R
Virtex™-II Platform FPGAs: Detailed Description
capability only available in the Xilinx design flow is “Modular
Design”, part of the Xilinx suite of team design tools, which
enables autonomous design, implementation, and verifica-
tion of design modules.
Design Implementation
The ISE Series development systems include Xilinx tim-
ing-driven implementation tools, frequently called “place
and route” or “fitting” software. This robust suite of tools
enables the creation of an intuitive, flexible, tightly inte-
grated design flow that efficiently bridges “logical” and
“physical” design domains. This simplifies the task of defin-
ing a design, including its behavior, timing requirements,
and optional layout (or floorplanning), as well as simplifying
the task of analyzing reports generated during the imple-
mentation process.
Incremental Synthesis
Xilinx unique hierarchical floorplanning capabilities enable
designers to create a programmable logic design by isolating
design changes within one hierarchical “logic block”, and
perform synthesis, verification and implementation pro-
cesses on that specific logic block. By preserving the logic in
unchanged portions of a design, Xilinx incremental design
makes the high-density design process more efficient.
The Virtex-II implementation process is comprised of Syn-
thesis, translation, mapping, place and route, and configu-
ration file generation. While the tools can be run individually,
many designers choose to run the entire implementation
process with the click of a button. To assist those who prefer
to script their design flows, Xilinx provides Xflow, an auto-
mated single command line process.
Xilinx hierarchical floorplanning capabilities can be speci-
fied using the high-level floorplanner or a preferred RTL
floorplanner (see the Xilinx web site for a list of supported
EDA partners). When used in conjunction with one of the
EDA partners’ floorplanners, higher performance results
can be achieved, as many synthesis tools use this more
predictable detailed physical implementation information to
establish more aggressive and accurate timing estimates
when performing their logic optimizations.
Design Verification
In addition to conventional design verification using static
timing analysis or simulation techniques, Xilinx offers pow-
erful in-circuit debugging techniques using ChipScope ILA
(Integrated Logic Analysis). The reconfigurable nature of
Xilinx FPGAs means that designs can be verified in real
time without the need for extensive sets of software simula-
tion vectors.
Modular Design
Xilinx innovative modular design capabilities take the incre-
mental design process one step further by enabling the
designer to delegate responsibility for completing the
design, synthesis, verification, and implementation of a hier-
archical “logic block” to an arbitrary number of designers -
assigning a specific region within the target FPGA for exclu-
sive use by each of the team members.
For simulation, the system extracts post-layout timing infor-
mation from the design database, and back-annotates this
information into the netlist for use by the simulator. The back
annotation features a variety of patented Xilinx techniques,
resulting in the industry’s most powerful simulation flows.
Alternatively, timing-critical portions of a design can be ver-
ified using the Xilinx static timing analyzer or a third party
static timing analysis tool like Synopsys Prime Time™, by
exporting timing data in the STAMP data format.
This team design capability enables an autonomous
approach to design modules, changing the hand-off point to
the lead designer or integrator from “my module works in
simulation” to “my module works in the FPGA”. This unique
design methodology also leverages the Xilinx hierarchical
floorplanning capabilities and enables the Xilinx (or EDA
partner) floorplanner to manage the efficient implementa-
tion of very high-density FPGAs.
For in-circuit debugging, ChipScope ILA enables designers
to analyze the real-time behavior of a device while operating
at full system speeds. Logic analysis commands and cap-
tured data are transferred between the ChipScope software
and ILA cores within the Virtex-II FPGA, using industry
standard JTAG protocols. These JTAG transactions are
driven over an optional download cable (MultiLINX or
JTAG), connecting the Virtex device in the target system to
a PC or workstation.
Configuration
Virtex-II devices are configured by loading application spe-
cific configuration data into the internal configuration mem-
ory. Configuration is carried out using a subset of the device
pins, some of which are dedicated, while others can be
re-used as general purpose inputs and outputs once config-
uration is complete.
ChipScope ILA was designed to look and feel like a logic
analyzer, making it easy to begin debugging a design imme-
diately. Modifications to the desired logic analysis can be
downloaded directly into the system in a matter of minutes.
Depending on the system design, several configuration
modes are supported, selectable via mode pins. The mode
pins M2, M1 and M0 are dedicated pins. An additional pin,
HSWAP_EN is used in conjunction with the mode pins to
select whether user I/O pins have pull-ups during configura-
tion. By default, HSWAP_EN is tied High (internal pull-up)
which shuts off the pull-ups on the user I/O pins during con-
figuration. When HSWAP_EN is tied Low, user I/Os have
Other Unique Features of Virtex-II Design Flow
Xilinx design flows feature a number of unique capabilities.
Among these are efficient incremental HDL design flows; a
robust capability that is enabled by Xilinx exclusive hierar-
chical floorplanning capabilities. Another powerful design
DS031-2 (v3.0) August 1, 2003
Product Specification
www.xilinx.com
1-800-255-7778
Module 2 of 4
36
XILINX [ XILINX, INC ]
