AN 311: ASIC-to-FPGA Design Methodology and Guidelines

AN 311: ASIC-to-FPGA Design Methodology and Guidelines
AN 311: Standard Cell ASIC to FPGA Design
Methodology and Guidelines
AN-311-3.1
© April 2009
Introduction
The cost of designing traditional standard cell ASICs is increasing every year. In
addition to non-recurring engineering (NRE) and mask costs, development costs are
increasing due to design complexity. Issues such as power, signal integrity, clock tree
synthesis, and manufacturing defects can add significant risk and time-to-market
delays. FPGAs offer a viable and competitive option to traditional standard cell ASIC
development by reducing the risk of re-spins, high NRE costs, and time-to-market
delays.
Programmable logic has progressed from being used as glue logic to today’s FPGAs,
where complete system designs can be implemented on a single device. The number
of gates and features has increased dramatically to compete with capabilities that
have traditionally only been offered through ASIC devices. Figure 1 illustrates the
evolution of FPGA applications that have led to higher density devices, intellectual
property (IP) integration, and high-speed I/O interconnects technology. All of these
elements have allowed FPGAs to play a central role in digital systems
implementations. With the availability of multimillion-gate FPGA architectures, and
support for various third-party EDA tools, you can use a design flow similar to that of
traditional standard cell ASIC devices to create system-on-a-programmable-chip
(SOPC) designs in FPGAs. In addition, with the Altera® Stratix® series of FPGAs, you
have a seamless path to HardCopy® ASICs.
Figure 1. Application of FPGA Devices from 1985 to the Present
ESL
Design
SOPC
Design
Increasing
PLD
Complexity
Complex
Control
Control
Logic
Glue
Logic
1985
© April 2009
Altera Corporation
C2H
Electronic System
Level Design
Block-based
Design
Second-generation Synthesis
IP Megafunctions
Synthesis
Macrofunctions
Equations
Schematics
1990
1995
2003
2006
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 2
Introduction
With the device sizes and architectures that are available today, FPGAs can effectively
implement systems that were once possible only in traditional standard cell ASICs.
Because of their programmable capability, FPGAs reduce the time to bring up a
system as well as minimize the financial risk involved with new designs. Some of the
newer FPGA devices have resources such as on-chip transceivers for different
physical layer (PHY) protocols, providing the capability to interface with external
memories and implement large blocks of internal memory. All of these aspects help to
reduce device count on a board, in turn bringing down the cost associated with the
product.
Cyclone® series FPGAs provide a low-cost alternative for applications that currently
use low-to-moderate-density traditional standard cell ASICs. A rich feature set makes
Cyclone series FPGAs suitable for a broad range of applications including displays,
wireless communication, video and image processing, automotive, and military at a
cost per device that is comparable to traditional standard cell ASICs.
Stratix series FPGAs provide a good alternative to more complex traditional standard
cell ASICs. These are high-performance, high-end FPGAs that have resources such as
large internal memory, fast external memory interfaces, on-chip transceivers, and a
large number of internal clock networks. Stratix GX series devices and Arria™ GX
devices have on-chip transceivers that support a number of industry-standard serial
interfaces in addition to other logic resources. Stratix series FPGA devices, which
includes Stratix, Stratix II, Stratix III and Stratix IV, also provides a cost effective
migration option for going to volume production with HardCopy ASIC offerings
from Altera. By choosing to go to HardCopy ASICs, you can utilize the programmable
feature during the design, verification, and prototyping stages, reduce the time to
market, and get the cost, power, and performance benefits in production. You can also
reduce the typical problems associated with traditional ASICs, such as low yield and
issues with testability and reliability.
f
For more details about HardCopy ASICs, and for design guidelines for targeting your
designs to HardCopy series devices, refer to the HardCopy device handbooks at
www.altera.com.
This document is intended for traditional standard cell ASIC designers considering
FPGA or HardCopy ASIC implementation of their designs, either for prototyping or
for production. This application note gives some guidelines for efficient FPGA
implementation of their designs. It also discusses the salient features of Altera’s
Quartus® II software that make it easy for building true
system-on-a-programmable-chip (SOPC) solutions. Comparisons between typical
FPGA and traditional standard cell ASIC design flow are also made whenever
applicable.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Traditional Standard Cell ASIC and FPGA Design Flows
Page 3
Traditional Standard Cell ASIC and FPGA Design Flows
Typical traditional standard cell ASIC and FPGA design flows are shown in Figure 2.
The back-end design of a traditional standard cell ASIC device involves a wide
variety of complex tasks, including placement and physical optimization, clock tree
synthesis, signal integrity analysis, and routing using different EDA software tools.
When compared to the design of traditional standard cell ASIC devices, the physical,
or back-end design of Altera FPGAs is very simple and is accomplished with a single
the Quartus II software. The Quartus II software is a fully integrated,
architecture-independent package offering a full spectrum of logic design capabilities
for designing with Altera FPGAs.
By using Altera FPGAs in place of traditional standard cell ASICs, you can potentially
reduce the complexity of the design process as well as significantly reduce cost. This
document discusses and compares each of the tasks involved in the design flows for
both FPGA and traditional standard cell ASIC devices.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 4
Design Specification
Figure 2. A Comparison of Traditional Standard Cell ASIC and FPGA Design Flows
FPGA
Design Flow
Tasks
Design
Specifiation
ASIC
Design Flow
Tasks
Design
Specification
RTL
(Verilog HDL
IP Instantiator)
Project Planning,
I/O Assignments
and Analysis,
Preliminary Power
Estimation
Design
Development/
Physical and
Power Planning
Functional
Simulation
Create a
Floor Plan
Synthesis
(Area, Power,
and Timing)
RTL
(Verilog HDL
IP Instantiator)
Functional
Simulation
Synthesis/Physical
(Area, Power, and
Timing)
Timing and
Power Analysis
Place and Route
Test Synthesis
(Scan Insertion,
BIST Synthesis)
Formal
Verification and
Timing Analysis
Sign-off
Placement and
Physical
Optimization
Formal
Verification and
Post-Layout
Timing Analysis
Power Analysis
Static Timing
Analysis
Board-Level
Signal Integrity
Formal
Verification
Quartus II
In-System
Debugging Tools
Clock Tree
Synthesis
In-System
Verification
Routing
Signal
Integrity
Skew and Timing
Analysis
DRC/ERC, Manual
Layout Fixes,
Hold Time Fixes
IR drops, X-talk
Sign-off
In-System
Verification
Design Specification
The design specification stage includes the following activities, which will impact the
choice of FPGA device:
■
I/O specification
■
System timing requirements and timing constraints
■
Memory requirements
■
Verification methodology
■
Selection of the FPGA family and device, including the speed grade
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Specification
Page 5
I/O Specification
Altera devices support a wide variety of I/O standards. I/O resources vary
depending on the device and family. When you design a traditional standard cell
ASIC, you can instantiate I/O pads for a design by specifying the technology I/O
buffers in a Verilog HDL or VHDL file to perform simulation and synthesis. At the
foundry, the I/Os specified in the RTL are replaced with the technology I/O pads.
