Shifting to an FPGA Data Center Future: How are FPGAs a Potential Solution?
April 04, 2022 by
Jake Hertz
As data centers are put under more pressure, EEs are looking at field-programmable gate arrays (FPGAs) as a potential solution. However, how could they be useful, and who is ramping up their research efforts?
Today more than ever, the data center is being put under enormous strain. Between the increasing popularity of cloud computing, the high rate of data creation, and new compute-intensive applications like machine learning, our current data center infrastructures are being pushed to their limits.
To help ensure that the data center of the future will be able to keep up with these trends and continually improve performance, engineers are reimagining data center computing hardware altogether.
From this, one of the most important pieces of hardware for the data center is the
FPGA.
View attachment 31878
A high-level overview of an FPGA. Image used courtesy of Stemmer Imaging
A
recently announced center, the
Intel/VMware Crossroads 3D-FPGA Academic Research Center, is hoping to spur the improvement of FPGA technology explicitly for data centers.
In this article, we’ll talk about the benefits of FPGAs for the data center and how the new research center plans to improve the technology even further.
A Shift to Accelerators
There are currently two major trends in the data center that are driving the future of the field: an increase in data traffic and an increase in computationally-intensive applications.
The challenge here is that, not only must the data centers be able to handle
increased data and tougher computations, but there is a greater demand to do this at lower power and higher performance than ever before.
To achieve this, engineers have shifted away from more general-purpose computing hardware, such as central processing units (CPUs) and graphics processing units (GPUs), and instead, employ hardware accelerators.
View attachment 31880
An example of heterogeneous architecture, which is becoming the norm in the data center. Image used courtesy of Zhang et al
Engineers can achieve higher performance and low power computation with application-specific computing blocks than previously possible. To many, a heterogeneous computing architecture consisting of accelerators, GPUs, and CPUs, is the widely accepted path forward for future data centers.
Benefits of FPGAs for the Data Center
FPGAs are uniquely positioned to
benefit the data center for several reasons.
First off, FPGAs are highly customizable, meaning that they can be configured for use as an application-specific hardware accelerator.
In the context of the data center,
engineers can configure FPGAs for applications like machine learning, networking, or security. Due to their software-defined nature, FPGAs offer easier design flows and shorter time to market accelerators than an application-specific integrated circuit (ASIC).
View attachment 31881
An example diagram showing how FPGAs can be dynamically reconfigured. Image used courtesy of Wang et al
Secondly, FPGAs can offer the benefits of versatility. Since an FPGA's functionality can be defined purely by HDL code, a single FPGA can serve many purposes. This functionality could help reduce complexity and create uniformity in a system.
Instead of needing a variety of different hardened ASICs, a single FPGA can be configured and reconfigured for various applications, opening the door to further optimization of hardware resources.
Thus, some FPGAs can be reconfigured in real-time based on the application being run, meaning a single FPGA can serve as many roles as needed.
A 3D-FPGA Academic Research Center
Recently, the Intel/VMware Crossroads 3D-FPGA Academic Research Center was announced as a multi-university effort to improve the future of FPGA technology.
The team, which consists of researchers from the University of Toronto, UT Austin, Carnegie Mellon, and more, focuses their efforts directly on the role of FPGAs in the data center. More specifically, the group will be investigating ways to achieve 3D integration within the framework of an FPGA.
The idea is that, by being able to stack multiple FPGA dies vertically, researchers should be able to achieve a higher transistor density while also balancing performance, power, and manufacturing costs.
Overall, the group hopes to use 3D-integration technology to create heterogeneous systems consisting of FPGAs and hardened logic- accelerators, all within a single package. The technology will
seek to combine a Network-on-Chip (NoC) in a layer beneath the traditional FPGA fabric such that the NoC can control data routing while the FPGA can provide the computation needed.
Overall, the group hopes to extend the rise of in-network computing into the server with their new technologies.
FPGAs for Future Data Centers
The FPGA will undoubtedly become a key player as the data center trends towards more data and more intensive computation.
With a new research group hoping to bolster the technology, it seems even more apparent now than ever that FPGAs are becoming a mainstay in the data center industry.
Amazon’s Xilinx FPGA Cloud: Why This May Be A Significant Milestone
/
AI and Machine Learning,
CPU GPU DSP FPGA,
Semiconductor / By
Karl Freund
Datacenters, especially the really big guys known as the Super 7 (Alibaba,
Amazon, Baidu, Facebook, Google, Microsoft and Tencent), are experiencing significant growth in key workloads that require more performance than can squeezed out of even the fastest CPUs. Applications such as Deep Neural Networks (DNN) for Artificial Intelligences (AIs), complex data analytics, 4K live streaming video and advanced networking and security features are increasingly being offloaded to super-fast accelerators that can provide 10X or more the performance of a CPU. NVIDIA GPUs in particular have benefited enormously from the training portion of machine learning, reporting a 193% Y/Y last quarter in their datacenter segment, which is now approaching a $1B run-rate business.
