wrapper, and choose a location to store the exported XSA,
preferably in a new directory.
To create an FSBL for the Cortex-A53 #0 (64-bit) automatically (and modify the
xfsbl_main.c/h files if a USB is present) using the provided script, use the following steps:
1. Copy the src/lab10/xsct_create_fsbl.tcl script to the directory where the XSA file
is located. You can modify the Tcl script if you change the default name of the XSA file in the
Vivado tool. You can also change the script if the compiler options are different.
2. Open a terminal on Linux or command prompt on Windows.
3. Change the directory into the directory where the XSA file is located.
4. Call xsct from the Vitis tool install area.
% xsct xsct_create_fsbl.tcl
5. The location of the generated ELF File prints out when the script completes.
1. Make sure the following jumpers are correctly set for USB to be in HOST mode (refer to
ZCU102 Evaluation Board User Guide (UG1182)).
a. J7 – ON
b. J113 – 1-2
c. J110 – 2-3
2. Refer to the following image for jumper settings on a ZCU102 Rev 1.0 board.
4. Verify the ARM_DAP is visible in the hardware device list and click Next, and click Finish.
5. Right-click the ARM_DAP device in the hardware tree and select Configure IBERT GTR.
6. When the dialog box opens, you must provide the FSBL ELF file created in the previous steps
and optionally a configuration file (a bitstream, if your design requires one). You can also reset
the system before configuring with the Reset Zynq option checked. Click OK when done.
Note: The Reset Zynq option leaves the ARM_DAP in a bad state on early versions of Zynq UltraScale+
devices (for example, ZU9EG es1). If that occurs, power cycle the board and keep the Reset Zynq
option unchecked.
8. The Auto-detect links option does not work for PS-GTR. You can manually create links by
using Create Links as shown in the following figure.
9. Create links for all four lanes with each lane’s TX connected to the same lane’s RX, as shown
in the following figure.
Click OK when done.
10. The following figure shows the Serial I/O Links view where Status shows all the four lanes as
linked.
Note: The Link 1 PLL Status shows Not Locked, because it uses the Link 0 PLL as required by PCIe
protocol.
12. Select the appropriate parameters for EyeScan and perform the EyeScan. For example, the
following figure is performing EyeScan on Lane L1 (Link 1). Once the EyeScan completes, the
eye from -1UI to +1UI is displayed.
Note: Although the Create Scan pop up shows -0.5UI to +0.5UI, the EyeScan displayed is from -1UI to
+1UI.
Note: The value reported by Open UI % is a percentage of the entire horizontal axis, which is 2UI wide
for the PS-GTR transceiver.
Troubleshooting
Known Issues
1. By default, FSBL does not enumerate USB as that is something Linux drivers would do. To put
USB in link state without Linux, a small modification is required in the FSBL C-code. This
modification still does not enumerate the device, it only brings the USB into link state.
2. The EyeScan does not have a built-in time-out mechanism. If your link is poor (for example, if
L*_TM_DIG_8.EYESURF_ENABLE != 1), the EyeScan hangs without providing a user. No
results are returned in this case.
3. If the EyeScan progress is not moving, ensure the following parameters for all lanes are set for
EyeScan to function correctly.
Note: * represents the lane number (as in, for Lane 0, the parameter would be L0).
Click on the lane in the hardware tree and click on the properties tab. There is a search
button you can use to find the following properties.
a. L*_TM_MISC3.CDR_EN_FPL = 0
b. L*_TM_MISC3.CDR_EN_FFL = 0
c. L*_TM_DIG_8.EYESURF_ENABLE = 1
Also check the values of the following parameters, which ensures the Eye Scan circuit is
operational.
d. L*_PLL_LOCK = 1
e. L*_TM_SAMP_STATUS4.E_SAMP_PH0_CALIB_CODE is non-zero value
f. L*_TM_SAMP_STATUS5.E_SAMP_PH180_CALIB_CODE is non-zero value
Notes
1. As mentioned in Assumptions, IBERT PS-GTR uses the psu_cortexr5_0 core, so no other
applications should use this core.
2. TCM0 and TCM1 memory are combined to form a unified memory for IBERT PS-GTR. Any
other processor core should not access this memory while IBERT PS-GTR is running.
3. The error counter is 16-bits, and the sample counter is 32-bits. Each sample can have 8 bits
of error count. Therefore on the edges, the error counter can saturate with a sample count
value of 8192. PS-GTR does not stop the sample counter even if the error counter saturates.
A prescale=0 produces 8192 samples and thus total samples of 8192 *8 (65536), and thus
the outside edges of the eye could show a BER of e-01 or less depending on the prescale
selected.
4. The EyeScan assumes there is a link present. If there is no link, the EyeScan cannot complete.
Canceling the EyeScan stops the command sequence, but the state of the previous point scan
is unknown.
5. If you run EyeScan and because of no link, the EyeScan does not complete, set the register
L*_TM_MISC_ST_0.EYE_SURF_RUN to 0 for the given lane before you run the EyeScan again.
6. If you run EyeScan on a lane that is either powered down or Display Port, it immediately
stops, and the scan is marked as incomplete. EyeScan does not work in either scenario.
Appendix A
The AMD Technical Information Portal is an online tool that provides robust search and
navigation for documentation using your web browser. To access the Technical Information
Portal, go to https://docs.amd.com.
Documentation Navigator
Documentation Navigator (DocNav) is an installed tool that provides access to AMD Adaptive
Computing documents, videos, and support resources, which you can filter and search to find
information. To open DocNav:
• From the AMD Vivado™ IDE, select Help → Documentation and Tutorials.
• On Windows, click the Start button and select Xilinx Design Tools → DocNav.
• At the Linux command prompt, enter docnav.
Note: For more information on DocNav, refer to the Documentation Navigator User Guide (UG968).
Design Hubs
AMD Design Hubs provide links to documentation organized by design tasks and other topics,
which you can use to learn key concepts and address frequently asked questions. To access the
Design Hubs:
Support Resources
For support resources such as Answers, Documentation, Downloads, and Forums, see Support.
Revision History
The following table shows the revision history for this document.
Copyright
© Copyright 2012-2024 Advanced Micro Devices, Inc. AMD, the AMD Arrow logo, Kintex,
UltraScale+, Versal, Vitis, Vivado, Zynq, and combinations thereof are trademarks of Advanced
Micro Devices, Inc. AMBA, AMBA Designer, Arm, ARM1176JZ-S, CoreSight, Cortex, PrimeCell,
Mali, and MPCore are trademarks of Arm Limited in the US and/or elsewhere. PCI, PCIe, and PCI
Express are trademarks of PCI-SIG and used under license. The DisplayPort Icon is a trademark
of the Video Electronics Standards Association, registered in the U.S. and other countries. Other
product names used in this publication are for identification purposes only and may be
trademarks of their respective companies.