At the Gigaom Structure conference in San Francisco today, Intel's Diane Bryant announced that the company will now offer customizable chips for the data center based on Intel Xeon processors. For the first time, Intel is integrating its Xeon processor with a coherent field-programmable gate array (FPGA) in a single package, socket compatible to the standard Xeon E5 processor offerings. The company's customers will be able to incorporate their own software and IP via a FPGA. The Xeon+FPGA solution can be programmed for specific functions and workloads, like search or video compression, particularly as the patterns change.

There are two ways customers can use the FPGAs, Bryant said. In one case, the online service provider deploys a Xeon package with the FPGA and tests a workload to ensure it delivers the desired benefits. If it does, Intel will burn the instructions onto the Xeon itself and manufacture the part without the FPGA.

The other use case takes advantage of the fact that FPGAs can be reprogrammed in the field. The service provider buys servers with the dual-chip package inside, and programs the FPGA depending on the workload they need to optimize. If their needs change later, they can reprogram the chip again.

Based on industry benchmarks FPGA-based accelerators can deliver >10X performance gains. By integrating the FPGA with the Xeon processor, Intel estimates that its customers will see an additional 2X in performance thanks to the low latency, coherent interface.

Intel plans to begin production of the Xeon-FPGA chip packages soon, Bryant said.

Intel has built custom chips for customers before. Last year it delivered about 15 unique designs, including processors for Facebook and eBay. But they involved relatively minor changes, such as disabling cores and adding extra I/O ports.