or product names are trademarks or registered trademarks of their respective holders. The specifications and information herein are subject to change without notice.
www.latticesemi.com
1
gdx2fam_13
Lattice Semiconductor
Figure 1. ispGDX2 Block Diagram (256-I/O Device)
sysIO Bank
sysHSI
Block
ispGDX2 Family Data Sheet
sysIO Bank
SERDES
FIFO
GDX Block
SERDES
FIFO
sysHSI
Block
sysHSI
Block
sysCLOCK
PLL
SERDES
FIFO
SERDES
FIFO
GDX Block
sysCLOCK
PLL
sysHSI
Block
GDX Block
GDX Block
GDX Block
GDX Block
SERDES
SERDES
SERDES
SERDES
SERDES
sysHSI
Block
FIFO
FIFO
sysIO Bank
Introduction
The ispGDX2™ family is Lattice’s second generation in-system programmable generic digital crosspoint switch for
high speed bus switching and interface applications.
The ispGDX2 family is available in two options. The standard device supports sysHSI capability for ultra fast serial
communications while the lower-cost “E-series” supports the same high-performance FPGA fabric without the
sysHSI Block.
This family of switches combines a flexible switching architecture with advanced sysIO interfaces including high
performance sysHSI Blocks, and sysCLOCK PLLs to meet the needs of the today’s high-speed systems. Through
a muliplexer-intensive architecture, the ispGDX2 facilitates a variety of common switching functions.
The availability of on-chip control logic further enhances the power of these devices. A high-performance solution,
the family supports bandwidth up to 38Gbps.
Every device in the family has a number of PLLs to provide the system designer with the ability to generate multiple
clocks and manage clock skews in their systems.
sysIO Bank
sysIO Bank
GDX Block
GDX Block
SERDES
sysCLOCK
PLL
FIFO
FIFO
FIFO
FIFO
Global Routing Pool
(GRP)
sysHSI
Block
GDX Block
GDX Block
GDX Block
FIFO
SERDES
GDX Block
GDX Block
SERDES
SERDES
sysHSI
Block
FIFO
FIFO
sysIO Bank
GDX Block
GDX Block
GDX Block
FIFO
SERDES
FIFO
SERDES
FIFO
SERDES
sysHSI
Block
sysCLOCK
PLL
sysIO Bank
sysIO Bank
ISP & Boundary Scan
Test Port
2
Lattice Semiconductor
ispGDX2 Family Data Sheet
The sysIO interfaces provide system-level performance and integration. These I/Os support various modes of
LVCMOS/LVTTL and support popular high-speed standard interfaces such as GTL+, PCI-X, HSTL, SSTL, LVDS
and Bus-LVDS. The sysHSI Blocks further extend this capability by providing high speed serial data transfer capa-
bility.
Devices in the family can operate at 3.3V, 2.5V or 1.8V core voltages and can be programmed in-system via an
IEEE 1149.1 interface that is compliant with the IEEE 1532 standard. Voltages required for the I/O buffers are inde-
pendent of the core voltage supply. This further enhances the flexibility of this family in system designs.
Typical applications for the ispGDX2 include multi-port multi-processor interfaces, wide data and address bus mul-
tiplexing, programmable control signal routing and programmable bus interfaces. Table 1 shows the members of
the ispGDX2 family and their key features.
Architecture
The ispGDX2 devices consist of GDX Blocks interconnected by a Global Routing Pool (GRP). Signals interface
with the external system via sysIO banks. In addition, each GDX Block is associated with a FIFO and a sysHSI
Block to facilitate the transfer of data on- and off-chip. Figure 1 shows the ispGDX2 block diagram. Each GDX
Block can be individually configured in one of four modes:
• Basic (No FIFO or SERDES)
• FIFO Only
• SERDES Only
• SERDES and FIFO
Each sysIO bank has its own I/O power supply and reference voltage. Designers can use any output standard
within a bank that is compatible with the power supply. Any input standard may be used, providing it is compatible
with the reference voltage. The banks are independent.
Global Routing Pool (GRP)
The ispGDX2 architecture is organized into GDX Blocks, which are connected via a Global Routing Pool. The inno-
vative GRP is optimized for routability, flexibility and speed. All the signals enter via the GDX Block. The block sup-
plies these either directly or in registered form to the GRP. The GRP routes the signals to different blocks, and
provides separate data and control routing. The data path is optimized to achieve faster speed and routing flexibility
for nibble oriented signals. The control routing is optimized to provide high-speed bit oriented routing of control sig-
nals.
There are some restrictions on the allocation of pins for optimal bus routing. These restrictions are considered by
the software in the allocation of pins.
GDX Block
The blocks are organized in a “block” (nibble) manner, with each GDX Block providing data flow and control logic
for 16 I/O buffers. The data flow is organized as four nibbles, each nibble containing four Multiplexer Register
Blocks (MRBs). Data for the MRBs is provided from 64 lines from the GRP. Figure 2 illustrates the groups of signals
going into and out of a GDX Block.
Control signals for the MRBs are provided from the Control Array. The Control Array receives the 32 signals from
the GRP and generates 16 control signals: eight MUX Select, four Clock/Clock Enable, two Set/Reset and two Out-
put Enable. Each nibble is controlled via two MUX select signals. The remaining control signals go to all the MRBs.
Besides the control signals from the Control Array, the following global signals are available to the MRBs in each
GDX Block: four Clock/Clock Enable, one reset/preset, one power-on reset, two of four MUX select (two of two in
64 I/O), four Output Enable (two in 64 I/O) and Test Out Enable (TOE).
3
Lattice Semiconductor
MUX and Register Block (MRB)
ispGDX2 Family Data Sheet
Every MRB Block has a 4:1 MUX (I/O MUX) and a set of three registers which are connected to the I/O buffers,
FIFO and sysHSI Blocks. Multiple MRBs can be combined to form large multiplexers as described below. Figure 3
shows the structure of the MRB.
