Late yesterday, Intel quietly announced one of the biggest ever changes to its chip lineup: It will soon offer a new type of Xeon CPU with an integrated FPGA. This new Xeon+FPGA chip will fit in the standard E5 LGA2011 socket, but the integrated FPGA will allow each chip to be customized to specific workloads. This move is almost certainly intended to make Intel-x86 a better all-round platform for a wider variety of workloads in enterprise and data center settings, and to dissuade customers from switching to GPGPU accelerators from the likes of Nvidia.
The Xeon+FPGA also raises the question of whether Intel would ever consider integrating an FPGA into its consumer Core line of chips — it’s exceedingly unlikely, but it it’s hard to deny how awesome it would be if next-gen games and apps had access to an FPGA to speed up core processes. But more on that at the end of the story.
What is an FPGA?First things first, let’s talk field-programmable gate arrays (FPGA). As the name implies, an FPGA is essentially a blank chip that can be repeatedly reprogrammed after manufacturing. With very few exceptions, every chip inside your computer is hard-coded (at the time of manufacturing) to perform just one set of functions. Your CPU can only do exactly what Intel or AMD designed it to do. You can’t take your CPU and turn it into a GPU. But you can take an FPGA, program it to perform one set of functions (say, graphics), and then reprogram it to handle another type of workload (say, sorting through databases).
An Altera FPGAThe main advantage of an FPGA, other than its customizability, is that it has monstrously high performance. In much the same way that an ASIC is by far the fastest and most efficient way of processing a specific workload (and thus why they’re used for Bitcoin farming), an FPGA is also very fast and efficient. They’re not quite as fast or efficient as ASICs, but what you lose in speed you gain in reprogrammability (again, ASICs are set in stone at manufacturing time).
Why is Intel pairing its Xeon CPU with an FPGA?Over the last few years, as more and more processing has moved to the cloud, Intel’s dominance of the server market has been attacked on a few fronts. Xeon chips are good all-round chips — but there are plenty of cases where another, more-workload-specific chip might make sense. This is why people are starting to look at cheaper, low-power chips for web servers (which don’t need expensive, hot Xeon chips), and GPGPU accelerators for highly parallel processing (as in supercomputers). For these more specific workloads, Intel is pitching Atom for lightweight stuff, the 50-core Xeon Phi for supercomputers, and now the Xeon+FPGA (it doesn’t seem to have an official name yet).
Intel has tried to fight back against GPGPU accelerators with its own massively parallel chips, such as the Xeon PhiIntel also notes that, in the last year alone, it has delivered “15 custom products” for big customers like Facebook and Ebay (probably Xeons with more/less cache, more/less cores, etc.) — and it will do over 30 custom designs this year.
What’s the purpose of this new Xeon+FPGA product? In the words of Intel: “The FPGA provides our customers a programmable, high performance coherent acceleration capability to turbo-charge their critical algorithms.” Intel estimates that the Xeon+FPGA will see massive performance boosts in the 20x range (for code executed on the FPGA instead of a conventional x86 CPU — but obviously there will be big overall speedups as bottlenecks are removed. The other advantage is that workloads change — so if your critical algorithms change, or your whole company pivots, the FPGA can be repurposed without having to buy lots of new hardware.
An eight-core Xeon CPU, with cache down the middle and various controllers/IO along the left and right edges.As far as tech specs go, we can sadly only guess at this point. The FPGA is probably manufactured by Altera, which Intel has a close working relationship (it was one of the first companies that was allowed to use Intel’s closely guarded chip fabs). FPGAs in general are very bulky (that’s the trade-off of using programmable gates), and there isn’t a whole lot of spare space on an LGA2011 Xeon E5 chip package, so we’re probably not looking at a very large FPGA. Though if the FPGA can leverage the Xeon’s cache, and perhaps some other low-level resources, some space-saving efficiencies might be realized. (Intel says there’s a “low-latency coherent interface” between the Xeon and the FPGA, but that’s about it.)
There’s no word on pricing or availability — but obviously it won’t be cheap (the top Xeons already cost thousands of dollars). Intel is hoping that the Xeon+FPGA package is enticing enough to convince enterprises to stick with x86, rather than moving over to a competing architecture (such as Nvidia’s Tesla GPGPUs). It is probably a lot easier to rewrite a small segment of critical code to run on an FPGA than to rewrite the entire application in OpenCL.
Finally, a fun thought experiment: If this little Xeon+FPGA experiment is a success, might we one day see a Core+FPGA chip? Or a GPU+FPGA setup? Imagine if you could load a core part of a game engine into the FPGA; the physics engine perhaps, or the enemy AI. Or a core part of your PC’s operating system. That could be very cool indeed.
