ARM based Opterons incoming!

[Phynaz]

http://techreport.com/news/25977/amd-reveals-arm-based-opteron-a1100-series

"Sampling in a few weeks"

 

[gdansk]

That's good news. Any other Cortex A57 sampling soon?

 

[Abwx]

http://techreport.com/news/25977/amd-reveals-arm-based-opteron-a1100-series

"Sampling in a few weeks"

Since it s supposed to be your domain of competence what is your take about thoses products that have no real equivalent in the market if any market at all for the time being.?

 

[Phynaz]

I can't really talk about unreleased server products.

For example, are these cores Big.little?

 

[Abwx]

AMD's first ARM-based server chip is set to start sampling in "a few weeks." The news comes from Andrew Feldman, GM of the company's Server Business Unit

http://techreport.com/news/25977/amd-reveals-arm-based-opteron-a1100-series

The AMD Opteron A1100 Series processors support:

• 4 or 8 core ARM Cortex™-A57 processors
• Up to 4 MB of shared L2 and 8 MB of shared L3 cache
• Configurable dual DDR3 or DDR4 memory channels with ECC at up to 1866 MT/second
• Up to 4 SODIMM, UDIMM or RDIMMs
• 8 lanes of PCI-Express® Gen 3 I/O
• 8 Serial ATA 3 ports
• 2 10 Gigabit Ethernet ports
• ARM TrustZone® technology for enhanced security
• Crypto and data compression co-processors

The AMD Opteron A-Series development kit is packaged in a Micro-ATX form factor and includes:

• An AMD Opteron A1100 Series processor
• 4 Registered DIMM slots for up to 128GB of DDR3 DRAM
• PCI Express® connectors configurable as a single x8 or dual x4 ports
• 8 Serial-ATA connectors
• Compatibility with standard power supplies
• Ability to be used stand-alone or mounted in standard rack-mount chassis
• Standard UEFI boot environment
• Linux environment based on Fedora, which provides developers with a rich set of tools and applications
  • Standard Linux GNU tool chain, including cross-development version
  • Platform device drivers
  • Apache web server, MySQL database engine, and PHP scripting language for developing robust web serving applications
  • Java 7 and Java 8 versions to provide developers to work in a 64-bit ARM environment

http://ir.amd.com/phoenix.zhtml?c=74093&p=irol-newsArticle_Print&ID=1894373&highlight=

 

[Abwx]

I can't really talk about unreleased server products.

For example, are these cores Big.little?

Look data centers dedicated , so low power and effort
put on I/O as suggested by the specs.

 

[ShintaiDK]

Looks to be a Q3 release if they first to sample in a few weeks?

 

[Khato]

Looks to be a Q3 release if they first to sample in a few weeks?

The Anandtech coverage wasn't even that optimistic.

AMD will be making a reference board available to interested parties starting in March, with server and OEM announcements to come in Q4 of this year.

 

[Phynaz]

what is your take about thoses products that have no real equivalent in the market if any market at all for the time being.?

Calxeda.

 

[WaitingForNehalem]

Calxeda.

Ouch. AMD could actually pull this off though as they have the experience and expertise.

 

[Shivansps]

Not sure, Calxeda had experience and died....

I dont like the estimated TDP either.

 

[Ajay]

I dont like the estimated TDP either.

For 8 A57 cores @ 2+ GHz and 8 MB of L3$?
I think that the performance/watt per core is pretty good; if it pans out as listed for the eight core SoC.


Hmm, "Open Data Center" instead of Freedom Fabric. AMD seems to be aiming for Facebook, Google, Yahoo and like minded internet businesses; not any sort of traditional enterprise (which makes sense). Still, they need to grab nearly all the ARM server marketshare to make this a viable business - and AMD needs to grow that marketshare. I have a feeling that real margins will have to be low for AMD to grab that kind of business - these are the cheap seats of the server world.

 

[jpiniero]

They did say Seattle was going to be pretty vanilla... it probably didn't cost that much, and gives them a product out there.

 

[liahos1]

does this compete against avoton or avoton successor?

 

[Soulkeeper]

If i'm not mistaken calxeda failed to be competitive on a perf/watt basis.
Hopefully AMD can, but i'm skeptical

 

[jpiniero]

does this compete against avoton or avoton successor?

Yeah.

If i'm not mistaken calxeda failed to be competitive on a perf/watt basis.

This does look like it is a more complete product than what Calxeda had though.

 

[Exophase]

Not sure, Calxeda had experience and died....

Experience doing what? They were started in 2008 and their first product was the Cortex-A9 product in 2011. They never had a successful server background.