In an Altera FPGA design flow, you choose the type, location, and I/O settings for all
the pins in your design with the Pin Planner, which is part of the Quartus II software.
With the Pin Planner, you can validate your I/O assignments by performing legality
checks on your design’s I/O pins and surrounding logic. These checks include proper
reference-voltage pin usage, valid pin location assignments, and acceptable mixed
I/O standards.
As part of I/O planning, especially with high-speed designs, you should take
board-level signal integrity and timing into account. When you have an FPGA device
with high-speed interfaces on a board, the quality of the signal at the far end of the
board route, as well as the propagation delay in getting there, are vital for proper
system operation.
The Quartus II software provides features to take these factors into consideration,
making the software “board-aware.” The Quartus II software can take into account
board routing and external devices to generate advanced timing reports and board
simulation modeling files.
The Quartus II software provides the following methods of signal integrity analysis:
f
■
I/O timing using a default or user-specified capacitive load with no signal
integrity analysis (default)
■
The Enable Advanced I/O Timing option utilizing a user-defined board trace
model to produce enhanced timing reports from accurate, “board-aware”
simulation models
■
Full board routing simulation in third-party tools using IBIS or HSPICE I/O
models generated by the Quartus II software
For additional information about using the Pin Planner, refer to the I/O Management
chapter in volume 2 of the Quartus II Handbook.
Some Altera device families have dedicated circuitry associated with I/O cells to
support multiple channels of serial transceivers. The on-chip transceiver circuitry
provides physical coding sublayer (PCS) and physical media attachment (PMA)
implementation for PHY protocols, such as PCI Express (PCIe), Gigabit Ethernet,
XAUI, SPI, and SONET. By using an FPGA that has integrated transceivers you may
be able to reduce the overall device count on your boards and reduce cost.
Number of I/O Pins
Determine the exact number of I/O pins used in your design. With this information,
you can select a specific device and package.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 6
Design Specification
Location of I/O Pins
Carefully analyze the impact of the I/O pin locations on the board layout to minimize
potential problems. On an FPGA, you must know the I/O standards that are
supported by a device so that modules in the design that require a certain I/O
standard can be physically placed near the corresponding I/O bank. Some Altera IP
cores (such as PCIe) specify pinout constraints automatically that may be needed to
meet any special requirement. Use the Pin Planner to specify the locations for all other
I/Os. If you anticipate a change in the target device at a later time in the design cycle,
the Pin Planner also helps you plan for device migration.
f
To learn more about the location of I/O banks and supported standards within an I/O
bank for a specific device, see the corresponding Altera device handbook.
I/O Timing
I/O timing is affected by the I/O standard and the drive strength and pin loading. In
addition, to meet I/O requirements, Altera FPGAs have I/O pins with fast, dedicated
registers. The device uses these registers, depending on the clock setup time (tSU) and
clock-to-output delay (tCO) requirements. You can enable the use of fast I/O registers
by setting the Fast Input Register and Fast Output register options using the
Assignment Editor in the Quartus II software.
1
Redefine I/O specifications for every new design because different FPGA families
may support different I/O standards. Even within a device family, different devices
have different numbers of I/O pins.
Starting with the Quartus II software version 7.0, you can use the Advanced I/O
Timing feature to specify the board parameters for each I/O individually so that the
timing analysis is very accurate.
f
For more details on this feature, refer to the Quartus II TimeQuest Timing Analyzer
chapter in volume 3 of the Quartus II Handbook.
Memory Requirements
You must identify embedded RAM capacity and speed requirements early in the
design specification process. The size of embedded memory you can implement in an
FPGA depends on the selected device. Stratix series devices support between 4 MBits
and 27 MBits of embedded memory. Cyclone series devices support between
288 KBits to 3.8 MBits of embedded RAMs.
f
Refer to the appropriate device handbook for details about available embedded RAM.
Memory capacity in traditional standard cell ASIC technology is higher than in FPGA
technology, but integrating and testing memory in the standard cell ASIC device
requires significantly more work, because it generally involves instantiating a special
module in the register transfer level (RTL) design, as well as creating custom circuitry
around the RAM blocks for testing. For Altera FPGAs, the Quartus II software allows
you to arrange memory blocks to meet the system requirement automatically without
having to create any special blocks for testing. The In-System Memory Content Editor
in the Quartus II software helps verify the contents of the memories in the FPGA
when you are in the test phase.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Specification
Page 7
For more information about how to address your memory requirements in Altera
FPGAs, refer to “Specification of External and Internal Memory” on page 15.
f
For more details about the memory organization in a specific device, refer to the
appropriate device handbook.
Verification Methodology
Most of the traditional standard cell ASIC projects start with an exhaustive
verification plan and design specifications that are completely frozen. A similar
strategy of starting a verification plan early in the design cycle is desirable for today’s
complex FPGA designs. Decisions such as whether to use formal verification methods
should be made early. The Altera design flow supports industry standard simulation
and formal verification tools to aid you with verification.
A key difference in verification methodology between standard cell ASIC-based
systems and FPGA/HardCopy ASIC-based systems is the ability to utilize in-system,
at-speed verification of your hardware and software, enabling true software and
hardware co-design and co-verification. This change in the verification methodology
of your chip and system designs is a key contributor to saving millions of dollars and
months of time in getting your system to market.
FPGA Device Sizing
Estimate the size of your design so that you can arrive at the power and heat
dissipation requirements. After the size of the logic is estimated and operating
frequencies are specified, the approximate size and speed grade of the device can be
selected.
Altera FPGAs also support vertical migration within the same package. In this
context, you can migrate between devices whose dedicated pins, configuration pins,
and power pins are the same for a given package across device densities. For example,
you can migrate a design from an EP1S10B672C6 device to an EP1S20B672C6 device
in the 672-pin package.
Number of Low-Skew Signals
Identify the number of clocks required in your design to choose the FPGA for your
implementation. There is usually an upper limit to the number of different low-skew
signals available in a device. Clock tree insertion is a manual process in the traditional
standard cell ASIC flow, whereas in an FPGA flow it is an automatic process. In most
cases, the Quartus II software can automatically choose which low-skew resources to
assign to which signals. In addition, it can also determine which signals should be
routed on low-skew resources. However, you can make these assignments manually.
Altera FPGAs provide a variety of clocking resources in each device. Clocks that are to
be used in all regions of the chip can be routed over global clock nets, which are
guaranteed to be low skew. The global clock networks span the entire general purpose
logic arrays, feeding all architectural structures. Internal logic, PLL outputs, or device
inputs might drive these low-skew global clock resources. You can also use these
low-skew global resources for other device-wide signals with large fan-outs, such as
asynchronous resets or clock enables. The number of device wide low-skew resources
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 8
Design Specification
available is device dependent. Stratix devices provide fast regional clocks in addition
to the global low-skew resources. Similarly, Stratix II, Stratix III, and Stratix IV devices
provide a hierarchical clocking structure consisting of global clocks (GCLKs), regional
clocks (RCLKs), and peripheral clocks (PCLKs). You can use a combination of these
low-skew signals to route clocks and other high fan-out signals in your design.
