Vitis™ Unified Software Platform
Vitis USP is a free development platform created from the need to unify the workflow on all Xilinx devices, regardless of their size or topology, providing a level of abstraction suitable for any type of application developers - even if they have no previous experience with FPGA design. With the help of Vitis USP, application-level software developers can easily migrate particular processing elements into the programmable fabric and move their algorithms across different Xilinx platforms - from Edge to Cloud - without having to change a single line of code.
Layered Topology
Vitis USP features a layered topology, where each layer (component) represents a different level of abstraction for the underlying complex functionalities, specific to FPGA design.
Vitis USP - block diagram
Vitis Target Platform defines hardware and software components in the context of application deployment and development, including external memory interfaces, custom I/O interfaces, and software runtime. The developers are free to use predefined Target Platforms for Xilinx evaluation boards or define their own using Vivado® Design Suite.
Xilinx Runtime Library (XRT) is the next layer in the Vitis USP topology. This is the key component that makes the user's code platform-agnostic, allowing seamless scalability and migration between different Xilinx platforms. XRT is an open-source software stack containing kernel-level functions and drivers used to dynamically interact with the FPGA-based accelerators, providing a link between the application code running on an embedded Arm or x86 Host and the underlying Target Platform. It is responsible for downloading compiled FPGA binaries onto the platform, managing data traffic between the host CPU and FPGA, controlling the execution sequence, and other management tasks.
Vitis Core Tools are the complete set of command-line and graphical developer tools, including Vitis compilers, analyzers, and debuggers. These tools are used to build, analyze performance bottlenecks, and debug accelerated algorithms. Vitis Core Tools are an integral part of the Vitis USP. Implemented on top of the XRT layer, these tools are the integral part of the Vitis Core Development Kit.
Vitis Accelerated Libraries represent the highest layers of abstraction, providing developers with a familiar environment that is very close to what they are used to working in their domains of expertise. Vitis libraries are grouped in several categories, from the most basic ones like Math, Linear Algebra, DSP, to very domain-specific ones, such as Data Compression, AI, Media Processing, and much more. Each library contains examples, benchmarks, tests, and of course - their source code (C, C++, Python…). Although Vitis accelerated libraries are open-source (Apache 2.0 license), Xilinx puts a lot of effort into ensuring best-in-class performance and advantage over fixed-function hardware competitors (GPUs/CPUs)
Vitis accelerated libraries contain functions that can be called directly from C, C++, or Python, just like any other software-based libraries. However, these libraries are used to build FPGA accelerators, also known as kernels. The entire FPGA synthesis process is automatically handled by the tools found in the Vitis Core Development Kit (Vitis compiler and linker).
Vitis AI Development Kit is a part of the Vitis AI Development Environment and contains specialized tools for working with AI/ML models, such as the AI compiler, AI quantizer, AI optimizer, AI profiler, and AI library with a set of lightweight C ++ and Python APIs.
Analogous to the Vitis Core Development Kit, these tools are implemented on top of the XRT layer, providing the same abstraction capabilities, which allow AI developers to accelerate, optimize, and fine-tune their models, choosing the level of abstraction they need.
Edge Hardware Platform
XILINX Adaptive System-on-Modules
The Kria™ K26 SOM is the fastest path to achieving whole application acceleration at the edge. Optimized for Vision AI applications, the K26 SOM combines high AI performance with the adaptability necessary to keep up with evolving algorithms and sensor requirements. With credit-card-sized form factor, and available in production-qualified and certified Commercial and Industrial grades, K26 SOMs are built for volume edge deployment.
Production SOMs for Edge Deployment
Based on the Zynq® UltraScale+™ MPSoC architecture, the K26 SOM is capable of up to 1.4TOPS AI processing and has an integrated H.264/265 video codec. With 245 I/Os, the K26 SOM can adapt to virtually any requirement – you can connect up to 15 cameras across multiple interfaces, connect to networks at up to 40Gb/s, and have access to a wide range of USB peripherals. The resulting platform is highly scalable, with many possible end applications, and expandable for evolving system requirements.
