Slashdot Mirror
SeaMicro Unveils 512 Atom-Based Server
1sockchuck writes "Stealthy startup SeaMicro has unveiled its new low-power server, which incorporates 512 Intel Atom CPUs, a load balancer and interconnection fabric into a 10u server. SeaMicro, which received a $9.3 million government grant from DOE to develop its technology, says its server uses less than 2 kilowatts of energy — suggesting that a single rack with four SeaMicro units and 2,048 CPUs could draw just 8 kilowatts of power. Check out the technical overview, plus additional coverage from Wired, GigaOm and VentureBeat."
Wow. Just think how many servers you could run in VMWare on that. A hundred would be a decent functional number.
The question is, how good is the performance for, say, intensive numerical computations? Is the gigaflop per watt convenient?
"I'm never quite so stupid as when I'm being smart" (Linus van Pelt)
...misread the headline as saying that somebody had made a server out of only 512 atoms (as in the particle, not the cpu)?
File under 'M' for 'Manic ranting'
Before it was Fifth Grade Kindergardeners, now it's Atom/Molecule sized Intel Atom processors? :smoketheweed:
Imagine how may of these servers you could fit in the space of even a single grain of rice, let alone a standard 1U enclosure!
---
Play Six Pack Man. I
OK, OK. At least $EDITOR gets it right at the second attempt (8 kW of power).
Oh, wait. Nevermind.
Chas - The one, the only.
THANK GOD!!!
They're going to power a Ferrari out of 34.5 Vespa scooter engines.
Yes, but can it play Duke Nukem?
I don't know if 8 kW is a lot or a little less than a normal rack would draw. The article doesn't say...how much power would a "normal" rack consume? Isn't there a measure of computing power vs power used?
Or it's only me who can't find it?
"No changes to software" or something like that.... And only tons of RFC* and "funny acronyms"... What software needs no change?
http://opencm3.net, http://www.nongnu.org/gm2/
In all of the benchmarks that I've seen, clock for clock a Core 2 gets about twice the score of an Atom, sometimes more. The Core 2 uses a bit more than twice as much power, but if you have two Atoms you also need twice as many north-bridge chips and this pushes the power usage up to over what the Core 2 will consume. The newer Xeons do even better.
The first benchmark results I found that compared the two were PassMark benchmarks, where a 2GHz Atom scored 386 and a Intel Xeon X5680 at 3.33GHz scored 10620. The fastest Atom, the D510 at 1.66GHz, scored 662. Even if your code scales linearly, you need more than 16 of the fastest Atom that you can buy to replace one Xeon. Or, to put it another way, this 512-Atom machine is about as powerful as a 32-CPU Xeon.
A single Atom D520 draws around 13W, so 16 of them draw 208W. The Xeon will draw 130W. Drawing under 2KW for 512 Atoms means that they probably aren't using the fastest available ones. Actually, it means that they're drawing under 4W per Atom, which means that they're probably using Z-series Atoms, getting about half the performance of the D-series ones, so you'd only need about 16 Xeons for the same performance.
For most workloads, if the server is not busy, you'll get much better power usage from the Xeon as well. Power usage drops off dramatically when the CPU is not 100% busy. Unless you are turning individual atoms off completely, you can't scale back power usage nearly as well with the Atoms, as single processes that would not be CPU-bound on the Xeon will require an Atom core to run at full speed.
In other words, it sounds a lot more like greenwashing than anything that's actually sensible.
I am TheRaven on Soylent News
SeaMicro: 512 Atom processors in 10U rack
Dell PowerEdge R815: 4 Opteron processors, each one up to 12 cores, in 2U rack. In a 10U you can include 5 of these servers, which will bring 4 x 12 x 5 = 240 Opteron cores
More info here
I think the main point here is trying to make a server that can do many non-intensive parallel computations. But then when I look at a GPGPU (such as the Tesla c2050), you can get the same type of performance using a 2kW server in 3U (which I have here). The Atom is ~3GFLOPS per processor making this cluster ~1500GFLOPS strong. A single C2050 has ~500GFLOPS and you can load 4 of them in a single server. nVidia's S2050 has that performance in a single U.
Custom electronics and digital signage for your business: www.evcircuits.com
following the recent single-atom transistor development, an actual simple CPU that is built from 512 atoms of various elements?
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
Let me know when someone really does build a server with 512 atoms.
How can they make a computer out of 512 atoms?
There must be someone to mention the Beowulf cluster of these ! :)
So here, I've mentioned it
This is a good start- SM10000 System Overview
Interconnect is 1.28 Tbps or 2.5 Gbps per core.
I/O includes a minimum or 8 gige or 2 10-gige, which can be increased to 64 gige or 16 10-gige links per chassis.
This unit runs as 512 system images using stock 32 bit OS's. Each CPU may have 1 or 2 GB's of ram and up to 64 local drives may be installed and divided among the CPU's with the included management software. The unit supports PXE boot, so the system images may run off local disk or from a ram image.