In addition to the clock networks described previously, Stratix GX series devices
feature separate clock distribution resources that connect directly to the clocking
resources of the device logic array. This architecture makes it flexible for reference
clock generation, clock domain translation, and support of multi-channel
functionality.
1
The Quartus II software turns off unused clock networks to reduce power dissipation
in the device. It can also disable unused portions of clock networks for Stratix II,
Stratix III, and Stratix IV devices, and can optimize the placement to minimize the
portion of the clock network used to further reduce power.
f
For more information about regional clocks and fast regional clocks, refer to the Area
and Timing Optimization chapter in volume 2 of the Quartus II Handbook.
f
For more information about the clocking resources such as GCLK, PCLK, and RCLK
within a specific device, refer to the appropriate device handbook.
For more information about clock tree synthesis and clock requirements, refer to
“Clock Tree Synthesis” on page 23.
Phase-Locked Loop (PLL) Requirements
Inserting PLL circuitry into a traditional standard cell ASIC device is typically a
manual process in which you instantiate special PLL blocks in the design. In the
Altera FPGA flow, you can create configurable PLLs with the MegaWizard® Plug-In
Manager, available in the Quartus II software. You can control PLL parameters such
as phase shift, clock switchover, and PLL bandwidth using the MegaWizard Plug-In
Manager. In the PLLs available in Altera devices, each output of the PLL can be
programmed independently, creating customizable clock frequencies that are
independent of other input or output clocks. Inherent jitter filtration and fine-grained
control of the configurable range help you generate the high-performance precision
clocks required in your system. The number of PLLs available in FPGA technology is
usually limited, whereas PLL quantities are virtually unlimited in traditional standard
cell ASIC technology.
f
For more information about the number and type of PLLs supported in a specific
device, refer to the appropriate chapter on Clock Networks and PLLs in the Altera
device handbooks.
Clock Frequencies
The frequency at which your design is expected to operate is an important factor in
choosing the appropriate Altera FPGA for your design. The maximum frequency of
operation is affected by several factors, such as logic utilization, routing congestion,
and the speed grade of the chosen device.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Development
f
Page 9
Refer to the appropriate Altera device handbook to determine the maximum clock
frequency for a given device and speed grade.
Number of Simultaneously Switching Outputs (SSOs)
The number and the placement of SSOs has a direct impact on the number of power
and ground pins required for a traditional standard cell ASIC. This is usually not an
issue with an FPGA, because the placement and number of power and ground pins is
pre-determined. The Quartus II software helps you with placement of user pins. You
can check the pin placement based on the I/O rules with I/O assignment analysis.
Power Requirements
The Quartus II software supports two power analysis methods: the PowerPlay Early
Power Estimator (EPE), which is a spreadsheet-based tool, and the PowerPlay Power
Analyzer, which is a part of the Quartus II software. Performing preliminary power
analysis, based on the estimated logic size and speed, helps you determine the device
power requirements. This usually leads to defining the device’s cooling and package
requirements.
When you design a traditional standard cell ASIC, you might have to use stand-alone
power analysis tools to estimate the power dissipation in your design, and a good
estimate might not be available until late in the design cycle. When you target your
design to an Altera FPGA, you can use the PowerPlay EPE to get an early estimate of
power dissipation in your design. You can run the more elaborate PowerPlay power
analysis later in the design cycle. You can also optimize synthesis and the Fitter to
focus on reducing power consumption by turning on the Extra Effort PowerPlay
option in power driven compilation.
f
The Quartus II PowerPlay Power Analyzer helps you get fast and accurate power
consumption in your design. For more details about using this tool, refer to the
PowerPlay Power Analysis chapter in volume 3 of the Quartus II Handbook.
f
For more details about power driven synthesis and other power saving optimization
techniques, refer to the Power Optimization chapter in volume 2 of the Quartus II
Handbook.
Design Development
Design development includes the following steps for both FPGAs and traditional
standard cell ASICs:
© April 2009
■
Top-down or bottom-up methodology selection
■
RTL coding
■
Specification of the external and internal memory
■
Synthesis
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 10
Design Development
Methodology Selection
Altera’s FPGA design methodology supports both top-down and bottom-up design
methodology. FPGA design flows support modular design approaches for bottom-up
methodology and hierarchical design partitioning for top-down design methodology,
similar to the process used for traditional standard cell ASIC devices. Altera’s design
software also supports newer standards such as SystemVerilog that are becoming a
part of traditional standard cell ASIC design methodology. For team-based designs
that are typical of ASIC-size designs, the Quartus II software supports incremental
compilation methodology. For designs using HardCopy ASICs to achieve higher
performance than the companion FPGA, it is recommended that you use the
HardCopy first flow.
Incremental Compilation
In an incremental compilation flow, you can split a large design into smaller
partitions. Team members can work on partitions independently, which can simplify
the design process and reduce compilation time. The Quartus II incremental
compilation feature preserves the results and performance for unchanged logic in
your design as you make changes elsewhere, allowing you to perform more design
iterations per day and achieve timing closure more efficiently.
If you want to take advantage of the compilation time savings and performance
preservation of Quartus II incremental compilation, plan for an incremental
compilation flow from the beginning of your design cycle. Good partition and
floorplan design helps lower-level design blocks meet top-level design requirements,
reducing the time spent integrating and verifying the timing of the top-level design.
f
For more information about design planning and different design approaches, refer to
the Design Planning with the Quartus II Software and the Quartus II Incremental
Compilation for Hierarchical and Team Based Design chapters in volume 1 of the
Quartus II Handbook.
f
For further information about using different approaches in design methodology
when using third-party synthesis tools, refer to the corresponding chapter in the
Synthesis section of the Quartus II Handbook, volume 1.
RTL Coding
RTL coding style becomes important when you target your design to an FPGA
because the resources in a device are finite. This application note provides some
coding guidelines to improve design performance by taking advantage of the FPGA
architecture.
This section provides the following guidelines for the RTL coding of your design
when you target an Altera FPGA:
■
Synchronous design practices versus asynchronous designs
■
Synchronous versus asynchronous resets
■
Gated clocks versus clock enables
■
Divided clocks
■
Using data pipelining
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Development
f
Page 11
■
Using encoding schemes
■
Using look-ahead techniques
■
Using logic duplication
■
Using internal buses
Only some of the RTL coding guidelines are discussed here. For more details about
coding guidelines, refer to the Design Recommendations for Altera Devices and the
Quartus II Design Assistant chapter and the Recommended HDL Coding Styles chapter in
volume 1 of the Quartus II Handbook.