Vision AI Starter Kit
Out-of-the-Box Ready with the KV260 Vision AI Starter Kit
With both hardware and software development requirements simplified, the KV260 Vision AI Starter Kit is the fastest and easiest platform for application development with the goal of volume deployment on Kria K26 SOMs
LEARN MORE
GETTING STARTED GUIDE
Pre-Built Hardware and Software Platform
Kria App Store
Developed with software developers in mind, the K26 SOM is out-of-the-box ready and enabled by pre-built accelerated applications for common vision functions. Whatever your FPGA knowledge, Kria SOMs will speed your development and improve time-to-market.
KRIA APP STORE
Avnet System-on-Modules
Avnet offers a broad range of development kits, Systems-on-Module (SoMs), tools, and resources to help developers jump-start their designs, regardless of the complexity. Whether looking for a ready-made System on Module solution that can be used in a custom design or just exploring the rich world of options offered by Xilinx powerful MPSoCs even without prior FPGA design experience, these ready-made solutions from Avnet Integrated supported by Vitis USP can help reduce the associated design costs and cut time to market, allowing designers to focus on what really matters.
In addition to the hardware itself, Avnet Integrated provides all the necessary resources in the form of preconfigured SD card images, allowing developers to take full advantage of Vitis USP and gain a competitive edge in the highly demanding High-Speed Compute (HSC) market.
Ultra96-V2 Development Board
 |
 |
| Block diagram |
The Ultra96-V2 Development Board is the successor of the very popular Ultra96 Development Board from 2018. It is a Single-Board Computer (SBC), compliant with the Linaro 96Boards Consumer Edition (CE) specification. It employs the ZU3EG A484, a powerful Xilinx Zynq® UltraScale+™ general-purpose MPSoC, 2 GB of Micron's low-power DDR4 (LPDDR4), and a plethora of connectivity options, both wired and wireless, including world-wide certified Microchip Wi-Fi/Bluetooth® 5 Low Energy, USB 3.0 upstream and downstream ports, mDP display port, microSD card reader, and much more. It also contains two expansion headers (40-pin low-speed and 60-pin high-speed), enabling additional connectivity and functionality over a broad range of expansion boards and shields.
Combined with the powerful Vitis USP, the Ultra96-V2 Development Board represents an entry-level Zynq UltraScale+ MPSoC development environment, ideal for developers from many different areas of expertise. It can be used for the development and prototyping of a broad range of applications, including AI/ML, industrial automation, robotics, and much more. Supported by a rich base of extension boards and shields, it can also be used as a rapid prototyping platform for IoT-based applications, providing powerful embedded computing capabilities for demanding edge applications. The UltraZed-EV SoM is available for both extended and industrial temperature ranges.
For more details, a full list of specifications, and ordering instructions, please visit the Ultra96-V2 landing page.
UltraZed System on Module
Introduction
The UltraZed™ Systems-On-Module (SoMs) are based on the highly-acclaimed Xilinx Zynq® UltraScale+™ MPSoCs. Designed in a small form factor, the UltraZed SoMs can be used with a custom carrier card or bundled with one of the carrier cards created by Avnet, providing the complete prototyping, development, and evaluation system. The UltraZed SoMs significantly reduce time to market while saving a lot of development effort and associated costs, delivering a proven and tested basis for a complex custom solution. Powered by the Vitis USP, these ready-made solutions provide a powerful tool for the broadest range of both software and hardware developers.