Just to note, the Atom z530 is a single core, 32 bit only CPU, if that matters.
I couldn't tell you if the 16 10-gige links would seriously limit this box or not. You'd have to show me a data center with more than 160 Gbps of internet connectivity first. :) And that's assuming you only purchased one of these suckers, because you'd need that much per chassis.
The Internet has no garbage collection
The point isn't so focused on how many processors you can fit in that space, it's power consumption.
The link you provide details 2 1100W PSUs. That's in 2U. The summary quotes just under 2000W for one 10U server. Just looking at that, you're running at 1/5th the power consumption.
A ad off topic I know but I need to know. What's the deal about all these hexagonal, honeycomb like venting grills that I first saw appear on IBM gear many years ago but now seems to be ubiquitous to server gear from all vendors. Is there some cool scientific reason behind it? Is it about maximizing airflow, creating nice vortexes inside the machine to spread the air, reduce noise, reduce vibrations in the chassis, weight concerns, heat spreading.. what's the deal?
- Henrik
- when the Shadows descend -
No it isn’t. Because as a poster above mentioned, and as we all should know by now, Atom is a power consumption scam and is designed as one.
When you look at an Atom mainboard, what you think is the CPU, because of the cooler, is the north bridge. And what you think is the north bridge, because of its smallness, is the Atom CPU.
They simply shifted features to the north bridge, so they could trick you into believing the whole system would be more efficient, by giving you small consumption numbers for the CPU.
The joke is, that even those numbers are still about 5 times that of a equivalent ARM CPU. Which does not need the large NB.
So the point here is eyewash, and to sell you a delusion. /. articles. [Especially the ones around anticompetitive behavior. Specifically against AMD when the Athlon came out. I personally suffered from this.])
But I did not expect better from the convicted criminal that Intel is. (Look up the lawsuits yourself, in case you lived in a cave and missed the
Any sufficiently advanced intelligence is indistinguishable from stupidity.
It's only a scam if you make outrageous claims that it can be used for everything, like building an efficient supercomputer. The Atom has a good purpose and that is low power CPU with some good capabilities you will need for (HD) media that will still allow your netbook to run for 10 hours (i've seen them actually work over 8 with WiFi and normal workload). The only other netbook option i've seen that can also do this is some underpowered Pentium M version... but I doubt that it can decode HD smoothly despite it's north- *and* south bridge, remember this optimization comes slow, and the Atom already is a big leap forward in lower power CPUs for regular use.
I've just ordered a little PC for downloading and some media playback with a total power use of 20Watt (that's including the ION chipset), which is still able to play HD media... that's exactly what the Atom is good for, but not even close to optimal for servers or really low power computers.
Crysis?
I've seen this posted before in other Atom stories and it's 100% BS. The Atom is a full CPU, they did not shift any features into the northbridge as you claim. The 945 chip often used on Atom motherboards has been around for a while, and was used in systems before the Atom even existed. The 945 is just not very power efficient so that's why it needs more cooling than the CPU.
The Atom can be paired with other northbridge chips, notably Nvidia's ION. If part of the CPU was in the 945 chip as you claim this would be impossible.
Did anyone else get disappointed when after reading the summary, they found out it wasn't about nano-technology?
Lawsuits are civil proceedings, not criminal. Losing a lawsuit does not make you a convicted criminal.
The scope of atomic manipulation recently could really get this crazy.
Been there, done that, paid for the T-shirt
and didn't get it
This seems to be the poor man's answer to Sun's Niagra/CMT kit.
Yeah, at least someone gets it. The entire server is a virtualization host. It can allocate up to 512 VMs each with very deterministic QoS parameters. The CPUs are so cheap and plentiful, you don't need to share them between guests. The real hardware to virtualize is the I/O hardware, hence the features around sharing a big ass disk array and network interconnect via custom ASICs.
So I must ask why this over Operons, Xeons, or Arm?
That was my first thought too. Considering some of the benchmarks that show ARM performance per watt that crushes the Atom. But according to the Wired article:
Though SeaMicro has used Atom processors for its chipset, the company says it has designed its architecture to be flexible and support any CPU. So any low-power chip included that from ARM, which runs on most smartphones today, can become a part of SeaMicro’s system. But Atom remains the best choice for now, says Feldman. ARM processors used in cellphones consume much lower power than an Atom chip but they also cannot deliver the same kind of computing performance, he claims.
So they say their system can handle it but they don't think ARM's have enough performance. I'd tend to doubt that and instead bet that they just haven't built their custom chip to handle the ARM yet. If they did, they could take advantage of ARM's extensive hardware power saving features. I would think it would be much cheaper too since you could have a SOC that would do what appears to take two Intel chips here, even after SeaMicro has bundled a lot of the rest of the board functions in their own custom ASIC. Less chips and cheaper licensing from ARM should make for a cheaper and lower power solution.
At the low end you want a battery to drive a mobile device ideally for weeks at a time. At the high end you want to tie zillions of devices together without spending a fortune on power and A/C. Maybe the high end solution will use larger number of low-end devices, but with an overall lower TCO.
