Intel details FPGA roadmap

Seven years after its $16.7 billion acquisition of FPGA maker Altera, Intel is expanding the acquired technology into new areas.

While the primary use of an FPGA processor has been for smartNICs that offload tasks from server processors, Intel is now looking to expand its application from the data center to remote, edge computing, and embedded systems.

It’s not like Altera processors have languished for the past few years, though. A major change is manufacturing. When Intel bought Altera, its chips were made by TSMC. Now they’re made by Intel, so hopefully that’s one less headache for the supply chain.

“Supply has just become critical,” said Patrick Dorsey, vice president of product marketing, FPGAs and power products in Intel’s Programmable Solutions Group. “The number one feature on the market is ‘have you got my share?’ So having the supply chain is going to become critical.

When it comes to chips, Intel has outlined a roadmap that breaks down new and future FPGAs into four distinct product categories, all under the Agilex brand.

Agilex D-Series

The new Agilex D-Series FPGAs join legacy brands Cyclone, Arria, and Stratix in the midrange. The D-series will be physically smaller overall, with lower thermal requirements and a lower price.

Dorsey says the D-Series will be sold for use in areas such as 5G baseband radios, in-vehicle applications such as the medical device industry and manufacturing. The D-series will also find its way into the data center through applications such as storage acceleration, where the FPGA would handle data movement instead of the CPU, according to Dorsey.

Multi-chip FPGA packages

The second product line consists of Agilex FPGAs supporting Compute Express Link (CXL) and PCIe 5. Designed for demanding processor workloads, they will also find their way into data centers for network processing. This family of FPGAs will include a multi-chip package – you will see the FPGA combined with an ASIC, a CXL and PCIe 5 chip, all in a single package linked by a high-speed interconnect.

This will be part of Intel’s smartNIC, or as Intel calls it, the Infrastructure Processing Unit (IPU) strategy.

One of the reasons for the multi-chip package strategy is that it allows Intel to upgrade bits of the chip without requiring a complete overhaul. The first Agilex FPGAs this year will come with CXL 1.2 support, for example. But next year they will be upgraded to CXL 2.0. With this design, Intel can simply replace the CXL component in the package without having to redesign and refactor the entire chip.

Sundance Mesa

Another part of Intel’s data center strategy focuses on artificial intelligence and support for AI applications both within the network and at the edge. The Sundance Mesa line of new FPGAs will come with advanced AI capabilities related to INT8 processing, which is used for the inference part of the AI.

“You’re not going to be training on an FPGA, but you can do very fast inferences and you can innovate very quickly with this programmable fabric,” Dorsey said.

Intel Direct RF Wallet

The final product line is the Intel Direct RF portfolio, which are dedicated processors for analog and RF signaling applications.

Another “big focus” for Intel is developers, according to Dorsey. It’s notoriously difficult to write applications for FPGAs, he says. Intel therefore has more than 35% of its engineering team creating software to ease the load of writing core code from developers. Intel also makes compilers and other development tools available, and includes FPGAs in its oneAPI strategy, which Intel designed to simplify programming by creating an API that supports a variety of processors, from CPUs to GPUs. by the FPGA and more.