Johan, what will additionally be interesting is AMD's Seattle cores launching around the beginning of next year. the multi-core ARM architectures, integrated FreedomFabric and 10GbE on SoC, etc. will be a powerful combination. those too will be integrated into Moonshot. (heck, even the AMD X-series compares pretty well here...)
Looking forward to that AMD's Seattle too. I hope however that they will not be limited to the current SeaMicro boxes: those servers are extremely well designed hyperscale servers, but they are a massive investment for a large part of the market (most of our readers :-). Few of the web services providers are willing to shell out so much money (see Google, Facebook's minimal servers)
These low power high core count processors like the C2730 are interesting for SOHO NAS units. The core count and RAM capacity are also really interesting for VM platforms that use simple point of sale interfaces, since they aren't demanding in terms of calculation performance. Also, blade racks for virtualizing call centers and basic productivity remote desktops would see a boon from a high RAM and core count setup.
There's a lot of potential, but they only support VT-x so advanced VM setups need not apply. Kind of disappointing, but I guess they want you to buy into Xeon. Still beats ARM though in terms of capacity (RAM, cores, etc).
64GB available for 8 lightweight cores is more than enough, I think. It's not like these systems will be used for heavy duty stuff like a Xeon-based server.
Anyone know of any small (around nano-itx or pico-itx sized) boards that have Intel Atom Rangeley chips and dual GigE ports? I don't need any video out. A serial port is good enough for the project I'm planning on doing.
Too bad it has the GPU in it that's going to go unused. I'm sure this industrial application targeted board will probably cost the same as a larger Haswell based setup.
Just be sure to stay clear of the NANO-5050 based on my experience with the Intel D2550MUD2. They're both running GMA3650 graphics ("Cedar Trail", originally developed by PowerVR). There is very limited support in Linux and Windows. In fact, Intel only provides Windows driver downloads for *32*-bit Windows (nothing for Linux). Personally, I'll be very careful before buying another Intel based or provided motherboard.
Yes, where can I buy micro-boards (smaller than mini-ITX) with these. At least 2 gigabit Ethernet and 2 cores. I don't need a Bay Trail chip since I won't be connecting it to a monitor.
Nope. Out of curiosity, what is you design logic for wanting less than mini-itx, but gual GigE ports? I can think of a few applications, but basically all of them seem to be served as well by a router with an open source firmware.
In some synthetic benchmarks the 8-core C2750 is more than 10x faster than the S1260. My general characterization is that it feels more like half of a Intel Xeon E3-1220 V3 in terms of performance. Power consumption for a mITX C2750 platform including 32GB (4x8GB @ 1.35v) cooling and a SSD is 33w with everything including the 4x Intel i354 Gigabit Ethernet ports running.
About the same both on the Supermicro and ASRock units we have tested thus far. Happy to run some benches if you want. Should be getting the fourth platform in the next two weeks so can spare.
Other cool news is that the C2750 runs ESXi fine. If Intel included VT-d it would be a category killer with passthrough. I think you are right that the C2550 or C2730 may be better NAS chips because even on the 12 disk ASRock C2750D4I you can run RAID-Z2 and have plenty of CPU power left over.
Interesting comparison. However, I have a few nits to pick. "There is a major myth that the Intel Atom C2750 is many times slower in single threaded performance", but most of your "closer to realworld" benches tell exactly the opposite :-). See the crafty bench for example.
Johan that's not really a fair thing to say. The Xeon E3 and Opteron turbo to 3.7 and 3.6Ghz respectively. The Atom will turbo to 2.6Ghz. Compared to the Xeon L5520 (2.26Ghz), it's only around 20% behind. That's quite a feat for a cpu core that uses around 2W. With the clocks normalized, it's almost 5x faster than the S1260 it replaces, only 35% behind the Opteron and 2x slower than the E3. He concludes the following: "performance generally around half of the Intel Xeon E3-1200 V3 series". Seems right to me.
1. The clockspeed is part of the µarchitecture trade-offs. So you can not simply normalize clocks. 2. The point I was making was about "single threaded performance". It is not even close to being half that of a Xeon... more like one fourth. Not bad, but it is *not* a myth that Atoms are "several times" slower single threaded.