Each of the three registers in the MRB can be configured as edge-triggered D-type flip-flop or as a level sensitive
latch. One register operates on the input data, the other output data and the last register synchronizes the output
enable function. The input and output data signals can bypass each of their registers. The polarity of the data out
and output enable signals can be selected.
The Output and OE register share the same clock and clock enable signals. The Input register has a separate clock
and clock enable. The initialization signals of each register can be independently configured as Set or Reset. These
registers have programmable polarity control for Clock, Clock Enable and Set/Reset. The output enable register
input can be set either by one of the two output enables generated locally from the Control Array or from one of the
four (two in 64 I/O) Global OE enable pins. In addition to the local clock and clock enable signals, each MRB has
access to Global Clock, Clock Enable, Reset and TOE nets.
4
Lattice Semiconductor
Figure 2. GDX Block
GRP
32 bits
MUX
Control Select
8
8
ispGDX2 Family Data Sheet
GDX Block
sysIO Bank
Control Array
Nibble 0
OE
8
2
4 bits
MUX and Register
Block (MRB)
0
IN
OUT
8
2
OE
MUX and Register
Block (MRB)
1
IN
OUT
4 bits
8
2
4 bits
MUX and Register
Block (MRB)
2
OE
IN
OUT
8
OE
2
4 bits
MUX and Register
Block (MRB)
3
IN
OUT
8
2
16 bits
4
8
2
Nibble 1
MRBs 4-7
OE
IN
OUT
OE
IN
OUT
OE
IN
OUT
Nibble 2
MRBs 8-11
16 bits
4
8
2
16 bits
4
Nibble 3
MRBs 12-15
The output register of the MRB has a built-in bi-directional shift register capability. Each output register correspond-
ing to MRB “n”, receives data output from its two adjacent MRBs, MRB (n-1) and MRB (n+1), to provide shift regis-
ter capability. Like the output register, each input register of the MRB has built-in shift register capability. Each input
register can receive data from its two adjacent MRB input registers, to provide bi-directional shift register capability.
The chaining crosses GDX Block boundaries. The chain of input registers and the chain of output registers can be
First, what is the status of the IO port? After I configured PxDIR, is PxOUT all 1 or all 0, or is it uncertain? This changes in the register from time to time. I don't understand. Please give me some...
Notes on program initialization process
1. CH395 is burned with a globally unique Mac address when it leaves the factory. If the CH395 Mac address is modified through a function, it cannot be retrieve...
[font=宋体][backcolor=white][color=#000][size=14px]When I adjusted the STM32F105 yesterday, the serial port always had garbled characters, and there was no response when changing the corresponding baud ...
The problem is this: In order to avoid wasting memory, only a relatively small memory is reserved in config.bib for framebuffer use. When the application requests a buffer, it first gets it from the f...
In the consumer, medical, automotive, and even industrial sectors, more and more electronic products utilize high-speed signaling for data and voice communications, audio, and imaging applications. Al...
"Have you set your calendar reminder?"
On August 24, Nvidia Robotics' official account posted a photo of a black gift box on a social media platform, with an attached greeting card sig...[Details]
As the number of cars increases, environmental pressures are also increasing. At this time, some people are saying that new energy vehicles are energy-efficient and environmentally friendly, and ar...[Details]
On August 21st, Zhiyuan Robotics revealed at its first partner conference that it expects shipments to reach thousands of units this year and tens of thousands next year. The company hopes to reach...[Details]
As AI accelerates across industries, the demand for data center infrastructure is also growing rapidly.
Keysight Technologies, in collaboration with Heavy Reading, released the "Beyo...[Details]
Electric vehicles are now widespread, but they've brought with them a host of problems, the most prominent of which is charging. Small electric vehicles (EVs) are a new form of transportation in a ...[Details]
Laird Thermal Systems has introduced the HiTemp ET series Peltier cooler modules, which can operate at high temperatures and provide on-site cooling for sensitive electronic devices.
Dig...[Details]
Learned the following information.
Customer product: industrial computer motherboard
Glue application area: CPU/BGA filling
Glue color requirements: black or t...[Details]
Compared to cloud databases, minicomputers are purpose-built for decentralized, rugged computing at the edge of the network. By moving applications, analytics, and processing services closer to the...[Details]
In the field of intelligent driving, regulations are becoming increasingly stringent, and the technical threshold continues to rise. Especially after the traffic accident in March 2025, the Ministr...[Details]
In daily life, when we purchase a transformer, we are faced with the installation and wiring procedures. Generally speaking, large transformers such as power transformers are equipped with speciali...[Details]
Inverter power supplies on the market can generally be divided into two categories: sine wave inverters and square wave inverters. Some engineers also like to categorize pure sine wave inverters as...[Details]
For new energy vehicles, the importance of batteries is unquestionable. Not only does it determine the performance of the vehicle, but the battery density also has a great relationship with the veh...[Details]
Renesas Electronics' new ultra-low-power RA4C1 MCU features advanced security and a dedicated peripheral set, making it ideal for metering and other applications.
The new product mee...[Details]
introduction
A common voltage regulator is a three-terminal one. Its function is to step down the voltage and stabilize it at a fixed output value. Voltage regulators are commonly available in...[Details]
The Radxa Cubie A7A is a single-board computer (SBC) powered by the Allwinner A733 octa-core Cortex-A76/A55 SoC, equipped with a 3 TOPS AI accelerator and up to 16GB of LPDDR5 memory.
Fo...[Details]
Embedded System
京公网安备 11010802033920号
EEWORLD