UltraZed-EV™ SoM
 |
 |
| Block diagram |
The UltraZed-EV™ SoM features the powerful XCZU7EV-FBVB900, a member of the EV family of Zynq® UltraScale+™ MPSoC devices, specialized in demanding video processing applications. This compact SoM is equipped with three high-density mezzanine-style micro headers installed on the back of the module (JX1 to JX3), enabling seamless connectivity to a carrier card, and easy access to all available I/O pins, including 152 Programmable Logic I/O pins (PL IO) and 26 Processing System Multiplexed I/O pins (PS MIO), among others. UltraZed-EV SoM also includes all the necessary supporting components, including dual system memory (4 GB DDR4 for PS and 1 GB DDR4 for PL side), 8 GB eMMC flash, high-speed transceivers, GbE Ethernet and USB PHYs, on-board voltage regulator, and other components necessary for its basic functionality.
Supported by the Vitis™ USP, the UltraZed-EV SoM enables a wide range of designers to develop demanding applications with ease and confidence. An integrated H.264/H.265 video codec unit capable of simultaneously encoding and decoding 4K2K video streams at 60 Hz makes the UltraZed-EV SoM the perfect solution for a wide range of applications that rely on high-bandwidth video stream processing, such as automotive ADAS, video conferencing, augmented reality, drones, machine vision, surveillance, and other similar applications.
For more details, a full list of specifications, and ordering instructions, please visit the UltraZed-EV landing page.
UltraZed-EG™ SoM
 |
 |
| Block diagram |
The UltraZed-EG™ SoM features a highly flexible general-purpose Xilinx Zynq® UltraScale+™ MPSoC device, labeled as ZU3EG (XCZU3EG-SFVA625). This compact SoM is equipped with three high-density mezzanine-style micro headers installed on the back of the module (JX1 to JX3), enabling seamless connectivity to a carrier card and easy access to all available I/O pins, including 180 Programmable Logic I/O pins (PL IO), and 26 Processing System Multiplexed I/O pins (PS MIO), among others. UltraZed-EG SoM also includes all the necessary supporting components, including 2 GB DDR4 SDRAM, 8 GB eMMC flash, high-speed transceivers, GbE Ethernet and USB PHYs, on-board voltage regulator, and other components necessary for its basic functionality.
Supported by the Vitis™ USP, the UltraZed-EG SoM enables a wide range of designers to develop demanding applications with ease and confidence. The UltraZed-EV SoM is the perfect solution for a wide range of HS computing applications such as industrial automation, maintenance prediction, motor control, test and measurement, and other similar applications that can benefit from this highly flexible SoM solution based on the Xilinx Zynq® UltraScale+™ EG series MPSoC.
For more details, full list of specifications, and ordering instructions, please visit the UltraZed-EG™ SoM landing page.
Related parts:
- UltraZed-EV™ Carrier Card
 |
 |
| Block diagram |
The UltraZed-EV™ Carrier Card provides support and additional functionalities for the UltraZed-EV SoM, expanding its connectivity via the three high-density mezzanine-style micro headers (JX1 to JX3), which allow easy docking to the carrier card. The carrier board provides functionalities from both PL and PS side of the SoM, such as dual USB-UART interface, USB 2.0/3.0 connectors, SATA 3.0 host interface, one PCIe® Gen2 x1 Root Port, HDMI input and output interfaces, DisplayPort connector, High Pin Count (HPC) FMC interface, 3G SDI interface, LEDs, pushbuttons, DIP switches, JTAG interface, among many others.
The carrier card also provides the voltage regulator for the docked SoM, along with the PM Bus and SYSMON header, necessary clock signals for different on-board peripherals, and everything else needed to turn the UltraZED-EV SoM into a powerful and feature-rich development platform. The UltraZed-EV Carrier Card can also be used as a reference and a good starting point for creating a custom UltraZED-EV SoM-based solution.
For more details, a full list of specifications, and ordering instructions, please visit the UltraZed-EV™ Carrier Card landing page.