Not saying the Atom C2000 is a dog (it is pretty amazing in fact), just pointing out that the C2000 is still "several times" slower per thread. The myth has not been debunked yet...
Even there 1/3 of the E3 V3 performance at < 1/4 the platform power consumption on a single threaded application. The difference is that the C2750 is not 1/10th or less the speed like the S1260 was.
@geekfoo - will look into adding the bash one. The new development version of the script adds a redis benchmark and sysbench CPU tests. It also now works on CentOS as well as Ubuntu from a clean install:
Awwwww, from the title I was all excited to see an actual review of these two server SoCs, not just a feature comparison. Oh well, just have to wait awhile longer for that to come eh?
Johan just couldnt wait :). I understand him. From this arm generaion on its starting to become interesting. When was the last time something this happened?
If by interesting you mean a performance bump in line with expectations and the ability to address enough memory for it not to be a major limitation, than sure, this generation becomes interesting. Personally I found the first iteration more interesting for the simple fact that it actually beat Intel to the punch and provided a superior solution for the particular niche it was good for. Whereas this time around, well, the C2530 is almost at the bottom of Intel's Avoton lineup - only the one dual core model is below it. The more important point being that for any reasonably threaded workload Intel offers 8 core SKUs which, due to the fixed power consumption per node, greatly increase performance while only bumping up power consumption by ~33%.
The next round might be a bit more interesting but I doubt it. Note how the ECX-2000 projected performance increase is pretty much exactly as much as the projected power increase over the ECX-1000? And that's with going from a 40nm process to 28nm. Why exactly should we expect anything more from the next generation?
Your reasoning has some good points. However: the higher RAM capacity and CPU performance really allows that new kind of apps run on this thing. Apps that can benefit from the excellent network fabric for example. It will be interesting to see which realworld apps are bottlenecked by the limited CPU power and which ones are bottlenecked by network. Realworld server applications can sometimes be extremely different from the typical benches used. The ECX-2000 will not win any raytracing bench, but Calxeda does have a fighting chance turn as a Content Delivery Server or Drupal based server. All I am saying is, don't judge on raw performance or perf/watt.
So, the point is "don't judge on the performance, don't argue on the perf / watt, don't judge on the software, don't judge on the availability, don't judge on the power consumption (no data is given)." But see, "we have 10 g eth"
I am saying that raw perf/watt does not translate in realworld perf/watt when running realworld server apps. The S1260 looked on par with the ECX-1000 in the benchmark: about the same raw performance (integer crunching benches) and 8.5W vs 6W. The reality was that a S1260 server node needed 15 to 20W (vs 8W) and performed up to twice as slow in real server loads. That is 4-5 (!!) worse perf/watt ratio than expected from integer benches and TDP. Not saying the C2000 will be the same, just illustrating how far real server workloads can be from the other benchmarks.
He's saying the ARM chip, like all good RISC chips is good for certain applications like CDN. Real world and on other applications it might not hit its stride.
Real world also means that, because the ARM has less RAM and fewer cores, it has the potential to be bogged down in scenarios where there is a higher demand than the cache / RAM / threads can handle. It may be that those requests only require a small amount of network bandwidth, but if the processor can't schedule those tasks...then the CPU is a problem. However, if the CPU is handling a simple decode of a massive file, it can stream the results faster than network bandwidth. Real world perf is subject to a large degree of variability in task type.
wow, whoever they have laying out these moonshot module/sleds wants shooting :)
you could easily get 8 or even 16 SoC in that same space today im sure, perhaps even more.
Oh and Johan as you say you cant really compare OC but you should probably go into petty cash and get a few cheap Hardkernel ARM big.LITTLE A7 + A15 Exynos 5410 Octa-Core ODROID-XU's for generic testing etc
Granted, it is in no way comparable as the multithreaded performance probably falls WAY short, but as an interesting power consumption comparitor, my home server is based on a G1620 Ivy Bridge chip, uATX board (I forget the make, but H77 based, I want to say Gigabit?), 8GB of memory (2x4GB 1.25v), a pair of Intel CT gigabit NICs (motherboard port disabled), 60GB SSD for a boot/app disc and a pair of 2TB Samsung F4EG drives in RAID0 and a 80+ Bronze 350w PSU.