Score:5, Bullshit
The CPU is a full-fledged CPU. The Northbridge, as others have pointed out, is simply old.
Stop lying, and get a grip. It's painful to read posts like yours. The level of whining is simply staggering.
While "shifting features" is indeed not accurate, the point stands. Compared to ARM (or even Athlon and Nehalem/Lynnfield processors), Atom (prior to Pinetrail and Moorestown) was less integrated. So to compare power numbers, you can't just look at the CPU itself but need to include the memory controller (when compared to Athlon and Nehalem/Lynnfield) and PCI-E controller (when compared to Nehalem/Lynnfield). When compared to most ARM SoC's, you'll need to include graphics, miscellaneous I/O, GPS, video processor, image processor, memory controller, USB, CPU bus, peripheral bus, chip-to-chip interconnect, and sometimes a 3G modem.
Even Moorestown is very misleading when they did their comparison with ARM. They conveniently showed off numbers for tasks that use the dedicated components (video decoding, audio playback, talk time on GSM/CDMA, etc.) -- much of which was done by the "helper" chip and not the chip Intel fabbed themselves -- but left out any tasks that would tax the main CPU. Comparing those numbers to ARM SoC's made it look competitive. Then they showed performance numbers of the CPU compared to ARM CPU's -- which, btw, seemed to increase linearly with number of threads, notice how a dual-core A9 was twice the performance of a single? -- but left out the power numbers for when the CPU was taxed heavily.
I mean, it's marketing. But that hardly means it's not underhanded.
Your claim of 13W for an Atom CPU is misleading because only the model with integrated graphics uses that much (which would indeed give it a terrible performance/watt rating, no better than the old 130nm Pentium M). A quick look at one of the 'FAs does not turn up which CPU they actually used but a 300-series dual core Atom at 8W would seem to perform almost as well as a ULV Core 2 Duo at 10W TDP. The only Atom CPU, IMO, that really has an impressive performance/Watt is the 2.0GHz single core, 2.5W Z series. Comparing only TDP may be ignoring the benefits of a lower idle power consumption for the Atom, however.
We can safely assume that the most scalable OS on this planet, Windows, is powering this thing right?
Because of the way VCPUs are scheduled (at least in VMWare) that 8-vCPU VM won't get a time-slice until such time as there are 8 real cores available for the duration of that slice.
That is insane. I seriously doubt it applies to anything other than VMWare, however. No doubt the VMWare folks are working to remove whatever limitation possessed them to make such a ridiculous design as well.
the missing components are overall powerdraw (looks like the Atom based server would win here) and what the task at hand is (if its something that can be massively parallelized it could come out faster on the Atom.) also price comparisons here are important an R815 starts at $13,000 without a 3rd or 4th CPU. there are definitely better options than both of these if you're looking only at the number crunching per U of rackspace, but thats no way to build an environment.
This was fixed in ESX 3.5 somewhere around update 2 or just after. I can't remember off the top of my head which update fixed it. We are using 2 VCPU and 4 VCPU vm's in Vsphere 4 Update 1 and they work quite nicely.
I have one question. If the Japanese Ministry of Agriculture is not in charge of Gundam, then who is?
Did you check the power consumption on 240 opteron cores? Each core consumes about 75% more power than each Atom CPU. In CPU power alone you are consuming 1.5kW. The opteron also needs a northbridge for each CPU. Add the northbridge and you're almost at 2kw (the SeaMicro's total consumption), not including the rest of the power consumed by 5 2U servers. The SeaMicro is under 2kw for the entire system.
You'd definitely have a point if you weren't talking out your ass. A simple comparison of actual figures shows you don't have a clue about what SeaMicro has done here. They specifically address the fact that ordinarily the Atom requires a northbridge for each CPU, however their design does not.
Security is mostly a superstition... Avoiding danger is no safer in the long run than outright exposure. - Helen Keller
Can I ask?
Forget Atom and the legacy x86 - where're the ARM servers?
Most human behaviour can be explained in terms of identity.
I read this and initially thought how it might compare to the Plug Computer's I've been using for the past year that use less than 5 watts each and which new models now include multiple USB, multiple eSATA, wifi, bluetooth, 1 Gb ethernet etc for $99. So far the Plug computer's I've used have been
terrific in price/performance/features. One system has 3 plugs running linux, internetworked via wifi, total 1.5TB mini usb, LAMP installed serving music/videos to my home for less than 30 watts pwr. Well, I thought it was cool.
When I read this article I initially thought when they mentioned Atom Powered Cloud Computer Server... there was some kind of XEN etc virtualization involved but have not yet found any mention of that so is this more of a large grid of small servers or have they done something to allow the Atom to provide some sort of virtualization.... if not then I think the "cloud" moniker is overblown.. but which would generate buzz. But they did mention the "secret sauce" perhaps that silicon has kvm, xen etc capabilities.