- UltraZed-EG™ IO Carrier Card
 |
 |
| Block diagram |
The UltraZed-EG™ IO Carrier Card provides support and additional functionalities for the UltraZed-EV SoM, expanding its connectivity via the three high-density mezzanine-style micro headers (JX1 to JX3), which allow easy docking to the carrier card. The carrier board provides functionalities from both PL and PS side of the SoM, such as dual USB-UART interface, USB 2.0/3.0 connector, SATA connector, two DisplayPort connectors, 12 x Digilent Pmod™ compatible interfaces (connected to the PL side), Arduino pin-compatible header (digitally isolated), RJ45 connector, LEDs, pushbuttons, DIP switches, touchpad interface, and JTAG interface, among many others.
The carrier card also provides the voltage regulator for the docked SoM, along with the PM Bus and SYSMON header, necessary clock signals for different on-board peripherals, and everything else needed to turn the UltraZED-EG SoM into a powerful and feature-rich development platform. The UltraZed-EG IO Carrier Card can also be used as a reference and a good starting point for creating a custom UltraZED-EG SoM-based solution.
For more details, a full list of specifications, and ordering instructions, please visit the UltraZed-EG™ IO Carrier Card landing page.
- UltraZed-EG™ PCIe Carrier Card
 |
 |
| Block diagram |
The UltraZed-EG™ PCIe Carrier Card provides support and additional functionalities for the UltraZed-EV SoM, expanding its connectivity via the three high-density mezzanine-style micro headers (JX1 to JX3), which allow easy docking to the carrier card. It also provides a PCIe® x1 Endpoint interface via the edge connector, enabling it to be plugged in the PCIe slot on the host device. Among many other connectivity options and functionalities from both the PL and PS side of the SoM, the PCIe Carrier Card also features a dual USB-UART interface, USB 2.0/3.0 connector, SATA connector, DisplayPort connector, 2 x Digilent Pmod™ compatible interfaces (connected to the PL side), Low Pin Count (LPC) FMC connector, RJ45 connector, LEDs, pushbuttons, DIP switches, touchpad interface, JTAG interface, and more.
The carrier card also provides the voltage regulator for the docked SoM, along with SYSMON and PM Bus interfaces, as well as all the necessary clock signals for different on-board peripherals, and everything else needed to turn the UltraZED-EG SoM into a powerful and feature-rich PCIe-based development platform. The UltraZed-EG IO Carrier Card can also be used as a reference and a good starting point for creating a custom UltraZED-EG SoM-based solution.
For more details, a full list of specifications, and ordering instructions, please visit the UltraZed-EG™ PCIe Carrier Card landing page.
Xilinx Alveo adaptable accelerator cards
As technology evolves at a tremendous pace, so does the demand for high-performance computing. In the everlasting pursuit for higher processing speed and faster throughput, developers shifted their focus from using a single, ultra-fast computing device such as a CPU, to multi-parallel computing platforms capable of distributing the workload across an array of application-specific processing cores. Indeed, these platforms enabled significant processing acceleration and allowed many new technologies to emerge.
However, fixed-function, instruction-based hardware is built for specific tasks and is not optimized for unstructured algorithms and data processing. With the introduction of emerging technologies, such as Artificial Intelligence (AI) and its subsets, it became apparent that the application-specific hardware is unable to adapt to rapidly evolving software algorithms. Also, with the demise of Moore's law, both power consumption and the chip size became critical factors for a wide range of applications, from cloud to edge. To maintain an optimal level of synchronization between hardware capabilities and software requirements, a significantly different approach is required.
Based on the advanced 16 nm UltraScale+™ field-programmable architecture, Xilinx Alveo accelerator cards deliver unparalleled flexibility and processing power through the deployment of domain-specific architectures, enabling tremendous acceleration for memory-bound, compute-intensive applications. The Xilinx Alveo accelerator card series can be customized for different workloads at a hardware level, thereby ensuring optimal performance and power consumption, while consequently avoiding unnecessary delays associated with the traditional fixed-function hardware implementation. Xilinx Alveo accelerator cards are the most efficient processing accelerators available to date, delivering the highest performance per watt, which makes them a perfect choice for deployment both in the cloud and in on-premises data centers.