Whole thing pulls down 21w at idle and 32w with the drives spun up streaming. During heavy computational loads (as a simple test) the whole system pulls down a hair over 50W with both cores at 90+% and the HDD spun up.
I'd personally love to see the system as a whole pull down half those numbers. I really am curious looking, more from a comodity/lower requirement stand point, the newer Atom based Celeron and Pentium chips.
I'd be curious to see if anything exists with at least four SATA 6Gbps ports, RAID0/1/5 support, soldered on Atom based Pentium chip and at least one Intel based GbE port (with the option to stick another CT NIC in the PCI-e slot). Something tells me though that the newest celeron and pentium based atoms don't have support for 4 SATA 6Gbps ports though.
At least IMO, Calxeda is sandbagging a little about power consumption. A15 core is very power hungry, expecially over 1.6Ghz and Silvermont core has clearly showed a far lower power consumption under load. Dear Johan i can see a lot of marketing in Calxeda data, its pretty clear that this round is won by Intel with a wide margin. Maybe we need to wait some SPECpower submissions before any judgment. "Rough estimates" means nothing beacause Calxeda do not gives TDP data but only "average consumption" figures.
Last time, Calxeda marketing was simply wrong about their servers being good memcached or CDN servers. But we found that their power numbers were pretty close to reality. So they deserve the benefit of the doubt...We won't know until we unleash our server workloads on real machine, of course.
The thing about the preannouncement of the A53/7 and Intel's roadmap is that it's always hard not to wonder how microservers based on the *next* uarch will do. :)
The problem is that the integrated fabric might be a bit too complicated (= expensive mobo) for your usage model. There are probably better solutions for you on the market.
Yes, the new 16 Gigabyte DDR3 DIMMs to upgrade to 64GB RAM with just 4 sockets are available from the company I'M Intelligent Memory, website: www.intelligentmemory.com These memory modules come as UDIMMs (240 Pin) or SO-DIMMs (204 Pin), with or without ECC. As they are made of just 16 pieces of 8 Gigabit DDR3 chips, the modules are dual-rank and still have total 16GByte. No special form factor, everything looks standard. Intel recently released a BIOS update for their C2000 Avoton series to support those Intelligent Memory modules with 16GB capacity.. But the modules might also work in Calxeda and other platforms, because: When I look at the JEDEC DDR3 documentation, a 8 Gigabit DDR3 chip uses the same amount of address lines as a 4 Gigabit chip (A0 to A15). This means there is no hardware-modification required to run the modules. As a logical consequence, such 16GB modules should be working everywhere, in all types of CPUs, as long as the BIOS is programmed correctly to read the modules ID and set the values for the memory-initialization into the memory controller inside the SOC.
For those of us that actually run hosting companies and datacenters, these servers are actually worthless, for several reasons.
1) x86 is still king in the hosting arena. Everything else, I would never risk Enterprise customers with SLAs of 99.999% uptime on. 2) Underpowered. A nice Ivy Bridge system may pull more power, but it will handle proportionally more traffic. 3) Licensing problems - most Enterprise grade Cloud OS's are licensed out by the Socket, not the core. Therefore your best bang for the buck is something like an Ivy Bridge E or high-end AMD Opteron. Everything else is just a waste of licensing $$$.
Get me Haswell in an HP Moonshot form factor with mSATA SSDs, and then I might be interested. Until then, these are just toys.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
45 Comments
Back to Article
davegraham - Tuesday, October 29, 2013 - link
Johan, what will additionally be interesting is AMD's Seattle cores launching around the beginning of next year. the multi-core ARM architectures, integrated FreedomFabric and 10GbE on SoC, etc. will be a powerful combination. those too will be integrated into Moonshot. (heck, even the AMD X-series compares pretty well here...)JohanAnandtech - Wednesday, October 30, 2013 - link
Looking forward to that AMD's Seattle too. I hope however that they will not be limited to the current SeaMicro boxes: those servers are extremely well designed hyperscale servers, but they are a massive investment for a large part of the market (most of our readers :-). Few of the web services providers are willing to shell out so much money (see Google, Facebook's minimal servers)hoboville - Tuesday, October 29, 2013 - link
These low power high core count processors like the C2730 are interesting for SOHO NAS units. The core count and RAM capacity are also really interesting for VM platforms that use simple point of sale interfaces, since they aren't demanding in terms of calculation performance. Also, blade racks for virtualizing call centers and basic productivity remote desktops would see a boon from a high RAM and core count setup.There's a lot of potential, but they only support VT-x so advanced VM setups need not apply. Kind of disappointing, but I guess they want you to buy into Xeon. Still beats ARM though in terms of capacity (RAM, cores, etc).
