datasheets.com EBN.com EDN.com EETimes.com Embedded.com PlanetAnalog.com TechOnline.com  
Events
UBM Tech
UBM Tech

Product Review

Achronix announces new 22nm Speedster22i FPGAs

Clive Maxfield
4/24/2012 10:10 AM EDT
The folks at Achronix Semiconductor have just announced the details of their forthcoming Speedster 22i HD and HP product families, the first FPGAs to be built on the Intel 22nm process technology.


The Speedster22i FPGA are described as being the industry’s only application targeted, high-end FPGAs – they consume half the power at half the cost of equivalent high-end FPGAs implemented at the 28nm technology node.

Speedster22i devices are the first FPGAs to include fully integrated hard IP protocol functions targeted for communications applications. The hard IP in Speedster22i devices includes the entire I/O protocol stack for 10/40/100 Gigabit Ethernet, Interlaken, PCI Express Gen 1/2/3 and memory controllers for 2.133 Gbps DDR3. In other FPGAs, these functions are implemented in the programmable fabric which consumes up to 500,000 equivalent look-up-tables (LUTs), makes timing closure difficult, and increases the cost and power consumption of the FPGA solution. Additionally, the embedded hard IP in Speedster22i FPGAs eliminates the cost of purchasing, integrating and testing these functions as soft IP.

If we look at the trends for increasing capacity/density and performance, we see that they are beginning to "tail off" with existing FPGA technologies:



Similarly, if we look at static and dynamic power consumption, we see that this is ramping up (note that all of these graphs reflect the highest density devices available at that technology node):


So the 22nm Speedster22i devices – with their fully integrated hard IP protocol functions – are targeted at the "sweet spot" between existing FPGAs (which are general-purpose and intended to address a range of applications) and ASSPs and ASICs (whose functions are "frozen in silicon):


“We worked closely with leading companies in the communications market segment to design our FPGAs to meet their performance, power, and cost requirements”
said Robert Blake, Achronix’s President and CEO. “The combination of Intel’s 22nm process leadership and our innovation in both the core fabric and embedded hard IP for targeted applications means that our customers will have a high end FPGA solution that is half the power and half the cost of competing FPGAs.”

“Part of our differentiating strategy is to integrate best-in-class, silicon-proven IP,” said John Lofton Holt, Achronix’s founder and Chairman of the Board. “For example, beyond the dramatic power and performance advantages afforded by Intel’s 22nm process, our Speedster22i devices also leverage an entire portfolio of industry-leading I/O, core and packaging IP that was developed by Intel.  This helped us achieve previously unreachable levels of performance and signal integrity, and at the same time, reduce our development time and development costs.”

Two FPGA families for different target applications
Speedster22i has two product families, high density (HD) and high performance (HP) that share the same I/O functionality and hard IP. Both families leverage the Achronix CAD Environment (ACE) design platform, which provides a seamless and familiar tool environment for designers.

 
HD Family: The HD FPGAs are a synchronous based family that offers the combination of the highest density FPGA with the lowest power consumption. There are 4 members in the HD family with the largest device having 1.7 million effective LUTs and 144 Mb of embedded RAM. Additionally, with up to sixteen 28-Gbps SerDes, sixty-four 12.75-Gbps SerDes and 960 general purpose 2.133-Gbps I/O, the HD family offers the industry’s highest I/O bandwidth, which is critical for high-end switch and bridging applications.

HP Family: The HP FPGAs use Achronix’s patented picoPIPE self-timed architecture, operate at up to 1.5 GHz and are 3x to 4x faster than synchronous FPGAs. Speedster22i HP FPGAs are designed to achieve maximum performance for feed-forward data flow and DSP applications. The largest member of the HP family has 250 thousand LUTs and 64 megabits of embedded RAM.

Power advantage
For the communications target applications, the embedded hard IP consumes up to 90% less power than implementing the same functionality in the programmable fabric of general purpose FPGAs. Additionally, Intel’s 22nm process technology with FinFET transistors consume 50% less power and are nearly 40% faster than transistors built on 28nm planar processes. For the HD devices, these combined factors result in up to 50% less total power consumption than mainstream FPGAs.

Robust 4th generation design tools available today
Both the HD and HP families are supported in Achronix’s mature and easy-to-use ACE design tools version 4.2 which are available now. The ACE design tools are built on the industry standard Eclipse Foundation open source platform and run under both Windows and Linux operating systems. Additionally, Achronix provides HDL Synthesis tools from both Synopsys and Mentor as part of the ACE tool suite.

“The Speedster22i application targeted FPGAs offer a clear cost and power advantage, but our customers are also focused on design productivity,” said Steve Mensor, Achronix’s Vice President of Marketing. “Our robust ACE design tools are now in their fourth generation and combined with the Speedster22i core performance advantage and embedded hard IP, in most cases, make timing closure a push-button operation.”

Early access program
Achronix is focused on the success of its customers and has an early access program to help customers learn the ACE design tools and complete their designs. This program includes evaluation boards, onsite training and technical support. Customers should contact Achronix to obtain more information about the early access program.

Device availability
Engineering samples of the HD1000 will begin shipping in Q3 2012. The HD1000 is the industry’s largest FPGA with over 1 million effective LUTs and 84 Mb of embedded RAM. The remaining HD and HP devices will be rolled out in the following 12 months.


If you found this article to be of interest, visit Programmable Logic Designline where you will find the latest and greatest design, technology, product, and news articles with regard to programmable logic devices of every flavor and size (FPGAs, CPLDs, CSSPs, PSoCs...).

Also, you can obtain a highlights update delivered directly to your inbox by signing up for my weekly newsletter – just Click Here to request this newsletter using the Manage Newsletters tab (if you aren't already a member you'll be asked to register, but it's free and painless so don't let that stop you [grin]).




Max the Magnificent

4/24/2012 10:28 AM EDT

Originally the folks at Achronix were focusing only on their Asynchronous technology. Personally I love Async, but I understand that the amazing FPGA performance (1.5GHz) is only available for "flow-through" type DSP algorithms ... my understanding (although I may be wrong) is that the performance is impacted when you have feedback.

So it makes a lot of sense that they would create two device families -- a synchronous family for regular processing and an asynchronous family for flow-through algorithmic c"crunching"

I also like the idea of application-targeted devices -- it's going to be interesting to see how this plays out over the next few years.

Sign in to Reply



Please sign in to post comment

Navigate to related information

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
Browse the technical library
Our technical library houses over 4,000 high-quality sponsored white papers, application notes, reference guides, use cases—all organized by company.