Intel Corp. has spent a lot of effort pushing the concept of its Field Programmable Gate Arrays, which can be used to accelerate various computing tasks such as artificial intelligence and machine learning.

FPGAs are a special kind of microprocessor used in data centers that can be reprogrammed on the fly and in real-time, in order to optimize them for specific tasks. The problem is that although they can be incredibly useful, they’re very difficult to reprogram in the first place, because they require expertise in little-known programming languages such as Verilog or VHDL.

To make its FPGAs more accessible and speed up adoption, Intel released a set of new software tools earlier this week that are designed to make them easier to program.

The idea is to make FPGAs easier for mainstream developers to use, in order to increase their adoption in the data center for workloads such as high-performance computing, artificial intelligence, data and video analytics, and 5G network processing.

The new tools include an Acceleration Stack for Intel Xeon CPUs with FPGAs. Intel said this enables code to be reused, while providing a common developer interface across all of its FPGA products. It works by abstracting hardware-specific FPGA resource details from the application developer, while system-optimized reference libraries are created for the target industries. Put more simply, it allows developers to program their FPGAs using library functions and flows they’re already familiar with.

Intel’s second new tool is something called Open Programmable Acceleration Engine technology, which is open code that’s packaged with the new interface, between the FPGA processor and the accelerator. Intel said this code provides a lightweight and consistent application programming interface across all FPGA accelerate generations and platforms.

Not least, Intel touted a new FPGA Software Development Kit for the OpenCL framework. OpenCL is used for writing programs that execute across heterogeneous platforms that are made up of central processing units, graphics processing units, field-programmable gate arrays and other processors or hardware accelerators. The new SDK enables developers to create custom accelerator functions that can run on Intel’s FPGAs, the company said.

What Intel is trying to do here is bring FPGA programming inside its more familiar Xeon frameworks, in order to reduce the learning curve for developers, one analyst said.

“Typically there is an inverse relationship between acceleration and programmability,” said Patrick Moorhead, president and principal analyst at Moor Insights and Strategy. “FPGAs are very hard to program, so Intel is trying to make that easier between generations and different implementations of their FPGAs. These improvements could increase the penetration of FPGAs and in concert increase their sales.”

Images: Intel