JohanAnandtech - Wednesday, October 30, 2013 - link
The RAM capacity is still very limited for VM platforms. Two low voltage Xeons can offer up to 384 GB, 768 GB with LR DIMMs.theduckofdeath - Wednesday, October 30, 2013 - link
64GB available for 8 lightweight cores is more than enough, I think. It's not like these systems will be used for heavy duty stuff like a Xeon-based server.dylan522p - Tuesday, October 29, 2013 - link
Why did you ignore the Xeon that is Silcermont based?JohanAnandtech - Tuesday, October 29, 2013 - link
I am not sure what you mean but I fixed the table where Silvermont moniker was missingdylan522p - Tuesday, October 29, 2013 - link
Ya sorry. That was ignorance on my part.pikeike - Tuesday, October 29, 2013 - link
Anyone know of any small (around nano-itx or pico-itx sized) boards that have Intel Atom Rangeley chips and dual GigE ports? I don't need any video out. A serial port is good enough for the project I'm planning on doing.Thanks!
idos422 - Tuesday, October 29, 2013 - link
The smalled Silvermont based board with dual GigE I've found is this nano-ITX one using the embedded versions (Bay Trail I): http://www.portwell.com/products/detail.asp?CUSTCH...Too bad it has the GPU in it that's going to go unused. I'm sure this industrial application targeted board will probably cost the same as a larger Haswell based setup.
texadactyl - Friday, November 1, 2013 - link
Just be sure to stay clear of the NANO-5050 based on my experience with the Intel D2550MUD2. They're both running GMA3650 graphics ("Cedar Trail", originally developed by PowerVR). There is very limited support in Linux and Windows. In fact, Intel only provides Windows driver downloads for *32*-bit Windows (nothing for Linux). Personally, I'll be very careful before buying another Intel based or provided motherboard.stagn - Tuesday, October 29, 2013 - link
Yes, where can I buy micro-boards (smaller than mini-ITX) with these. At least 2 gigabit Ethernet and 2 cores. I don't need a Bay Trail chip since I won't be connecting it to a monitor.ShieTar - Tuesday, October 29, 2013 - link
Nope. Out of curiosity, what is you design logic for wanting less than mini-itx, but gual GigE ports? I can think of a few applications, but basically all of them seem to be served as well by a router with an open source firmware.pjkenned - Tuesday, October 29, 2013 - link
Hi Johan,If you want me to run tests on the C2750 let me know. Published these a while back: http://www.servethehome.com/Server-detail/intel-at...
In some synthetic benchmarks the 8-core C2750 is more than 10x faster than the S1260. My general characterization is that it feels more like half of a Intel Xeon E3-1220 V3 in terms of performance. Power consumption for a mITX C2750 platform including 32GB (4x8GB @ 1.35v) cooling and a SSD is 33w with everything including the 4x Intel i354 Gigabit Ethernet ports running.
About the same both on the Supermicro and ASRock units we have tested thus far. Happy to run some benches if you want. Should be getting the fourth platform in the next two weeks so can spare.
Other cool news is that the C2750 runs ESXi fine. If Intel included VT-d it would be a category killer with passthrough. I think you are right that the C2550 or C2730 may be better NAS chips because even on the 12 disk ASRock C2750D4I you can run RAID-Z2 and have plenty of CPU power left over.
