The developer’s dilemma for any new FPGA-based acceleration platform, comprising FPGA hardware design, host software stack and application workloads, centers on how much to develop from scratch versus reuse or license. Schedules are often tight and the design team may not have all of the necessary hardware, software, and application development expertise.
The new Intel® Open FPGA Stack (Intel® OFS) now offers hardware, software, and application developers a scalable, source-accessible infrastructure with standard interfaces and APIs that enables them to build custom acceleration platform solutions. Developers are already using this second-generation hardware and software infrastructure to develop solutions based on Intel and third-party platforms featuring the Intel® Stratix® 10 and Intel® Agilex™ FPGA solutions. All Intel OFS hardware and software code has been developed using an open source development methodology. Users are granted access to the source code – which Intel refers to as source accessible.
Intel OFS also enables scale and eases deployment. From a hardware perspective, this scalability comes from the code having been developed in a modular fashion. Developers need only implement the blocks relevant to their design. The Intel OFS hardware code is composable, meaning it is easy to build application-specific FPGA designs using this IP. These “shells” cover key memory, networking, CPU, and datapath elements needed to allow communication to and from the FPGA. Intel OFS hardware code employs industry standard AXI interfaces to make this IP easy to reuse. Intel OFS also addresses software scalability by supporting both bare-metal and virtualized deployment models. In addition, Intel OFS supports remote updates, which can be downloaded without the need for host reset or re-initialization.
Both the hardware and software elements of Intel OFS employ standard software APIs and driver interfaces, which enables the use of common application frameworks. These APIs and driver interfaces are available in the Linux kernel at kernel.org so that third-party OS software development and distribution companies can build native support for Intel OFS-based hardware directly into their own distributions.
The base Intel OFS code is Intel’s responsibility. Intel will continue to upstream and package the user space and pull the code forward with the Linux Foundation’s latest long-term support initiative (LTSI) kernels. Additionally, Intel OFS includes high-level design board support and Intel is sharing simulation and validation models to jump start validation efforts. Hardware developers can develop their platform, take and clone the Intel OFS software code, and port this code to their hardware. Along the way they may add or modify drivers as needed for their custom designs and they can tune the board support package (BSP) for their target applications.
Intel is working with software distribution vendors to incorporate this base code via the kernel and user space to provide a foundation for native Intel OFS support in their distributions. These vendors can collaborate with hardware acceleration platform providers to incorporate specific BSP support such as drivers into their distributions for these third-party hardware platforms.
Application development generally strives to maximize the ROI, which includes two critical components: development cost and revenue. To minimize development cost, development teams must consider whether they should develop an application from ground up or port an existing workload, which will likely require at least some modification. With Intel OFS, teams now have the option of choosing from Intel-developed, third-party, or proprietary acceleration platforms based on the same infrastructure with standard interfaces and APIs.
The recently announced Silicom FPGA SmartNIC N5010 is the first platform to leverage Intel OFS for its BSP. Silicom is one of the first Early Access Program (EAP) members for Intel OFS. (See “SmartNICs based on Intel® FPGAs Boost Converged Broadband Network Performance” and “The Next Platform discusses the latest Intel Networking Innovations including new Intel® SmartNICs based on Intel® FPGAs” for more information about this product.)
EAP for Intel OFS runs through 2020 and most of 2021. Intel OFS EAP members benefit from early access to Intel OFS code and documentation and receive direct support from the Intel Customer Experience Group (CEG). If you are interested in trying Intel OFS for your next project or interested in details about the Intel OFS Early Access Program, please reach out to your Intel sales representative to get started.
Once Intel OFS transitions to general availability, it will be available under a robust self-service support model, as one would expect from any open-source development methodology.
For the Intel News Byte about this announcement, click here.
Notices & Disclaimers
Intel technologies may require enabled hardware, software or service activation.
No product or component can be absolutely secure.
Your costs and results may vary.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.