Tuesday, June 18th 2013

AMD Unveils Server Strategy and Roadmap

AMD today publicly disclosed its strategy and roadmap to recapture market share in enterprise and data center servers by unveiling innovative products that address key technologies and meet the requirements of the fastest-growing data center and cloud computing workloads.

Additionally, AMD revealed details of its 2014 server portfolio including best-in-class Accelerated Processing Units (APUs), two- and four-socket CPUs, and details on what it expects to be the industry's premier ARM server processor. This is on the heels of announcing the general availability of the AMD Opteron X-Series processor, code named "Kyoto," which dominates the small-core server market on every performance benchmark. These forthcoming AMD Opteron processors bring important innovations to the rapidly changing compute market, including integrated CPU and GPU compute (APU); high core-count ARM servers for high-density compute in the data center; and substantial improvements in compute per-watt per-dollar and total cost of ownership.

"Our strategy is to differentiate ourselves by using our unique IP to build server processors that are particularly well matched to a target workload and thereby drive down the total cost of owning servers. This strategy unfolds across both the enterprise and data centers and includes leveraging our graphics processing capabilities and embracing both x86 and ARM instruction sets," said Andrew Feldman, general manager of the Server Business Unit, AMD. "AMD led the world in the transition to multicore processors and 64-bit computing, and we intend to do it again with our next-generation AMD Opteron families."

In 2014, AMD will set the bar in power-efficient server compute with the industry's premier ARM server CPU. The 64-bit CPU, code named "Seattle," is based on ARM Cortex-A57 cores and is expected to provide category-leading throughput as well as setting the bar in performance-per-watt. AMD will also deliver a best-in-class APU, code named "Berlin." "Berlin" is an x86 CPU and APU, based on a new generation of cores named "Steamroller." Designed to double the performance of the recently available "Kyoto" part, "Berlin" will offer extraordinary compute-per-watt that will enable massive rack density. The third processor announced today is code named "Warsaw," AMD's next-generation 2P/4P offering. It is optimized to handle the heavily virtualized workloads found in enterprise environments including the more complex compute needs of data analytics, xSQL and traditional databases. "Warsaw" will provide significantly improved performance-per-watt over today's AMD Opteron 6300 family.

Seattle
"Seattle" will be the industry's only 64-bit ARM-based server SoC from a proven server processor supplier. "Seattle" is an 8- and then 16-core CPU based on the ARM Cortex-A57 core and is expected to run at or greater than 2 GHz. The "Seattle" processor is expected to offer 2-4X the performance of AMD's recently announced AMD Opteron X-Series processor with significant improvement in compute-per-watt. It will deliver 128 GB DRAM support, extensive offload engines for better power efficiency and reduced CPU loading, server caliber encryption, and compression and legacy networking including integrated 10GbE. It will be the first processor from AMD to integrate AMD's advanced Freedom Fabric for dense compute systems directly onto the chip. AMD plans to sample "Seattle" in the first quarter of 2014 with production in the second half of the year.

Berlin
"Berlin" is an x86-based processor that will be available both as a CPU and APU. The processor boasts four next-generation "Steamroller" cores and will offer almost 8X the gigaflops per-watt compared to current AMD Opteron 6386SE processor. It will be the first server APU built on AMD's revolutionary Heterogeneous System Architecture (HSA), which enables uniform memory access for the CPU and GPU and makes programming as easy as C++. "Berlin" will offer extraordinary compute per-watt that enables massive rack density. It is expected to be available in the first half of 2014.

Warsaw
"Warsaw" is an enterprise server CPU optimized to deliver unparalleled performance and total cost of ownership for two- and four-socket servers. Designed for enterprise workloads, it will offer improved performance-per-watt, which drives down the cost of owning a "Warsaw"-based server while enabling seamless migration from the AMD Opteron 6300 Series family. It is a fully compatible socket with identical software certifications, making it ideal for the AMD Open 3.0 Server -- the industry's most cost effective Open Compute platform. It is expected to be available in the first quarter of 2014.
Add your own comment