Synchronous Design Practices Versus Asynchronous Designs
Use synchronous design practices to avoid glitches and race conditions inherent in
asynchronous design. Conforming to a synchronous design style also makes it easier
to perform timing analysis and achieve timing closure.
f
For a detailed discussion about synchronous design practices, refer to the Design
Recommendations for Altera Devices and the Quartus II Design Assistant chapter in the
Synchronous FPGA Design Practices section in volume 1 of the Quartus II Handbook.
Another problem caused by asynchronous signals is that the sampling clock at the
receiving end might be too slow compared to the initiating clock. The slow clock at
the destination might cause a transition on a signal to be completely missed.
Implement a good handshake protocol between asynchronous blocks to ensure
proper operation.
The following typical design styles contribute to asynchronous designs:
■
Gated clocks
■
Latch inferences
■
Multiple clocks
■
Derived clocks
Synchronous Versus Asynchronous Resets
An asynchronous reset is defined as a way of clearing the contents of a register,
independent of the associated clock. Standard cell ASIC libraries consist of registers
with and without a built-in reset/clear pin. A register with a built-in reset/clear pin is
generally bigger than a register without one.
Sometimes, traditional standard cell ASIC designers use registers without
asynchronous reset pins to obtain extra speed and reduce area in a design by using an
external gate on the data path of the register for a reset. When the reset is routed
through the data pin, the clock must be running when the reset is asserted.
Additionally, a synchronous reset signal is treated as any other data signal, so no extra
care is required during the routing and timing optimization phase.
In traditional standard cell ASIC designs, internally generated asynchronous resets
from a state machine can cause problems in scan testing. A typical problem results
from the shifting of test vectors through the flipflops of the state machine, which
triggers unintended resets. This is not an issue for FPGAs, which do not have to be
scan tested.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 12
Design Development
You can also control the asynchronous reset from an input pin on the device. Such a
reset signal should be buffered like a clock tree to reach all of the asynchronous reset
pins of all the device registers. In traditional standard cell ASICs, a reset tree is
generated just like a clock tree. You must use static timing analysis or use an
independent start signal to prevent different state machines from becoming
out-of-sync immediately after the reset release when you are designing a standard cell
ASIC.
In FPGAs, a reset tree is already in place. All registers have a built-in asynchronous
reset, so no area savings is attained by not utilizing the available reset capability.
Gated Clocks Versus Clock Enables
Clock gating is used in traditional standard cell ASIC designs for power optimization.
However, gated clocks can introduce glitches and register incorrect data due to the
delay on the combinational path.
Registers in FPGAs have a clock enable pin. You can avoid clock gating by using the
clock enable pin. Using the clock enable pin to disable the clocking of the flipflop does
not mean that the flipflop is not clocked. It means that the current state of the flipflop
is clocked continuously. Using this clock enable does not reduce the power
consumption of your design. An example of this implementation of the flipflop is
shown in Figure 3.
Figure 3. A Clock Enable Pin Disabling Clocking of a Flipflop
DFF Cell
D
Din
Q
Q
CE
clk
clk
The following RTL code segment shows how to infer a flipflop with the clock enable
shown in Figure 3 on page 12:
module clock_en(in,out,clk,data_en);
input in,clk,data_en;
output out;
reg out;
always @(posedge clk) begin
if (data_en)
out <= in;
else
out <= out;
end
endmodule
f
Refer to the appropriate device datasheet for device specific information about
clocking structures.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Development
Page 13
If reducing the power is the primary reason for using clock gating in your design, you
can use dedicated resources available in Altera devices to perform clock gating. Altera
FPGAs from the Cyclone III, Cyclone II, Stratix IV, Stratix III, and Stratix II families
have dedicated clock control blocks to perform clock gating. You can use these blocks
to shut down selected clock networks. Altera recommends that you use the dedicated
circuitry when available rather than using a multiplexing logic structure.
1
Stratix IV and Stratix III devices use Altera’s innovative Programmable Power
Technology, which lets you select higher performance or lower power on a logic cell
basis and reduce the overall power.
f
For more details about power management with Stratix III devices, refer to the
Stratix III Power Management Design Guide. For more details about power management
for Stratix IV devices, refer to AN 514: Power Optimization in Stratix IV FPGAs.
f
If you must use clock gating, follow the recommended clock gating method described
in the “Gated Clocks” section in the Design Recommendations for Altera Devices and the
Quartus II Design Assistant chapter in volume 1 of the Quartus II Handbook.
Divided Clocks
Many designs require clocks generated by division of a master clock. If you have to
create a divided clock, Altera recommends that you use the dedicated PLL circuitry
for clock division. If you use clock division logic, make sure you use synchronous
counters or state machines to perform clock division. Using ripple clock division can
make timing analysis difficult.
f
For more details on internally generated clocks, refer to the Design Recommendations
for Altera Devices and the Quartus II Design Assistant chapter in volume 1 of the
Quartus II Handbook.
Using Data Pipelining
Because FPGAs have many registers, you can use the registers for a pipelined
architecture to improve the device performance without incurring any area penalty.
1
When you pipeline your design, make sure you equalize data and control path
latency, and modify the testbenches to capture the outputs at the right time.
Using Encoding Schemes
Almost all of today’s designs contain a number of state machines. While designing
traditional standard cell ASICs it is common to use some sort of binary encoded state
machine to reduce area. On the other hand, FPGAs are register rich, and using a
one-hot encoded state machine to reduce the combinational logic required between
different states. Because of this, one-hot encoding generally produces better
performance results for your state machines in your designs.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 14
Design Development
Using Look-Ahead Techniques
Look-ahead techniques force a portion of the large combinational logic function into
the previous clock cycle, and force the remaining logic function to be performed in the
next clock cycle. This technique, also known as register balancing, balances the levels
of logic between registers, resulting in much faster execution. Thus, by splitting the
combinational logic over two clock cycles, your logic can run faster without
additional latency.
f
The Quartus II software can perform register retiming without any changes in the
RTL if the corresponding physical synthesis option is set. For more details about
setting this option, refer to the “Gate Level Register Retiming” section in the Power
Optimization chapter in volume 2 of the Quartus II Handbook.
Using Logic Duplication
You can improve FPGA performance by minimizing the routing delays. High fan-out
is one of the sources of routing delays. You can reduce the number of high fan-out
registers by logic duplication. The Quartus II software supports logic duplication as
part of the netlist optimization. You can choose to do logic duplication in your RTL as
well. Some synthesis tools might see the logic replication in the RTL code as
redundant, and might optimize the replicated logic. Synthesis attributes might be
required to keep the intended replicated logic.
f
For further information about the available Physical Synthesis Optimization options,
refer to the Netlist Optimization and Physical Synthesis chapter in volume 2 of the
Quartus II Handbook.
f
For more details about register duplication and register preservation attributes for
synthesis tools, Refer to the Quartus II Integrated Synthesis chapter in volume 1 of the
Quartus II Handbook.
Using Internal Buses
Internal buses in traditional standard cell ASIC devices allow various internal
modules and external devices to communicate. It is not good design practice to have
internal tristate buses. Implement tristate functions only at the I/O level.