Tagged In hardwarecomponentsintelsiliconx86serversenterprisexeonfpgaXeon E5alteradata centerShare This Article .article {margin:0px !important;}.AR_1 {margin :0 0 20px 0 !important;}.AR_2 {margin:0 0 20px 0;} CommentPost a Comment chojin999FPGA advantages over DSP are more marketing than reality…
Most benchmarks don’t give such excellent results on FPGA devices. After all.. FPGA devices actually are nothing more than old architecture DSP units… fixed, hard-wired, functionality limited units that already existed before the DSP era began.
Nowadays manufacturers are selling FPGA as the Second Coming Of Jesus .. but the marketing hype is too much and unjustified.
FPGA devices are better than DSP on some embedded systems and pretty limited cases.. it’s nothing magic and nothing revolutionary.
Indeed in some cases it’s better using FPGA over DSP for lower power consumptions and lower production costs ..but it’s not true that all of a sudden DSP now are obsolete and FPGA so much better.
So adding FPGA devices directly inside Xeon CPUs might be useful for some algorithms but all the DSP units like the SSE ones and AES-NI already there can already do wonders with properly coded algorithms.
Huh? Who’s talking about DSPs?
And FPGAs aren’t hard-wired — that’s the whole point?
ShanieOneillnucmy Aunty Allison
recently got a nice 6
month old Jaguar by working from a macbook.this website Cashduties.COM
So if FPGAs are so much faster and flexible how come there aren’t desktop chips designed entirely with this stuff? You said it was a bit of a space hog, but otherwise why not? Too power hungry and could fry eggs? Cost prohibitive for the general public?
http://www.mrseb.co.uk/ Sebastian AnthonyThey’re not for general-purpose computing (which is what you’re doing on a desktop PC). That’s what a CPU is for. There are hundreds of millions of logic gates in a modern CPU. There might only be a few thousand in an FPGA.
+ yes, they’re large, and power hungry, and all the other usual stuff.
You could do general desktop computing on an FPGA — but if you have a firm workload that doesn’t change, why use an FPGA? Much better to use a conventional CPU.
http://www.classicdosgames.com DOSGuyThe traditional, permanent gates of CPUs and GPUs are faster than FPGAs. The problem with a CPU is that it’s a general purpose processor. It has gates devoted to doing a lot of different things, many of which may not be useful to you.
An ASIC (application specific integrated circuit) will always be faster than a general purpose processor for the specific task that it was designed to perform. For instance, Deeper Blue (the computer that beat Kasparov at chess) was full of ASICs that were specifically designed to analyze chess positions, which dramatically increased the number of moves per second that the computer could analyze. If it had been forced to use only general purpose processors, it wouldn’t have been able to consider as many moves, and consequently would have played a weaker game.
What Intel is doing here is giving customers a middle ground between a general purpose processor and an ASIC. The permanently gated part of the chip will be way faster at performing any given task than an identically configured FPGA ever could, so that’s why they’re not making Core i7s that are pure FPGA. But the Core i3/5/7 only has so many floating point units, so many registers, so many WHATEVER, and that number is set in stone. If you have a workload that would benefit from having a huge number of [insert resource here], you can program the FPGA to have those resources. They’ll be slower than if those resources were permanently gated, but without the FPGA you couldn’t have them at all. The FPGA allows you to supplement the CPU’s existing resources with whatever additional resources you need.
massaudon’t forget that creating an fpga “program” is much harder than just running a program on a general purpose CPU.
it is not just going from code to debug phase and done.
you have to do a logic simulation than a real simulation and than you could still do a custom rooting to get the fpga even faster. if the algorithm fit in the FPGA than it is faster than most GP processor.
one good note for FPGAs to get new ground are memresistors. for example mrFPGA it significantly reduces area (x5) power(1.63x) and performance(2.28x) at the same time while not going to 3d logic.
chojin999http://www.bdti.com/MyBDTI/pubs/info_stanford02_fpgas.pdf
“Comparing FPGAs and DSPs for Embedded Signal Processing
[...]
Performance Analysis
•Comparing performance of off-the-shelf DSP
to that of FPGAs is tricky
•Common MMACS metric is oversimplified to
the point of absurdity
•FPGAs vendors use distributed-arithmetic
benchmark implementations that require fixed
coefficients
•MMACS metric overlooks need to dedicate
resources to non-MAC tasks
•Many important DSP algorithms don’t use MACs at
all!”
http://www.ti.com/lit/wp/spra879/spra879.pdf
“
Choosing the Right Architecture for Real-Time Signal Processing Designs
[...]