29 Comments on AMD Unveils Server Strategy and Roadmap

#1
cdawall
where the hell are my stars
steamroller 1st half of 2014...hmmm, compatable with the existing server socket....hmmmm.
Posted on Reply
#2
james888
1. Steamroller cores!
2. They improved everything? I want to see some reviews.
3. HSA is here. I wonder how useful it will actually be and want to see some benchmarks and reviews.
Posted on Reply
#3
seronx
No, 2P/4P Steamroller means no Steamroller FX CPUs.
Posted on Reply
#4
cdawall
where the hell are my stars
by: seronx
No, 2P/4P Steamroller means no Steamroller FX CPUs.
Or they are trying to burn through as many piledriver cores as they can.
Posted on Reply
#5
esrever
by: seronx
No, 2P/4P Steamroller means no Steamroller FX CPUs.
my guess is waiting for ddr4.
Posted on Reply
#6
Vinska
>still using 32nm in 2H2014
My heart is sobbing. :(
Indeed, what AMD does with the 32nm process power-efficiency-wise is simply magic. But they should really be moving on to a smaller process. Why? Oh why do You keep on using the old 32nm? (well, we all do know why, but still...)
Posted on Reply
#7
repman244
by: Vinska
Oh why do You keep on using the old 32nm?
Because GloFo doesn't have anything else? Because they don't have as much money as Intel? Because the process doesn't mean anything if the design is bad?
And you should never compare the processes between different manufacturers...
Posted on Reply
#8
Vinska
by: repman244
Because GloFo doesn't have anything else? Because they don't have as much money as Intel? Because the process doesn't mean anything if the design is bad?
And you should never compare the processes between different manufacturers...
*points at*
by: Vinska
(well, we all do know why, but still...)
i.e. it was rhetorical, man :)
Posted on Reply
#9
vinibali
The "Seattle" processor is expected to offer 2-4X the performance of AMD's recently announced AMD Opteron X-Series processor with significant improvement in compute-per-watt.
the Jaguar core has the same size,consumption and powerratio like the best ARM cores like A9 and A12. how the **** is A57 much better?
Posted on Reply
#10
Ikaruga
People are talking about 32nm like it's something bad, while good old Sandy 2600k is still doing just fine against the latest AMD chips ;)
Posted on Reply
#11
Vinska
by: Ikaruga
People are talking about 32nm like it's something bad, while good old Sandy 2600k is still doing just fine against the latest AMD chips ;)
You mean, the Intel's last 32nm chips against the chips that were AMDs first 32nm ones?
Posted on Reply
#12
Ikaruga
by: Vinska
You mean, the Intel's last 32nm chips against the chips that were AMDs first 32nm ones?
The 2600K is still doing "OK" against the latest 32nm AMD ones as well (let alone the old ones). They might suck at some newly optimized heavily multithreaded loads perhaps, but those games/apps are still the exceptions tbh.

Well things will shift to favor 8+ cores more rapidly in the future ofc, but we are still not there...yet (if you take the average of all the games and apps a desktop user might run)
Posted on Reply
#13
Vinska
@Ikagura Sorry, just a dreadful habit which surfaces up after spending too much on /g/
Posted on Reply
#14
drdeathx
by: Ikaruga
The 2600K is still doing "OK" against the latest 32nm AMD ones as well (let alone the old ones). They might suck at some newly optimized heavily multithreaded loads perhaps, but those games/apps are still the exceptions tbh.

Well things will shift to favor 8+ cores more rapidly in the future ofc, but we are still not there...yet (if you take the average of all the games and apps a desktop user might run)
2600K is dead.
Posted on Reply
#15
Ikaruga
by: Vinska
@Ikagura Sorry, just a dreadful habit which surfaces up after spending too much on /g/
Why would you say sorry?:confused: You said nothing wrong.:toast:
I just pointed out that the old 2600k is a 32nm cpu, while it doesn't really need much power (in AMD terms), and it still performs very well both on stock and especially when it's overclocked.

by: drdeathx
2600K is dead.
Thank you for your insightful input...... again....much appreciated. :shadedshu
Posted on Reply
#16
drdeathx
by: Ikaruga
Why would you say sorry?:confused: You said nothing wrong.:toast:
I just pointed out that the old 2600k is a 32nm cpu, while it doesn't really need much power (in AMD terms), and it still performs very well both on stock and especially when it's overclocked.


Thank you for your insightful input...... again....much appreciated. :shadedshu
Anytime but why you bring it up? New Piledriver FX processors are much better anyway..... Sandy is gone and Ivy is otw out. Haswell is here!:slap:
Posted on Reply
#17
james888
by: drdeathx
Anytime but why you bring it up? New Piledriver FX processors are much better anyway..... Sandy is gone and Ivy is otw out. Haswell is here!:slap:
If its so dead I'll take that 3770k off your hands :rolleyes: ;)
Posted on Reply
#18
acerace
by: james888
If its so dead I'll take that 3770k off your hands :rolleyes: ;)
Just a mindless troll sometimes whose goal is to cause a flame war. Yeah, very typical. :)
Posted on Reply
#19
Ikaruga
by: drdeathx
Anytime but why you bring it up? New Piledriver FX processors are much better anyway..... Sandy is gone and Ivy is otw out. Haswell is here!:slap:
I use a lot of applications which are not really multi-threaded and where single threaded performance matters a lot. For example, I like to play with Dolphin-Emulator which can't and would not take benefit from more than 2.5 cores, or an other example: I run servers like source or doom(chillax),etc which are all also single threaded (network load can be bound to an other core, but that's almost nothing)... so the "dead" 2600K is still a much better option for me than any(!) AMD cpu out there (But I wish they weren't tho, I like competition tbh)
Posted on Reply
#20
drdeathx
by: james888
If its so dead I'll take that 3770k off your hands :rolleyes: ;)
LOL, its gone. I havent updated my sig yet....... I am on the 3930K 12 core LOL:nutkick:

by: acerace
Just a mindless troll sometimes whose goal is to cause a flame war. Yeah, very typical. :)
LOL a wise ass. Me and James are good friends and he isn't trolling if you read my sig....... Bet you feel silly now trolling?:):laugh::)
Posted on Reply
#21
acerace
by: drdeathx
LOL, its gone. I havent updated my sig yet....... I am on the 3930K 12 core LOL:nutkick:



LOL a wise ass. Me and James are good friends and he isn't trolling if you read my sig....... Bet you feel silly now trolling?:):laugh::)
Yes, I noticed that. But, isn't it your thing sometimes love to make people angry? Yes, I know you. :toast:
Posted on Reply
#22
Vinska
by: drdeathx
LOL, its gone. I havent updated my sig yet....... I am on the 3930K 12 core LOL:nutkick:
Squeeze me? Dude, it's a hexacore (six cores).
Posted on Reply
#23
DaJMasta
by: Vinska
>still using 32nm in 2H2014
My heart is sobbing. :(
Indeed, what AMD does with the 32nm process power-efficiency-wise is simply magic. But they should really be moving on to a smaller process. Why? Oh why do You keep on using the old 32nm? (well, we all do know why, but still...)
Agree 100%. Even if it's a foundry limitation... you're getting trounced in the performance/watt because of the process. It would be so much closer to competition with intel and the chips themselves would cost less if you could move to a smaller node. APUs and whatnot with less emphasis on x86 single core performance is just fine if you have the TDP to back it up, right now it looks like they're just trying to market their way out of poorer performing parts even if the move towards more GPU compute is the right choice in the long run.

If you get your TDP down the market opens up to you AMD (mobile, workstation, small form factor). You need a smaller process node and you need it before 2015. :ohwell:


...and don't tell me you can't when you've been below 32nm on your GPUs for quite some time now.
Posted on Reply
#24
de.das.dude
Pro Indian Modder
a lot of stuff being talked about. Since most things launched are already in research atm, this means that the new staff is working on these... Hmm. A LOt of server and arm related things. Hope they are sure. Looks to me as if amd is about to close the AMx lineup, continue with the FM sockets and specialize in server and arm. Atleast steam roller will be supported on my socket XD
Posted on Reply
#25
Vinska
Well, the roadmap does show moving to 28nm. That's an improvement, yes. But they'd better jump to 22 or 20 nm by 2015, yo!
Posted on Reply
Add your own comment