If your code contains internal tristate inference, the Quartus II software implements it
using multiplexers. However, if your design contains multiple partitions, and you use
incremental compilation, the Quartus II software might not be able to infer the correct
multiplexing logic due to the lack of visibility. Therefore, it is a good practice not to
use internal tristates. Altera FPGAs support tristate buses through the I/O interface to
communicate with various on-board devices.
f
For more details about coding guidelines, refer to the Design Recommendations for
Altera Devices and the Quartus II Design Assistant and the Recommended HDL Coding
Styles chapters of the Quartus II Handbook, volume 1.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Development
Page 15
Specification of External and Internal Memory
Memory block sizes and configurations are major considerations for SOPC designs.
Performance requirements determine whether to have device memory on- or off-chip.
The versatile high-memory bandwidth internal memory can implement a variety of
memory functions such as RAM, CAM, FIFO, true dual-port memory, synchronous
memory, and asynchronous memory. The amount of internal memory that can be
implemented in an Altera FPGA is device dependent.
f
To learn more about the available memory configurations and features available in a
specific device, refer to the corresponding device handbook.
f
For more details on using the internal memory as FIFOs, refer to the Single- and
Dual-Clock FIFO Megafunction User Guide.
Memory Implementation—Flexibility and Efficiency
If a memory configuration does not fit in a single memory block, the Quartus II
software implements the required configuration by combining two or more memory
blocks. For example, if the design requires a 512 × 18-bit memory block, but the only
blocks available are 512 × 9-bit, the Quartus II software can combine two 512 × 9-bit
blocks.
f
For more information about memory configurations and the Quartus II software
settings, see the Altera website and AN 207: TriMatrix Memory Selection Using the
Quartus II Software.
Instantiating Altera RAM in Place of a Standard Cell ASIC RAM
In traditional standard cell ASIC design flow, memory is instantiated using models
provided by the foundry.
The memory model provided by the foundry contains timing information for
synthesis and a behavioral model for simulation. If the design has to be ported to an
Altera FPGA, you must configure the memory using the MegaWizard Plug-In
Manager. The wizard generates all the files required for synthesis and simulation of
the memory blocks.
Timing information for the memory models, also known as clear box models, can be
generated by the Quartus II software. Third-party synthesis tools can perform an
accurate area and timing estimate with the clear box models generated by the
Quartus II software.
f
1
For a more detailed discussion of RTL coding for efficient mapping of memory in
Altera devices, refer to the Design Planning with the Quartus II Software chapter in
volume 1 of the Quartus II Handbook.
When designing with Altera FPGAs, you must select the device during the design
planning process, because the available memory resource varies according to the
family and device.
The Stratix series of devices support only synchronous memory interfaces. However,
it is possible to operate these synchronous memories in a pseudo-asynchronous mode
if the design already contains asynchronous references to memory blocks.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 16
Design Development
f
For more information about synchronous memories, refer to AN 210: Converting
Memory from Asynchronous to Synchronous for Stratix & Stratix GX Designs.
External Memory Interfaces
For many of today’s high-memory bandwidth applications, the memory must reside
outside the standard cell ASIC or FPGA to provide the required buffer size. The Altera
IP MegaStore has pre-verified memory controllers that make it easy for you to design
a system with external memory. Altera’s complete memory interface design solutions
address today’s high-speed memory interface challenges such as memory controller,
I/O design, and board level signal integrity issues. Altera’s solutions include
advanced FPGA architectures, customizable MegaCore® functions, Quartus II design
software, reference designs, demonstration boards, and simulation models. Altera
supports interfaces to SDR SDRAM, DDR SDRAM, DDR2 SDRAM, DDR3 SDRAM
and reduced latency DRAM (RLDRAM) with clock frequencies of up to 400 MHz,
achieving a throughput up to 1600 Mbps.
f
For more details about different external memory interfaces available with different
Altera FPGA families, refer to the documentation and reference designs at Altera’s
External Memory Solutions Center.
Synthesis
Synthesis is the process of converting a design representation from RTL code to a
gate-level netlist. The Quartus II software has an integrated synthesis engine. Altera
also supports the use of third-party synthesis tools such as Synplify Pro and Precision
RTL from Mentor Graphics. Figure 4 shows a typical synthesis flow. Compared to
traditional standard cell ASIC tools, FPGA synthesis tools are much easier to use in
terms of complexity and scripting. These tools support all of the popular standard cell
ASIC synthesis techniques, including the following techniques:
f
■
Top-down or bottom-up approach
■
Modular design flow
■
Scripting
For more information about design methodologies with different synthesis tools
supported in the Altera design flow, refer to the Synthesis section in volume 1 of the
Quartus II Handbook.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Development
Page 17
Figure 4. Typical Synthesis Design Flow
VHDL
Verilog HDL
Functional
Simulation
Tech.lib
Constraints
Synplify Pro,
Precision Synthesis,
or Quartus II
Gate-level
Simulation
Synthesis
Post-synthesis
simulation files
Technologyspecific netlist
Forward-annotated
timing constraints
Quartus II
software
Post Placeand-Route
Simulation
Placeand-Route
Post place-and-route
simulation files
No
Timing
requirements
met?
Yes
Configuration files
( .sof / .pof )
Configure device
Third-Party EDA Tool Support for Synthesis
In addition to the Quartus II integrated synthesis engine, you can use the following
third-party tools to perform synthesis:
f
© April 2009
■
Synplify Pro and Synplify from Synopsys’ Synplicity Business Group
■
LeonardoSpectrum and Precision Synthesis from Mentor Graphics
■
Design Compiler from Synopsys
For more details about using third party synthesis tools, refer to the corresponding
chapter in the Synthesis section in volume 1 of the Quartus II Handbook.
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 18
Design Development Tools
Design Development Tools
Altera provides you with the following development tools, all of which are available
within the Quartus II software:
■
SOPC Builder
■
DSP Builder
■
MegaWizard Plug-In Manager
■
C2H Compiler
You can use these tools to ease SOPC design development and potentially reduce
time-to-market. The following sections present an overview of these tools, and their
advantages.
SOPC Builder
SOPC Builder provides a standardized, graphical environment for creating SOPC
designs composed of components including CPUs, memory interfaces, standard
peripherals, and user-defined peripherals.
SOPC Builder enables the combination of components such as embedded processors,
standard peripherals, IP cores, on-chip memory, interfaces to off-chip memory, and
user-defined logic into a custom system module. This tool generates a single system
module that instantiates these components, and automatically generates the necessary
bus logic to connect them. SOPC Builder also generates RTL design files, which you
can use for the functional simulation of your system.
Figure 5 shows an example of a typical set of system components with which SOPC
Builder can be used to generate a system-level module in minutes. SOPC Builder also
generates the test bench.