The Xilinx Alveo accelerator card series offers an adaptive and scalable solution for a broad range of different workloads through reconfigurable logic, including machine learning inference, quantitative finance, video transcoding, database search and analytics, and more. Starting with the power-efficient, passively cooled Alveo U50 model, packaged in an efficient 75-watt, small form factor, and armed with 100 GbE networking, PCIe Gen4, and HBM2, Xilinx offers solutions for many different use cases. With the capability to quickly adapt to continuously evolving workload algorithms, Xilinx Alveo accelerator cards offer a future-proof concept, reducing the total ownership cost.
Supported by Vitis™, a unified software platform that features a comprehensive core development kit, a rich set of hardware-accelerated open-source libraries, industry-leading domain-specific development environments, and a growing ecosystem of third-party libraries and pre-built applications, Xilinx Alveo Acceleration Cards allow the widest range of developers to build their own custom hardware-accelerated applications, using familiar high-level frameworks.
Alveo U50
Xilinx Alveo U50 is an adaptable accelerator card designed for a broad range of different workloads. This accelerator card can be used for optimized workload acceleration of many demanding applications, including financial computing, machine learning, computational storage, data search and analysis, and more. Packaged in an efficient 75-watt, featuring Ultrascale+™ programmable architecture, and armed with features such as PCI Express 4.0, 8GB of HBM2, 100 GbE networking, Alveo U50 accelerator card brings power-efficient and highly scalable adaptive hardware acceleration for both in the cloud and in on-premises data centers.
Due to its low power consumption, Alveo U50 is ideally suited for deployment in any server. Alveo U50 accelerator cards are supported by the Vitis™ Unified Software Platform, allowing a broad range of engineers and developers to unlock the full potential of adaptable hardware acceleration in their applications.
For ordering information and more details on Xilinx Alveo U50 accelerator cards with passive cooling, please visit the links below:
| XILINX ALVEO U50 FEATURES |
| Dimensions |
Form factor |
½ Height
½ Length
|
| Width |
Single Slot |
| Logic resources |
Look-up Tables |
872,000 |
| Registers |
1,743,000 |
| DSP Slices |
5,952 |
| Compute Performance |
INT8 TOPs |
16.2 |
| DRAM Memory |
HBM2 Memory |
8 GB |
| HBM2 Bandwidth |
316 GB/s (peak)
201 GB/s (nominal)
|
| Internal SRAM |
Total Capacity |
28 MB |
| Total Bandwidth |
24 TB/s |
| Interfaces |
PCI Express® |
Gen3 x16
2 x Gen4 x8
CCIX
|
| Network Interface |
1 x QSFP28 (100 GbE) |
| Timestamp |
Clock Precision |
IEEE 1588 |
| Power and Thermal |
Thermal cooling |
Passive |
| Typical power |
50 W |
| Maximum Power |
75 W |
| Links |
Machine Learning Solution Brief |
| Alveo U50 Product Brief |
Alveo U200/U250
Xilinx Alveo U200/U250 are adaptable accelerator cards designed for more demanding workloads. These cards can be used for optimized workload acceleration of many different applications that require high data throughput and low processing latency, including machine learning inference, video transcoding, database search and analytics, genomics, and similar highly demanding workloads. Based on the programmable 16nm Ultrascale+™, Alveo U200/Alveo U250 accelerator cards bring adaptable and customizable hardware acceleration with up to 90x performance increase over the fastest fixed-function solutions available up to date.
Alveo U200/U250 cards feature a set of high-speed options, enabling full-scale acceleration across the entire application, reducing any performance bottlenecks. These accelerator cards are available with both active and passive cooling, depending on specific deployment needs or constraints. Alveo U200/U250 accelerator cards are supported by the Vitis™ Unified Software Platform, allowing a broad range of engineers and developers to unlock the full potential of adaptable hardware acceleration in their applications.