View attachment 31882
But GPU’s aren’t the only acceleration game in town.
Microsoft has recently announced that Field Programmable Gate Array (FPGA) accelerators have become pervasive in their datacenters. Soon after,
Xilinx announced that Baidu is using their devices for acceleration of machine learning applied to speech processing and autonomous vehicles. And
Xilinx announced last month a ‘reconfigurable acceleration stack’ that reduces the time to market for FPGA solutions with libraries, tools, frameworks and OpenStack support for several datacenter workloads. Now Amazon has announced the addition of Xilinx FPGAs to their cloud services, signaling that the company may be seeing market demand for access to these once-obscure style of chips for parallel processing. This announcement may be a significant milestone for FPGAs in general, and Xilinx in particular.
What did Amazon Announce?
Amazon is not the first company to offer FPGA cloud services, but they are one of the largest. Microsoft uses them internally but does not yet offer them as a service to their Azure customers. Amazon, on the other hand, built custom servers to enable them to offer new public F1 Elastic Cloud instances supporting up to eight 16nm Xilinx Ultrascale+ FPGAs per instance. Initially offered as a developer’s platform, these instances can target the experienced FPGA community. Amazon did not discuss the availability of high-level tools such as OpenCL or the Xilinx
reconfigurable acceleration stack. Adding these capabilities could open up a larger market for early adopters and developers. However, I would expect Amazon to expand their offering in the future, otherwise I doubt they would have gone to all the expense and effort to design and build their own customized, scalable servers.
Why this announcement may be significant
First and foremost, this deal with the world’s largest cloud provider is a major design win for Xilinx over their archrival Altera, acquired last year by Intel, as Altera was named as Microsoft’s supplier for their FPGA enhanced servers. At the time of the Altera acquisition, Intel had predicted that over one third of cloud compute nodes would deploy FPGA accelerators by 2020. Now it looks like Xilinx is poised to benefit from the market’s expected growth, in part since Xilinx appears to enjoy at least a year lead in manufacturing technology over Altera with Xilinx’s new 16nm FinFET generation silicon, which is now shipping in volume production. Xilinx has also focused on providing highly scalable solutions, with support for PCIe and other capabilities such as the
CCIX interconnect. Altera, on the other hand, has been focusing on integration into Intel, including the development of an integrated multichip module pairing up one low-end FPGA with a Xeon processor. Surely, Intel wants to drag as much Xeon revenue along with each FPGA as possible. While this approach has distinct advantages for some lower end applications (primarily through faster communications and lower costs), it is not ideal for applications requiring accelerator pooling, where multiple accelerators are attached to a single CPU.
Second, as I mentioned above, Amazon didn’t just throw a bunch of FPGA cards into PCIe servers and call it a day; they designed a custom server with a fabric of pooled accelerators that interconnects up to 8 FPGAs. This allows the chips to share memory and improves bandwidth and latency for inter-chip communication. That tells us that Amazon may be seeing customer demand for significant scaling for applications such as inference engines for Deep Learning and other workloads.
Finally, Amazon must be seeing demand from developers across a broader market than the typical suspects on the list of the Super 7. After all, those massive companies possess the bench strength and wherewithal to buy and build their own FPGA equipped servers and would be unlikely to come to their competitor for services.
Amazon named an impressive list of companies endorsing the new F1 instance, spanning a surprising breadth of applications and workloads.
Where do we go from here?
The growing market for datacenter accelerators will be large enough to lift a lot of boats, not just GPUs, and Xilinx appears to be well positioned to benefit from this trend. It will now be important to see more specific customer examples and quantified benefits in order to gauge whether the FPGA is going mainstream or remains a relatively small niche. We also hope to see more support from Amazon for the toolsets needed to make these fast chips easier to use by a larger market. This includes support for application developers to use their framework of choice (e.g, Caffe, FFMPEG) with a simple compile option to target the FPGA, a goal of the recently introduced
Xilinx acceleration stack.
Datacenters, especially the really big guys known as the Super 7 (Alibaba, Amazon, Baidu, Facebook, Google, Microsoft and Tencent), are experiencing significant growth in key workloads that require more performance than can squeezed out of even the fastest CPUs. Applications such as Deep Neural...
moorinsightsstrategy.com