Figure 5. Elements of a Typical SOPC Design
Nios
Processor
Memory
Controller
Memory
DMA
Controller
UART
PCI
Figure 5 shows the SOPC Builder user interface after a set of peripherals are
connected to a Nios® processor.
f
For more details about the SOPC Builder, refer to SOPC Builder in volume 4 of the
Quartus II Handbook.
f
You can get more details about various IP Megafunctions available on the Altera.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Design Development Tools
Page 19
DSP Builder
Altera FPGAs have digital signal processing (DSP) blocks, such as dedicated
multiplier blocks, and support embedded processors. These features make Altera
FPGAs very suitable for use in such applications. In addition, the Altera design flow
provides tools that make designing for DSP much easier.
The Altera DSP portfolio consists of proven, high-performance, standard algorithms
and functions created to meet today’s rapidly evolving technologies. Every function
in Altera’s MegaCore function library has been rigorously tested and meets the
exacting requirements of various industry standards. Altera provides an extensive
portfolio of drop-in DSP functions. The DSP portfolio includes everything you require
to build SOPC solutions. You can choose blocks of IP from a comprehensive range of
standard DSP functions to create a complete SOPC solution. You can instantiate each
function multiple times in different designs. The IP library includes functions such as
filters, transforms, encoders, error detection and correction circuits, and video and
image processors.
Altera’s DSP Builder integrates the Quartus II software with high-level algorithmic
development tools such as MATLAB and Simulink software. The DSP Builder
software helps you create the hardware representation of a DSP design in an
algorithm-friendly development environment. Figure 6 shows a simple design, built
from concept-to-implementation, using the DSP Builder. Several ready-to-use
mathematical functions are available within both the MATLAB and Simulink tools,
along with simulation models, which can be used to create a schematic. DSP Builder
can generate a RTL description of the design, along with the test bench.
Figure 6. DSP Builder Design Flow
Drag and drop
Generate VHDL,
Synthesize, and
place-and-route
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 20
IP Availability and Flow
The MegaWizard Plug-In Manager
The MegaWizard Plug-In Manager helps you create or modify design files that
contain custom megafunction variations, which can then be instantiated in a design
file.
C2H Compiler
With the evolution of HDLs, digital systems are being described at increasingly higher
levels of abstraction. Newer standards such as SystemVerilog and SystemC are
offshoots of this development. Many traditional standard cell ASIC designers now
model their designs in a high-level language such as C for verification of the
architectural design and model implementation. Several design tools are now
available to help you to take your concept from a high-level language to hardware
implementation, thereby reducing the design cycle.
To support the high-level design flow, Altera provides the C2H Compiler as part of
the Nios II Embedded Design Suite. This tool creates custom hardware for
functionality that would otherwise require processor usage. Creating a dedicated
logic in hardware can improve the execution performance. You can decide on the
blocks required for the acceleration requirement you may decide to generate. The
C2H Compiler takes code written in ANSI-style C and maps it into resources in an
Altera FPGA.
f
For more details on using the Nios II C2H Compiler, refer to the Nios II C2H Compiler
User Guide.
IP Availability and Flow
IP blocks that are pre-verified reduce the design time, solve many time-to-market
issues, and simplify verification.
Altera has an extensive offering of specialty IP cores which you can incorporate
directly into your design. These include embedded processors, specialty
communications interfaces, and memory controllers. These pre-verified cores help
you reduce the time to market. These IP cores are designed to take advantage of
Altera’s device architecture, thus ensuring optimal fit results. You can also find many
reference designs using these IP cores on Altera’s website.
f
For a list of IP cores available for Altera devices, go to
www.altera.com/literature/lit-ip.jsp.
Altera delivers IP cores as a single MegaCore® IP library package that is built into the
Quartus II software download file and installation process. The MegaWizard Plug-In
Manager allows you to quickly and easily view documentation, specify parameters,
set up third-party tools, and generate the files necessary for integrating the
parameterized Altera IP core into a design
f
For more information about the IP design flow, refer to the Designing With Altera
Intellectual Property page on the Altera website.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Functional Simulation
Page 21
Functional Simulation
Functional simulation verifies the functionality of the RTL design. The following
third-party EDA tools are supported:
■
NC-Sim from Cadence
■
VCS and VCS-MX from Synopsys
■
ModelSim® from Mentor Graphics®
■
Active-HDL and Riviera-PRO from Aldec
f
When designing ASICs, you may run gate-level simulation and/or static timing
analysis on a synthesized netlist even before you perform place and route. However,
for Altera FPGA designs, run static timing analysis with the TimeQuest Timing
Analysis only after place and route. Refer to “Gate-Level Simulation” on page 22 and
“Static Timing Analysis”. You can generate a functional simulation netlist that does
not contain any timing information if you want to run a functional simulation of the
design after place and route to use with your RTL simulation test bench.
f
For additional and detailed information about performing functional simulation with
different industry standard simulators, refer to the Simulation section in volume 3 of
the Quartus II Handbook.
Test Synthesis
The manufacturer tests its FPGA devices for manufacturing defects, eliminating the
memory BIST, SCAN insertion, or other tests requirements typically used to detect
manufacturing faults in a traditional standard cell ASIC. This completely eliminates
the complex task of test synthesis.
Place-and-Route
Use the Quartus II Fitter to place-and-route your design targeted to an Altera FPGA.
The Quartus II software reads standard EDIF, VHDL, and Verilog HDL netlist files,
and generates VHDL and Verilog HDL netlist files for a convenient interface to other
industry-standard EDA tools.
The Quartus II software can perform various netlist optimizations during the Fitter
stage to meet performance and area requirements.
f
For more details about optimization techniques available in Quartus II for the
place-and-route stage of design compilation, refer to the Area and Timing Optimization
chapter in volume 2 of the Quartus II Handbook.
The Chip Planner in the Quartus II software allows you to assign logic to a specific
location or range of locations within a device or logic cell (including embedded and
I/O cells). The Chip Planner also lets you view logic placement made by the Fitter
and/or by user assignments, make and view LogicLock region assignments, and view
critical path information, physical timing estimates, and routing congestion.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 22
Post Place-and-Route Verification
Post Place-and-Route Verification
Typical design verification tasks after place-and-route include static timing analysis,
logic equivalence checking, and power estimation. The Quartus II software supports
various third-party EDA tools to perform simulation and formal verification tasks.
Gate-Level Simulation
The Quartus II software can generate a netlist to run simulation with industry
standard simulators and has the option of generating either of the following netlists:
f
■
Verilog HDL or VHDL netlist to perform functional simulation—You can generate
a functional simulation netlist that does not contain any timing information, and
use this netlist to simulate your design on any of the supported industry standard
simulators.
■
Verilog HDL or VHDL netlist with timing information to perform timing
simulation—You can generate a netlist with all the timing information included in
the Standard Delay File (.sdf). With this netlist, you can run a gate-level timing
simulation of your design using any of the supported industry standard
simulators.
For additional and detailed information about performing gate-level simulation with
different industry standard simulators, refer to the Simulation section in volume 3 of
the Quartus II Handbook.
