Advantages of AMD 7 Series FPGA and SoC Devices​

​​Xilinx, now part of AMD, expanded the definition of FPGAs at the 28 nm node and delivered not only the industry’s most advanced FPGAs but also a game-changing line of SoC and 3D ICs. NI played a key role in helping define the requirements for AMD 7 series devices and was a lead partner in the SoC program.

Rather than using the high-performance/high-power process tailored for graphics chipsets in PCs or the low-power process targeted to mobile phone devices, AMD worked with TSMC to develop a process tailored to the requirements of these devices. Thus, the TSMC 28 nm high performance low power (HPL) process technology includes the right mix of performance and power savings for all the markets AMD devices serve. Designs that need to run at high performance do not have to come at an extreme cost of higher power consumption. Conversely, designs that need to meet strict power requirements can still achieve relatively high-performance goals. With the entire 7 series portfolio targeting the same 28 nm HPL silicon process, AMD concentrated on architecture innovation. For the 7 series, AMD introduced a full line of scalable FPGAs, which includes the low-cost Artix™ 7 family, the midrange Kintex™ 7 family, and the high-end Virtex™ 7 family. The base FPGA building blocks of logic cells, DSP blocks, Block RAM, and so on are all consistent across the 7 series, making it much simpler to migrate designs.​

​​AMD invented the FPGA in 1988 and has delivered state-of-the-art FPGA technology ever since. The Kintex 7 device family features a balance of FPGA fabric clock rate performance versus power consumption, high-speed I/O, capacity, security, and reliability.

​This highly balanced feature set makes the device well-suited to serve a broad number of test and high-performance embedded applications, such as custom triggering, hardware-timed test sequencing, medical imaging, big physics control and monitoring of ultra-wide bandwidth communications and radar, signal intelligence, protocol-aware digital test, real-time vision algorithms, and software defined radio.

​The Kintex 7 FPGAs are ideal for many products that use the LabVIEW reconfigurable I/O (RIO) architecture. This device family has the capacity and performance of previous high-end FPGAs, but at half the power consumption. The impressive reduction in power means devices can pack more than twice the digital signal processing capability per device as the previous generations of FPGA-enabled NI hardware. With this increase in logic and DSP resources, engineers can implement more complex algorithms with more extensive use of signal processing and real-time analysis, and address the growing I/O data rates and complexity of today’s applications.

​In addition to more DSP resources, the DDR3 memory controller helped implement a 10 GB/s (theoretical) interface to temporary storage compared to the previous generation’s rates of 3.2 GB/s. Furthermore, the built-in PCI Express controller bandwidth increased from 800 MB/s to 1,600 MB/s, enabling even faster data transfers from the FPGA to the host. 

​NI hardware that capitalizes on Kintex 7 advantages include reconfigurable multifunction and digital I/O, NI FlexRIO modules, CompactRIO Controller, and Universal Software Radio Peripheral (USRP) software defined radio devices.

​​Artix 7 uses the same FPGA resources as Kintex 7, but it is optimized for even lower power consumption and smaller size packages, delivering similar advantages at a value-line price point. Migration between these devices is straightforward because their FPGA building blocks are functionally identical to the Kintex 7. This compatibility is particularly helpful when using the LabVIEW FPGA Module. ​

​​In LabVIEW 2014 and later, developers using AMD 7 series FPGAs in their FPGA-programmable hardware can also benefit from Vivado, the latest compilation technology from AMD. Vivado compilation tools offer numerous benefits, which include:

• ​More consistent, reliable timing closures
• ​Improved resource utilization
• ​Faster compiles for Kintex 7 FPGA targets previously using AMD ISE (NI cRIO-9068 and NI PXIe-7975R) 

^{Figure 1: Vivado compilation technology from AMD offers reduced compilation times for Kintex 7 previously using AMD ISE.}

Explore NI targets and devices that use AMD 7 series FPGAs and SoCs:

• CompactRIO
• Single-Board RIO
• FlexRIO
• PXI platform
• USRP Software Defined Radio Devices

Learn about the LabVIEW RIO architecture and LabVIEW FPGA: LabVIEW FPGA