Intel is gearing up to release its Xeon processors with field-programmable gateway arrays (FPGAs) in early 2016.
For the unfamiliar, in June 2014, Intel revealed that it was working on an integrated FPGA-Xeon solution. On Wednesday, Nov. 18, at the Structure Conference, the company's Data Center Group head Diane Bryant divulged that the new server chips would be arriving in Q1 2016.
"Combining Xeon with FPGAs gives Intel a more powerful and programmable chip that can plug into existing Xeon slots," noted Bryant.
The integrated programmable chips will initially ship to cloud-scale companies like Facebook, Amazon Web Services, Google, Baidu, Alibaba, Tencent and Microsoft which run the biggest data centers.
The server chips are expected to aid Intel's customers such as PayPal and Microsoft to run both search and transaction algorithms.
"We'll be shipping it to the largest cloud service providers in Q1 so they can begin tuning their algorithms," said Bryant.
The Intel Xeon processors with FPGAs are poised to bring a lot of flexibility. The server chips are a part of the company's efforts to widen the use of the accelerators deployed in the server processors, as well as the workloads the chips can take.
The major advantage of FPGAs is that software can be used to reprogram them once they are manufactured. This is the primary reason they are fast becoming the significant accelerators for both web-scale and cloud environments as workloads in these settings alter swiftly.
The FPGAs are programmed in such a manner that they run only a specific algorithm. They are said to be more efficient than the regular CPUs when gauged on the performance-per-watt criteria.
According to Intel's predictions, in the next five years, the FPGAs could potentially be deployed in nearly 30 percent data centers.
The company is yet to divulge details on the chips, with Bryant only going as far as to disclose that its "footprint" will be compatible with the current crop of Xeon processors.
Photo: Jiahui Huang | Flickr