JohanAnandtech - Wednesday, October 30, 2013 - link
Interesting comparison. However, I have a few nits to pick. "There is a major myth that the Intel Atom C2750 is many times slower in single threaded performance", but most of your "closer to realworld" benches tell exactly the opposite :-). See the crafty bench for example.milli - Wednesday, October 30, 2013 - link
Johan that's not really a fair thing to say. The Xeon E3 and Opteron turbo to 3.7 and 3.6Ghz respectively. The Atom will turbo to 2.6Ghz. Compared to the Xeon L5520 (2.26Ghz), it's only around 20% behind. That's quite a feat for a cpu core that uses around 2W.With the clocks normalized, it's almost 5x faster than the S1260 it replaces, only 35% behind the Opteron and 2x slower than the E3.
He concludes the following: "performance generally around half of the Intel Xeon E3-1200 V3 series". Seems right to me.
JohanAnandtech - Wednesday, October 30, 2013 - link
1. The clockspeed is part of the µarchitecture trade-offs. So you can not simply normalize clocks.2. The point I was making was about "single threaded performance". It is not even close to being half that of a Xeon... more like one fourth. Not bad, but it is *not* a myth that Atoms are "several times" slower single threaded.
Not saying the Atom C2000 is a dog (it is pretty amazing in fact), just pointing out that the C2000 is still "several times" slower per thread. The myth has not been debunked yet...
pjkenned - Thursday, October 31, 2013 - link
Even there 1/3 of the E3 V3 performance at < 1/4 the platform power consumption on a single threaded application. The difference is that the C2750 is not 1/10th or less the speed like the S1260 was.geekfoo - Friday, November 1, 2013 - link
Patrick, how come you dont run x264 and ffmpeg tests given these are quite popular things to do on servers both home and industrial ?something along the lines of
FOR %%A in ("*.mkv") DO (
ffmpeg.exe -threads 0 -i "%%A" -f mp4 -vcodec libx264 -crf 16 -minrate 900k -maxrate 2100k -bufsize 2100k -refs 3 -preset slow -vprofile high -strict -2 -acodec aac -ac 2 -ab 192k -vf yadif,hqdn3d,unsharp=5:5:0.5,gradfun "%%~nA.MP4"
)
for .bat and easy to edit slightly for bash
for f in *.mkv
do
name=`echo "$f" | sed -e "s/.mp4$//g"`
./ffmpeg -threads 0 -i "$f" -f mp4 -vcodec libx264 -crf 16 -minrate 800k -maxrate 1800k -bufsize 1800k -refs 3 -preset slow -vprofile high -strict -2 -acodec aac -ac 2 -ab 192k -vf yadif,gradfun,hqdn3d,unsharp=5:5:0.5 "$name.MP4"
done
pjkenned - Sunday, November 3, 2013 - link
@geekfoo - will look into adding the bash one. The new development version of the script adds a redis benchmark and sysbench CPU tests. It also now works on CentOS as well as Ubuntu from a clean install:http://forums.servethehome.com/processors-motherbo...
Feel free to give it a try to compare. Will add the above as an open item once I get back to the US this week.
Khato - Tuesday, October 29, 2013 - link
Awwwww, from the title I was all excited to see an actual review of these two server SoCs, not just a feature comparison. Oh well, just have to wait awhile longer for that to come eh?krumme - Tuesday, October 29, 2013 - link
Johan just couldnt wait :). I understand him. From this arm generaion on its starting to become interesting. When was the last time something this happened?Khato - Tuesday, October 29, 2013 - link
If by interesting you mean a performance bump in line with expectations and the ability to address enough memory for it not to be a major limitation, than sure, this generation becomes interesting. Personally I found the first iteration more interesting for the simple fact that it actually beat Intel to the punch and provided a superior solution for the particular niche it was good for. Whereas this time around, well, the C2530 is almost at the bottom of Intel's Avoton lineup - only the one dual core model is below it. The more important point being that for any reasonably threaded workload Intel offers 8 core SKUs which, due to the fixed power consumption per node, greatly increase performance while only bumping up power consumption by ~33%.The next round might be a bit more interesting but I doubt it. Note how the ECX-2000 projected performance increase is pretty much exactly as much as the projected power increase over the ECX-1000? And that's with going from a 40nm process to 28nm. Why exactly should we expect anything more from the next generation?