Static Timing Analysis
Altera provides the TimeQuest Timing Analyzer, an ASIC strength SDC-based timing
analysis tool that is part of the Quartus II software. In addition to the TimeQuest
Timing Analyzer, you can also use PrimeTime from Synopsys, which is the
industry-standard STA sign-off tool used for standard cell ASICs. If you select the
PrimeTime EDA tool for timing analysis, the Quartus II software writes the Verilog
HDL or VHDL gate-level netlist (.sdf) and the Tcl script to run static timing analysis in
PrimeTime. In addition, PrimeTime libraries are shipped with the Quartus II software.
f
For more details about using Synopsys PrimeTime for your FPGA designs, refer to the
Synopsys PrimeTime Support chapter in volume 3 of the Quartus II Handbook.
f
For additional information about static timing analysis with TimeQuest, refer to the
Quartus II TimeQuest Timing Analyzer and Switching to the Quartus II TimeQuest Timing
Analyzer chapters in volume 3 of the Quartus II Handbook.
Formal Verification
Formal verification is widely used to verify traditional standard cell ASIC designs. By
running formal verification in conjunction with static timing analysis, you can
confirm that the post-route netlist is the same as the RTL design in functionality. This
reduces the need to run resource intensive and time-consuming gate-level
simulations.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Post Place-and-Route Verification
Page 23
Altera design flow supports industry-standard Cadence Conformal LEC for formal
verification. If you use Quartus II integrated synthesis to synthesize your design, you
can verify the equivalence between the RTL and post-fit design using Conformal LEC.
If you use Synplify Pro to synthesize your design, you can verify the equivalence
between the .vqm netlist and the Quartus II post-fit netlist using Conformal LEC.
f
For additional information about formal verification support, see the Formal
Verification section in volume 3 of the Quartus II Handbook.
Power Estimation
As design and device sizes grow, power becomes a very important consideration in
system design. The Altera FPGA design flow supports several methods to estimate
power at different stages of the design. These estimations help you determine the
power consumption accurately and develop the right power budget for your system.
Use the PowerPlay EPE to calculate the estimated power consumption based on
typical conditions.
f
For more information about the PowerPlay Early Power Estimator, go to
www.altera.com/support/devices/estimator/pow-powerplay.jsp.
The PowerPlay Power Analyzer, which is typically run in a later stage in the design
cycle, uses simulation data to accurately predict the power dissipation. By choosing
simulation data that is a true representation of the design’s operation, you can predict
the power consumption fairly accurately.
The Stratix III device family of devices provides you with an architecturally
advanced, high-performance solution together with low power consumption. In these
devices, you can choose to increase performance in regions where you require it, and
decrease power consumption everywhere else. With architecturally advanced features
and Programmable Power Technology, Stratix III devices are a good alternative to
traditional standard cell ASICs.
f
For more details about running the PowerPlay Early Power Estimator and PowerPlay
Power Analysis, refer to the PowerPlay Power Analysis chapter in volume 3 of the
Quartus II Handbook. Additional information on power estimation is also available in
the Quartus II software Help.
Clock Tree Synthesis
Clock tree synthesis (CTS) is an important step in the traditional standard cell ASIC
design methodology and is performed after placement. CTS builds a clock network to
reduce the clock skew between registers in the design. CTS synthesis is performed
either by third-party EDA tools that have the capacity to perform physical synthesis,
or by the foundry tools. CTS can take a great deal of time, and also might require
several iterations before the required clock skew is achieved. Because Altera FPGAs
include pre-built, low-skew networks, you do not have to perform CTS.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 24
Scripting
Test Methodology
Testing and fault coverage is an important aspect of the traditional standard cell ASIC
development process. In an ASIC design flow, all the following must be considered
and analyzed: scan insertion, built-in self-test (BIST), signature analysis, IDDQ, and
automatic test pattern generation (ATPG). ASIC testing typically involves test vector
generation, using ATPG tools to test the device for manufacturing defects under the
“single stuck-at” model. ATPG is usually performed after completion of scan
insertion. Scan insertion is performed to improve the obtained fault coverage by
reducing the sequential testing problem to a combinational testing problem. Using
FPGA technology, you do not have to worry about device testing, because FPGA
devices are pretested at the factory. Therefore, compared to standard cell ASIC
devices, FPGA devices are much more likely to be free of manufacturing defects.
System designers use boundary scan testing to ensure pin connectivity and
functionality on a board. In traditional standard cell ASIC design flow, you have to
spend considerable manual effort to insert and simulate boundary scan logic.
Boundary scan insertion normally requires the use of third-party EDA tools.
However, boundary scan logic (JTAG tap controller) is built into all Altera FPGAs.
You do not have to use any third-party tools or do additional work to access the
boundary scan logic available on the device.
Scripting
Traditionally, many standard cell ASIC designers use scripts to run their compilation,
simulation, synthesis, and verification. Altera’s design software supports the use of
scripts. The TimeQuest Timing Analyzer uses scripts written in Tcl and allows you to
specify timing constraints in SDC format. FPGA synthesis tools such as Synplify and
Synplify Pro, and Precision Synthesis support scripts written in Tcl and SDC format.
You can generate a Tcl Script File (.tcl) from an existing project created with the
Quartus II software, or you can use Quartus II Tcl templates to create Tcl scripts. You
can run Tcl scripts or individual Tcl commands from within the Quartus II software
and other EDA software tools.
f
For more details about using Tcl Scripting with Altera’s design flow and Quartus II
software, refer to the Command-Line Scripting and Tcl Scripting chapters in volume 2 of
the Quartus II Handbook.
Quartus II Software On-Chip Debugging Features
When you design a standard cell ASIC, you might have limited options for on-chip
debugging. The number of test nodes on a standard cell ASIC depends on the spare
I/O on the chip. Also, once the chip is fabricated, you have no way to bring out any
more test points. However, when you design with an Altera FPGA, you have a
number of on-chip debugging options for system-level debugging. Based on your
requirements, you can use one of the following tools:
■
SignalTap® II Embedded Logic Analyzer
■
SignalProbe™ Incremental Routing
■
Logic Analyzer Interface
■
In-System Memory Content Editor
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Quartus II Software On-Chip Debugging Features
■
Page 25
In-System Sources and Probes Editor
SignalTap II Embedded Logic Analyzer
In traditional standard cell ASIC devices, test points and pins are used to probe
various nodes to help identify the source of a problem. With these test points/pins
and a logic analyzer, you can effectively determine the source of most problems.
However, standard cell ASICs are less flexible to troubleshoot because once a chip is
fabricated, you cannot bring out additional internal signals to the pin level. Any logic
for test modes, and assignment of specific signals to the I/Os, must be designed
beforehand.
The SignalTap II Embedded Logic Analyzer captures the data of internal nodes and
transfers the data in real time, via a download cable, to the Quartus II software. This
data transfer takes place because the SignalTap II Embedded Logic Analyzer supports
trigger positions and trigger events, including the device’s power-up sequence, in the
same way that a standard external logic analyzer does.
Because the SignalTap II Embedded Logic Analyzer uses the programming interface
to communicate with the computer, you do not require extra test pins. Acquired data
is saved to the device’s internal RAM and then streamed off-device via the JTAG
communication port. This is advantageous when BGA packages are used and access
to pins is difficult, or impossible.