Nothing Calxeda actually released was ever going to be successful, it was all proof of concept/brand building work that was too expensive to sustain.

 

[IlllI]

is this the result of the sseamicro purchase a few years back? if not, i'm wondering what ever came from that acquisition?

 

[Khato]

does this compete against avoton or avoton successor?

I'm not certain where Anandtech got the numbers in their table from as I haven't poured over AMD's PR materials, but they have a SPECint_rate Estimate for the A1100 of 80 with that estimated 25W TDP. Whereas according to Intel the C2750 gets a SPECint_rate_base of 106 with its 20W TDP.

Who knows where performance metrics will actually stand as I've long since given up on AMD providing accurate information in that respect when it comes to future products. The I/O portions certainly look promising, but even they aren't vastly superior to Avoton: Total PCI-E bandwidth is the same so long as the A1100 is driving Gen3.0 devices, which combined with having 4 controllers instead of 2 likely gives Avoton the edge there. On integrated ethernet A1100 wins out with 2x 10Gbe compared to Avoton's 4x 2.5Gbe (I doubt the lesser flexibility matters in the least.) Same is the case for SATA ports with A1100 having 8x SATA3 compared to Avoton's 2x SATA3 and 4x SATA2. And lastly A1100 supports twice as much memory as Avoton at 128GB instead of 64GB.

Real question with all of the above, of course, is what A1100's actual competition from Intel is going to be? If it's still just Avoton there's a chance it might find a niche. But if it's going up against Denverton?

 

[erunion]

Real question with all of the above, of course, is what A1100's actual competition from Intel is going to be? If it's still just Avoton there's a chance it might find a niche. But if it's going up against Denverton?

I don't think we know the exact launch of time frame, At least I don't. But it looks like A1100 will arrive about 1 year after Avaton.

Intel has two server SOCs vaguely due for 2014.

 

[Khato]

I don't think we know the exact launch of time frame, At least I don't. But it looks like A1100 will arrive about 1 year after Avaton.

Intel has two server SOCs vaguely due for 2014.

Yup, which could well mean that Denverton based systems would be showing up in Q4 just like A1100. It's pretty much a given that Denverton will increase efficiency over Avoton by a marked amount (I believe Intel said something like 30% for HSW -> BDW?) but I'm not aware of even the slightest hint about what the rest of the Denverton SoC will look like.

Regardless, here's to hoping that AMD PR didn't give Anandtech an optimistic projection of performance - we don't want Intel to beat up on A57 too badly, right?

 

[erunion]

Yup, which could well mean that Denverton based systems would be showing up in Q4 just like A1100. It's pretty much a given that Denverton will increase efficiency over Avoton by a marked amount (I believe Intel said something like 30% for HSW -> BDW?) but I'm not aware of even the slightest hint about what the rest of the Denverton SoC will look like.

Regardless, here's to hoping that AMD PR didn't give Anandtech an optimistic projection of performance - we don't want Intel to beat up on A57 too badly, right?

Intel is just the beginning of AMDs worries. Intel has a big head start and a strong road map. It will be difficult to pull marketshare from them. But then whatever marketshare ARM does manage to get will be fiercely fought over by the ARM vendors.

AMD will be launching a stock ARM core on GF 28nm around the same time TSMC has it on 20nm.

You could say AMD has an up hill battle.

 

[mrmt]

Just to put things in perspective, you can today build a microserver with both higher CPU performance and smaller power consumption with Avoton. By Q414 we'll have Denverton with 50% more cores with less frequency or same number of cores with 30% less power consumption. As Khato pointed out, it doesn't have necessarily an edge on the I/O part. The only advantage they have over Avoton is that they can address double the amount of RAM, but that isn't helping them much in traditional server markets, I wouldn't bet that it would on microservers.

This product seems DOA to me. It appears to be a product to rather test the assumptions in the business case and get a market foothold ASAP than something that will actually make money for AMD.

On a side note, why didn't they compare the ARM chip with Puma and instead went for Jaguar?

 

[jpiniero]

This product seems DOA to me. It appears to be a product to rather test the assumptions in the business case and get a market foothold ASAP than something that will actually make money for AMD.

Oh, this isn't going to make AMD much money. But it could be compelling if the price is right. A couple more ARM manufacturers jump in on this and you could have an old fashioned price war.

 

[ShintaiDK]

Oh, this isn't going to make AMD much money. But it could be compelling if the price is right. A couple more ARM manufacturers jump in on this and you could have an old fashioned price war.

Price war in a tiny halfdead niche segment?

Calxeda didnt go bankrupt due to a growing profitable segment.

Who else is actually trying this segment, besides AMD and Intel?