JohanAnandtech - Wednesday, October 30, 2013 - link
Your reasoning has some good points. However: the higher RAM capacity and CPU performance really allows that new kind of apps run on this thing. Apps that can benefit from the excellent network fabric for example. It will be interesting to see which realworld apps are bottlenecked by the limited CPU power and which ones are bottlenecked by network. Realworld server applications can sometimes be extremely different from the typical benches used. The ECX-2000 will not win any raytracing bench, but Calxeda does have a fighting chance turn as a Content Delivery Server or Drupal based server. All I am saying is, don't judge on raw performance or perf/watt.wsw1982 - Wednesday, October 30, 2013 - link
So, the point is "don't judge on the performance, don't argue on the perf / watt, don't judge on the software, don't judge on the availability, don't judge on the power consumption (no data is given)." But see, "we have 10 g eth"JohanAnandtech - Wednesday, October 30, 2013 - link
I am saying that raw perf/watt does not translate in realworld perf/watt when running realworld server apps. The S1260 looked on par with the ECX-1000 in the benchmark: about the same raw performance (integer crunching benches) and 8.5W vs 6W. The reality was that a S1260 server node needed 15 to 20W (vs 8W) and performed up to twice as slow in real server loads. That is 4-5 (!!) worse perf/watt ratio than expected from integer benches and TDP. Not saying the C2000 will be the same, just illustrating how far real server workloads can be from the other benchmarks.hoboville - Thursday, October 31, 2013 - link
He's saying the ARM chip, like all good RISC chips is good for certain applications like CDN. Real world and on other applications it might not hit its stride.Real world also means that, because the ARM has less RAM and fewer cores, it has the potential to be bogged down in scenarios where there is a higher demand than the cache / RAM / threads can handle. It may be that those requests only require a small amount of network bandwidth, but if the processor can't schedule those tasks...then the CPU is a problem. However, if the CPU is handling a simple decode of a massive file, it can stream the results faster than network bandwidth. Real world perf is subject to a large degree of variability in task type.
JohanAnandtech - Wednesday, October 30, 2013 - link
I understand. We simply have to wait until the ODM solutions on ECX-2000 are validated.geekfoo - Friday, November 1, 2013 - link
wow, whoever they have laying out these moonshot module/sleds wants shooting :)you could easily get 8 or even 16 SoC in that same space today im sure, perhaps even more.
Oh and Johan as you say you cant really compare OC but you should probably go into petty cash and get a few cheap Hardkernel ARM big.LITTLE A7 + A15 Exynos 5410 Octa-Core ODROID-XU's for generic testing etc
http://www.hardkernel.com/renewal_2011/products/pr...
geekfoo - Friday, November 1, 2013 - link
i find these older cortex tests interesting toohttp://www.phoronix.com/scan.php?page=article&...
and PandaBoard bearing an OMAP4430 with 1.0GHz Cortex-A9 MPCore)
http://www.phoronix.com/scan.php?page=article&...
azazel1024 - Tuesday, October 29, 2013 - link
Granted, it is in no way comparable as the multithreaded performance probably falls WAY short, but as an interesting power consumption comparitor, my home server is based on a G1620 Ivy Bridge chip, uATX board (I forget the make, but H77 based, I want to say Gigabit?), 8GB of memory (2x4GB 1.25v), a pair of Intel CT gigabit NICs (motherboard port disabled), 60GB SSD for a boot/app disc and a pair of 2TB Samsung F4EG drives in RAID0 and a 80+ Bronze 350w PSU.Whole thing pulls down 21w at idle and 32w with the drives spun up streaming. During heavy computational loads (as a simple test) the whole system pulls down a hair over 50W with both cores at 90+% and the HDD spun up.
I'd personally love to see the system as a whole pull down half those numbers. I really am curious looking, more from a comodity/lower requirement stand point, the newer Atom based Celeron and Pentium chips.
I'd be curious to see if anything exists with at least four SATA 6Gbps ports, RAID0/1/5 support, soldered on Atom based Pentium chip and at least one Intel based GbE port (with the option to stick another CT NIC in the PCI-e slot). Something tells me though that the newest celeron and pentium based atoms don't have support for 4 SATA 6Gbps ports though.