2.1.1.4 FPGA (Field Programmable Gate Array)
FPGA was judged good in the matter of time to market. FPGAs avail field-ready modifications to achieve functions. Their flexibility is not as high as software programmable alternatives, thus they fall below DSP, MCU and RISC in their rating for time to market. However, they have better support and faster cycle time than ASSP, configurable processors or ASIC, and thus can claim faster time to market than those alternatives.
FPGA was judged excellent in the matter of performance. With FPGAs, developers can tune hardware gates specific to the application, delivering high application-specific performan
FPGA was judged poor in the matter of price. FPGA is by far the most expensive alternative discussed here, achieving a poor rating.
FPGA was judged poor in the matter of power. FPGA is the bottom of the pack in power efficiency, an inherent trait in FPGA circuit technology as well as the overhead power of unused gates in the array. Technology advances will lower FPGA power, but likely not enough to change its place in the relative ranking on power efficiency.
FPGA was judged excellent in the matter of development ease. FPGA would rate the best on development cost assuming two situations: that the toolset for FPGA programming is not too expensive; and, assuming the developer is dealing primarily with hardware, that the engineer is involved in the development. If development leans towards software engineers, then FPGA would increase in effort and relative cost. In terms of development help, the FPGA tools and support structure for FPGA-based designs seems to be well established and acceptable to OEMs.
FPGA was judged good in the matter of feature flexibility. It can be field reconfigured for additional features or changes. However, this act of hardware reprogramming of an FPGA is more difficult and the achievable feature sets are more limited than software programmable
solutions like DSP.
“
FPGAs gates aren’t fully programmable software wise, they can be reconfigured on some architectures even a lot but not as much as DSP or CPUs. FPGAs must be designed for specific functions, even if thru hacks someone manages to make them run algorithms they weren’t designed for the performance hit would then be so high that is not worth the trouble. While modern DSP can be fully programmed to compute algorithms they weren’t designed for (and still getting a performance hit when trying to run tasks common on general purpose CPUs). Older DSP architectures weren’t so flexible…
Ivor O’ConnorSeems like this might spawn a whole new specialist niche that could command million dollar figures.
Matt MenezesI was just thinking this. I bet this is a boon to Altera FPGA programmers.
carol argoOutch !it is a nice idea on paper but gees only the bigguns are likely to bother(Google,ms,facebook,)
massauCore+FPGA this already exists. zync and sockit fpga is an arm core with fpga but it is damn expencive and there are no dimms used. so it isn’t really an upgradable system.
FollowFollow @ExtremeTech!function(d,s,id){var js,fjs=d.getElementsByTagName(s)[0],p=/^http:/.test(d.location)?'http':'https';if(!d.getElementById(id)){js=d.createElement(s);js.id=id;js.src=p+'://platform.twitter.com/widgets.js';fjs.parentNode.insertBefore(js,fjs);}}(document, 'script', 'twitter-wjs');ExtremeTech Newsletter
Subscribe Today to get the latest ExtremeTech news delivered right to your inbox.
More Articles ET deals: Today only! Toshiba Satellite Core i5 laptop for $450 Jun 19How to bring back the Start menu and button to Windows 8 Jun 19The great flying car dream: With Terrafugia, Toyota, and Musk on board, has their time finally come? Jun 19Who in their right mind would buy Amazon’s Fire Phone? Jun 19Hacker hijacks thousands of Synology storage devices, forces them to mine 500 million Dogecoins Jun 19 About ExtremeTechAdvertisingContact ExtremeTechET ForumsTerms of UsePrivacy PolicyZiff DavisJobs AdChoice/**/var _bap_p_overrides=_bap_p_overrides||{};_bap_p_overrides[8]={new_window:true}; (function(s, p, d) { var h=d.location.protocol, i=p+"-"+s, e=d.getElementById(i), r=d.getElementById(p+"-root"), u=h==="https:"?"d1z2jf7jlzjs58.cloudfront.net" :"static."+p+".com"; if (e) return; e = d.createElement(s); e.id = i; e.async = true; e.src = h+"//"+u+"/p.js"; r.appendChild(e);})("script", "parsely", document);Use of this site is governed by our Terms of Use and Privacy Policy. Copyright 1996-2014 Ziff Davis, LLC.PCMag Digital Group All Rights Reserved. ExtremeTech is a registered trademark of Ziff Davis, LLC. Reproduction in whole or in part in any form or medium without express written permission of Ziff Davis, LLC. is prohibited.
This post was made using the Auto Blogging Software from WebMagnates.org This line will not appear when posts are made after activating the software to full version.
No comments:
Post a Comment