For ordering information and more details on Xilinx Alveo U200/U250 accelerator cards with active and passive cooling, please visit the links below:
| XILINX ALVEO FEATURES |
ALVEO U200 |
ALVEO U250 |
| Dimensions |
Form factor (passive)
Form factor (active) |
Full Height, ¾ Length
Full Height, Full Length
|
Full Height, ¾ Length
Full Height, Full Length
|
| Width |
Dual Slot |
Dual Slot |
| Logic resources |
Look-up Tables |
1,182,000 |
1,728,000 |
| Registers |
2,364,000 |
3,456,000 |
| DSP Slices |
6,840 |
12,288 |
| Compute Performance |
INT8 TOPs |
18.6 |
33.3 |
| DRAM Memory |
DDR Format |
4 x 16 GB 72b DIMM DDR4 |
4 x 16 GB 72b DIMM DDR4 |
| DDR Total Capacity |
64 GB |
64 GB
|
| DDR Total Bandwidth |
77 GB/s |
77 GB/s |
| Internal SRAM |
Total Capacity |
43 MB |
57 MB |
| Total Bandwidth |
37 TB/s |
47 TB/s |
| Interfaces |
PCI Express® |
Gen3 x16 |
Gen3 x16
|
| Network Interface |
2 x QSFP28 (100 GbE) |
2 x QSFP28 (100 GbE) |
| Power and Thermal |
Thermal cooling |
Passive
Active
|
Passive
Active
|
| Typical power |
100 W |
110 W |
| Maximum Power |
225 W |
225 W |
| Links |
Machine Learning Solution Brief |
| Alveo U200/U250 Product Brief |
Alveo U280
Xilinx Alveo U280 acceleration card is designed to meet the continuously changing needs of a modern data center. Based on the proven Ultrascale+™ programmable architecture, and armed with 8 GB of HBM2 up to 460 GB/s bandwidth, it provides high-performance, adaptable acceleration for memory-bound, compute-intensive applications such as database, analytics, and real-time machine learning inference.
Besides 8 GB of HBM2 memory, the Alveo U280 accelerator card also includes two DDR 4 slots, each supporting up to 16 GB, reaching total bandwidth of 38 GB/s. Also, it has an internal Synchronous RAM capacity of 41 MB, reaching up to 30 TB/s.
Alveo U280 is Xilinx's top tier accelerator card. It offers many additional features, including PCI Express 4.0 interface with CCIX support, allowing it to leverage the latest server interconnect infrastructure for high-bandwidth, low latency, cache-coherent shared memory access with CCIX host processors. 100 GbE is also available, enabling utilization of the hardware acceleration for network performance optimization on a large scale. Xilinx Alveo U280 can be volume-deployed in data centers and used for application development and prototyping. Alveo U280 accelerator cards are supported by the Vitis™ Unified Software Platform, allowing a broad range of engineers and developers to unlock the full potential of adaptable hardware acceleration in their applications.
For ordering information and more details on Xilinx Alveo U280 accelerator cards, please visit the links below:
| XILINX ALVEO U280 FEATURES |
| Compute |
INT8 TOPs (peak) |
24.5 |
| Dimensions |
Width |
Dual Slot |
| Form factor |
Full Height, Full Length |
| DRAM Memory |
HBM2 Total Capacity |
8 GB |
| HBM2 Total Bandwidth |
460 GB/s |
| DDR Format |
2x 16 GB 72b DIMM DDR4 |
| DDR Memory Capacity |
32 GB |
| DDR Total Bandwidth |
38 GB/s |
SRAM
Memory |
Internal SRAM Capacity |
41 MB |
| Internal SRAM Total Bandwidth |
30 TB/s |
| Interfaces |
PCI Express® |
Gen4x8 with CCIX |
| Network Interfaces |
2x QSFP28 (100GbE) |
| Logic resources |
Look-up Tables (LUTs) |
1,079,000 |
| Power |
Maximum Total Power |
225 W |
| Links |
Alveo U280 Product Brief |