Gondalf - Tuesday, October 29, 2013 - link
At least IMO, Calxeda is sandbagging a little about power consumption. A15 core is very power hungry, expecially over 1.6Ghz and Silvermont core has clearly showed a far lower power consumption under load.Dear Johan i can see a lot of marketing in Calxeda data, its pretty clear that this round is won by Intel with a wide margin.
Maybe we need to wait some SPECpower submissions before any judgment. "Rough estimates" means nothing beacause Calxeda do not gives TDP data but only "average consumption" figures.
JohanAnandtech - Wednesday, October 30, 2013 - link
Last time, Calxeda marketing was simply wrong about their servers being good memcached or CDN servers. But we found that their power numbers were pretty close to reality. So they deserve the benefit of the doubt...We won't know until we unleash our server workloads on real machine, of course.Drumsticks - Tuesday, October 29, 2013 - link
Does calxeda do any custom changes to the A15? Or do they just straight up use what arm effectively gives them?JohanAnandtech - Wednesday, October 30, 2013 - link
They do not alter the core AFAIK, most of the engineering goes to the uncore. And IMHO that is the right way to do it for a company like Calxedatwotwotwo - Tuesday, October 29, 2013 - link
The thing about the preannouncement of the A53/7 and Intel's roadmap is that it's always hard not to wonder how microservers based on the *next* uarch will do. :)DARBYOTHRULL - Tuesday, October 29, 2013 - link
My question is, where can you buy these processors? I wouldn't mind using one for a personal server.JohanAnandtech - Wednesday, October 30, 2013 - link
The problem is that the integrated fabric might be a bit too complicated (= expensive mobo) for your usage model. There are probably better solutions for you on the market.DARBYOTHRULL - Wednesday, October 30, 2013 - link
That is probably true, and thank you for your comment. If only I knew where to find them, it just seems difficult to build a custom Atom system.chavv - Wednesday, October 30, 2013 - link
What OS/apps are these server SOCs supposed to run on?JohanAnandtech - Wednesday, October 30, 2013 - link
Everything that Linaro (dev organization) makes available on Xen, KVM, LinuxTanclearas - Wednesday, October 30, 2013 - link
How is the ECX-2000 "limited to one" DIMM slot, but have a 128-bit memory controller?texadactyl - Saturday, November 2, 2013 - link
The C2550 and C2750 can support up to 64MB RAM (not part of this article). The C2530 and C2730 (depicted in this article) are limited to 32MB RAM. Source: http://ark.intel.com/compare/77977,77982,77980,779... .TommyVolt - Thursday, December 19, 2013 - link
Yes, the new 16 Gigabyte DDR3 DIMMs to upgrade to 64GB RAM with just 4 sockets are available from the company I'M Intelligent Memory, website: www.intelligentmemory.comThese memory modules come as UDIMMs (240 Pin) or SO-DIMMs (204 Pin), with or without ECC. As they are made of just 16 pieces of 8 Gigabit DDR3 chips, the modules are dual-rank and still have total 16GByte. No special form factor, everything looks standard.
Intel recently released a BIOS update for their C2000 Avoton series to support those Intelligent Memory modules with 16GB capacity..
But the modules might also work in Calxeda and other platforms, because:
When I look at the JEDEC DDR3 documentation, a 8 Gigabit DDR3 chip uses the same amount of address lines as a 4 Gigabit chip (A0 to A15). This means there is no hardware-modification required to run the modules. As a logical consequence, such 16GB modules should be working everywhere, in all types of CPUs, as long as the BIOS is programmed correctly to read the modules ID and set the values for the memory-initialization into the memory controller inside the SOC.
brentpresley - Saturday, November 2, 2013 - link
For those of us that actually run hosting companies and datacenters, these servers are actually worthless, for several reasons.1) x86 is still king in the hosting arena. Everything else, I would never risk Enterprise customers with SLAs of 99.999% uptime on.
2) Underpowered. A nice Ivy Bridge system may pull more power, but it will handle proportionally more traffic.
3) Licensing problems - most Enterprise grade Cloud OS's are licensed out by the Socket, not the core. Therefore your best bang for the buck is something like an Ivy Bridge E or high-end AMD Opteron. Everything else is just a waste of licensing $$$.
Get me Haswell in an HP Moonshot form factor with mSATA SSDs, and then I might be interested. Until then, these are just toys.