SignalProbe Incremental Routing
The SignalProbe feature gives you quick access to internal design signals for
system-level debugging. You can use a logic analyzer to probe these internal signals.
The SignalProbe feature supports incremental routing, and allows you to route signals
to I/O pins for efficient signal verification without affecting the rest of the design.
When device memory is limited and there is no access to a JTAG communication port,
you can perform signal debugging and hardware verification using the SignalProbe
feature.
The Quartus II software lets you select the nodes to be routed to pre-specified
SignalProbe pins. You can then use a logic analyzer to analyze your selected internal
nodes.
Logic Analyzer Interface
Using this interface, you can connect and transmit FPGA internal signals to an
external logic analyzer. You can use this feature to connect a large number of internal
signals to a small number of output pins for debugging.
In-System Memory Content Editor
This feature provides read and write access to FPGA memories through the JTAG
interface. This makes it easy to test changes to memory contents when the device is
functioning in a system.
In-System Sources and Probes Editor
You can use the In-System Sources and Probes editor to set up register chains to drive
or sample instrumented nodes in your design. You can input virtual stimuli and
capture the value of instrumented nodes. Thus, you can create simple test vectors to
exercise your design without external test equipment.
© April 2009
Altera Corporation
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Page 26
Device Programming
f
For more information about the system-level debug options for Altera FPGAs, refer to
the In-System Design Debugging section in volume 3 of the Quartus II Handbook.
Device Programming
After a standard cell ASIC returns from the foundry, you cannot change the
functionality of the device. However, FPGA devices can be reconfigured many times,
providing more flexibility. The ability to change the functionality of a device and add
enhancements and fixes is a very useful feature during the prototyping stages.
Altera FPGAs can be configured using a download cable or by using a configuration
device. A download cable is used more frequently during prototyping stages. Using a
download cable makes design changes very simple and fast, but the board must be
reconfigured using the Quartus II software every time the power is applied to the
board. There are various download cables you can use to connect to the serial, USB, or
parallel port of your computer, and each download cable supports different
configuration methods.
In systems where it is not practical to configure a device using a download cable,
Altera FPGAs can be configured using flash-based or EEPROM-based configuration
devices that store the configuration data and configure the FPGA. When the board is
powered on, the configuration device sends the configuration data from memory to
the FPGA device.
f
For more information about configuration using the download cable, refer to
volume 1 of the Configuration Handbook and the appropriate device handbook.
f
For a comprehensive list of configuration options available for different device
families, refer to the Configuration Center.
HardCopy ASIC Devices
Altera’s low power, cost-effective HardCopy ASICs are supported by Stratix series
seamless prototyping—delivering the benefits of FPGAs in design and the benefits of
ASICs in production. Stratix series FPGAs enable software and hardware co-design,
which cuts 9 to 12 months from your design cycle. The low NRE, coupled with the fast
and predictable turnaround time for HardCopy ASICs, delivers your products to
market faster than any other methodology.
Altera’s Quartus II software allows you to create one design, with one RTL and one
suite of IP, while delivering two implementations: the FPGA prototype and the
HardCopy ASIC. This dramatically reduces total cost and risk in your ASIC projects.
Using HardCopy ASICs, Altera’s SOPC solutions can be leveraged from prototype to
production, while reducing cost and speeding time-to-market. Unlike traditional
standard cell ASIC device development, the HardCopy ASIC design does not require
the generation of testbenches, test vectors, or timing and functional simulation. The
HardCopy conversion process requires only the Quartus II software-generated output
files. Altera HardCopy Design Center performs the conversion and delivers
functional prototypes in eight weeks.
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
© April 2009 Altera Corporation
Conclusion
Page 27
f
For more information about HardCopy ASICs, and for details about migrating Altera
FPGA designs to HardCopy ASICs, refer to the HardCopy device handbooks at
www.altera.com.
Conclusion
There are several advantages to using Altera FPGAs or HardCopy ASICs in place of
traditional standard cell ASICs in your systems. You can reduce the time spent on
prototyping and have more time for volume production.You can start the firmware
development early in the design cycle because you can program the FPGA even while
your design is not yet complete and bug free. Modifications, to either the firmware or
the design, can be accomplished with a short turnaround time. Because devices can be
reconfigured, there is no NRE cost associated with design iterations, and design
modifications are easy to accommodate—taking minutes as compared to months.
The Quartus II software performs place-and-route and generates all the files
necessary to build the FPGA. The Quartus II software provides an easy interface with
industry-standard simulation and timing verification EDA tools. Altera also offers the
migration from FPGAs to cost effective HardCopy ASIC for high volume applications.
f
For more information about Altera products, refer to the Literature and Technical
Documentation page on the Altera website.
Document Revision History
Table 1 shows the revision history for this application note.
Table 1. Document Revision History
Date and Version
Changes Made
Change Summary
April 2009
■
Updated “IP Availability and Flow” on page 20
v3.1
■
Updated “Functional Simulation” on page 21
■
Updated “Gate-Level Simulation” on page 22
■
Updated “Static Timing Analysis” on page 22
■
Updated “Post Place-and-Route Verification” on page 22
■
Updated to follow the new naming convention for
HardCopy ASICs
■
Removed FPGA Financial Benefit section
June 2008
v3.0
© April 2009
Altera Corporation
Restructured sections for better
clarity.
—
AN 311: Standard Cell ASIC to FPGA Design Methodology and Guidelines
Document Revision History
Table 1. Document Revision History
Date and Version
October 2007
v2.0
July 2003
v1.0
101 Innovation Drive
San Jose, CA 95134
www.altera.com
Technical Support
www.altera.com/support
Changes Made
Updated sections:
■
FPGA Financial Benefit
■
Design Specifications
■
Design Development
■
Design Development Tools
■
Gate-Level Simulation and Timing Analysis
■
Post Place-and-Route Verification
■
Quartus II Software On-Chip Debug Features
■
Device Programming (Prototype)
■
HardCopy Devices
■
Added Referenced Documents
Change Summary
Updated for the Quartus II software
version 7.2.
Initial release.
Copyright © 2009 Altera Corporation. All rights reserved. Altera, The Programmable Solutions Company, the stylized
Altera logo, specific device designations, and all other words and logos that are identified as trademarks and/or service
marks are, unless noted otherwise, the trademarks and service marks of Altera Corporation in the U.S. and other
countries. All other product or service names are the property of their respective holders. Altera products are protected
under numerous U.S. and foreign patents and pending applications, maskwork rights, and copyrights. Altera warrants
performance of its semiconductor products to current specifications in accordance with Altera's standard warranty,
but reserves the right to make changes to any products and services at any time without notice. Altera assumes no
responsibility or liability arising out of the application or use of any information, product, or service
described herein except as expressly agreed to in writing by Altera Corporation. Altera customers are
advised to obtain the latest version of device specifications before relying on any published
information and before placing orders for products or services.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement