AMD Cancels 10-Core, 20-Core Server Processors. Prolong Lifespan of Current Opteron Server Platform
[font=Arial][left]Advanced Micro Devices has cancelled its multi-core Opteron server chips code-named Sepang (up to 10 cores) and Terramar (up to 20 cores) that were supposed to be launched in 2012. Moreover, the company decided to keep the existing server platform for its 2012 - 2013 server refreshes, which means that AMD-based machines will not support numerous new features.[/left][/font]

[font=Arial][left]In the 2012 - 2013 timeframe AMD plans to release Opteron code-named Abu Dhabi microprocessor with up to 16 Piledriver cores for 4-socket servers, Seoul central processing unit with up to 8 Piledriver cores for 2-socket servers as well as Delhi chip with up to 8 Piledriver cores for 1-socket servers. All the new CPUs will fit into existing socket G34, socket C32 and socket AM3+ and will therefore bring no improvements on the platform level. The new processors will be made using 32nm SOI process technology at Globalfoundries.[/left][/font]

[left][font="Arial"]...[/font][/left]
[font=Arial][left]The reason for a major change of the Opteron roadmap was not revealed, but AMD indicated claimed that the new lineup "delivers better performance", which is a rather questionable claim. [/left][/font]
[font=Arial][left]
[/left][/font]
[font=Arial][left]
[/left][/font]Story at [url="http://www.xbitlabs.com/news/cpu/display/20120202124109_AMD_Cancels_10_Core_20_Core_Server_Processors.html"]XBitLabs[/url]
[font=Arial]
Advanced Micro Devices has cancelled its multi-core Opteron server chips code-named Sepang (up to 10 cores) and Terramar (up to 20 cores) that were supposed to be launched in 2012. Moreover, the company decided to keep the existing server platform for its 2012 - 2013 server refreshes, which means that AMD-based machines will not support numerous new features.
[/font]



[font=Arial]
In the 2012 - 2013 timeframe AMD plans to release Opteron code-named Abu Dhabi microprocessor with up to 16 Piledriver cores for 4-socket servers, Seoul central processing unit with up to 8 Piledriver cores for 2-socket servers as well as Delhi chip with up to 8 Piledriver cores for 1-socket servers. All the new CPUs will fit into existing socket G34, socket C32 and socket AM3+ and will therefore bring no improvements on the platform level. The new processors will be made using 32nm SOI process technology at Globalfoundries.
[/font]



...


[font=Arial]
The reason for a major change of the Opteron roadmap was not revealed, but AMD indicated claimed that the new lineup "delivers better performance", which is a rather questionable claim.
[/font]

[font=Arial]


[/font]

[font=Arial]


[/font]Story at XBitLabs



ImageImage


Image

Join the NVidia Forum Team - Please Help Medical Research - Folding@home


#1
Posted 02/02/2012 11:02 PM   
I hope AMD comes through and can give intel some comp.
I hope AMD comes through and can give intel some comp.

#2
Posted 02/03/2012 09:03 AM   
[quote name='Xavier Zepherios' date='02 February 2012 - 05:02 PM' timestamp='1328223720' post='1364355']
[font="Arial"][left]The reason for a major change of the Opteron roadmap was not revealed, but AMD indicated claimed that the new lineup "delivers better performance", which is a rather questionable claim.[/left][/font][/quote]

The reason is pretty obvious. AMD makes it's money off of planned upgrades and socket consistency. The biggest reason to buy AMD is rolling upgrades for lower costs than replacing an entire Intel server setup/datacenter --drop in replacements for G34 are a better idea than a new socket for AMD's customers.
[quote name='Xavier Zepherios' date='02 February 2012 - 05:02 PM' timestamp='1328223720' post='1364355']

The reason for a major change of the Opteron roadmap was not revealed, but AMD indicated claimed that the new lineup "delivers better performance", which is a rather questionable claim.




The reason is pretty obvious. AMD makes it's money off of planned upgrades and socket consistency. The biggest reason to buy AMD is rolling upgrades for lower costs than replacing an entire Intel server setup/datacenter --drop in replacements for G34 are a better idea than a new socket for AMD's customers.

Done.

#3
Posted 02/03/2012 06:49 PM   
[quote]...For the past seven years, Intel and AMD have emphasized multi-core CPUs as the answer to scaling system performance, but there are multiple reasons to think the trend towards rising core counts is largely over. First and foremost, there’s the fact that adding more CPU cores never results in perfect scaling. In any parallelized program, performance is ultimately limited by the amount of serial code (code that can only be executed on one processor). This is known as Amdahl’s law. Other factors, such as the difficulty of maintaining concurrency across a large number of cores, also limit the practical scaling of multi-core solutions...[/quote]

[i][url="http://www.extremetech.com/computing/116561-the-death-of-cpu-scaling-from-one-core-to-many-and-why-were-still-stuck?print"][i]The death of CPU scaling: From one core to many — and why we’re still stuck[/i][/url][/i]
...For the past seven years, Intel and AMD have emphasized multi-core CPUs as the answer to scaling system performance, but there are multiple reasons to think the trend towards rising core counts is largely over. First and foremost, there’s the fact that adding more CPU cores never results in perfect scaling. In any parallelized program, performance is ultimately limited by the amount of serial code (code that can only be executed on one processor). This is known as Amdahl’s law. Other factors, such as the difficulty of maintaining concurrency across a large number of cores, also limit the practical scaling of multi-core solutions...




The death of CPU scaling: From one core to many — and why we’re still stuck

Intel Siler DX79SI Desktop Extreme | Intel Core i7-3820 Sandy Bridge-Extreme | DangerDen M6 and Koolance MVR-40s w/Black Ice Stealths | 32 GB Mushkin PC3-12800LV | NVIDIA GTX 660 Ti SLI | PNY GTX 470 | 24 GB RAMDisk (C:\Temp\Temp) | 120 GB Intel Cherryville SSDs (OS and UserData)| 530 GB Western Digital VelociRaptor SATA 2 RAID0 (C:\Games\) | 60 GB G2 SSDs (XP Pro and Linux) | 3 TB Western Digital USB-3 MyBook (Archive) | LG BP40NS20 USB ODD | LG IPS236 Monitor | LogiTech X-530 Speakers | Plantronics GameCom 780 Headphones | Cooler Master UCP 1100 | Cooler Master HAF XB | Windows 7 Pro x64 SP1

Stock is Extreme now

#4
Posted 02/03/2012 10:26 PM   
[quote name='jaafaman' date='03 February 2012 - 04:26 PM' timestamp='1328307964' post='1364880']
[i][url="http://www.extremetech.com/computing/116561-the-death-of-cpu-scaling-from-one-core-to-many-and-why-were-still-stuck?print"][i]The death of CPU scaling: From one core to many — and why we’re still stuck[/i][/url][/i]
[/quote]

I don't agree with everything in this article, but it does have a point. You cannot indefinitely add cores, but this article oversimplifies and makes a lot of false assumptions. Interlargos' problems with many OSes are specifically caused by the need to rewrite the the threading from nearly the ground up --not from shoving too many cores onto the die.


That said, there is a looming barrier we are fast approaching and odds are we're going to hit it within my lifetime. When we do hit the absolute computing threshold the only way to improve performance with binary computing will be to reduce transistor count and the number of logic pathways in order to increase electrical throughput. Instruction sets will also be in the way and we'll need to move into load-store RISC or even ZISC arches and use real time operating systems for everything.

In the back rooms of foundries and in the basements of colleges these things are already being discussed, but we live in a capitalist system. There is more money in incremental improvements [i]now[/i] and [i]now[/i] is all most suits or investors care about. It's hard to justify research for something you'll need 40 years, 10 years or even 5 years in the future when the main thing most people care about is the last quarter versus LY.
[quote name='jaafaman' date='03 February 2012 - 04:26 PM' timestamp='1328307964' post='1364880']

The death of CPU scaling: From one core to many — and why we’re still stuck





I don't agree with everything in this article, but it does have a point. You cannot indefinitely add cores, but this article oversimplifies and makes a lot of false assumptions. Interlargos' problems with many OSes are specifically caused by the need to rewrite the the threading from nearly the ground up --not from shoving too many cores onto the die.





That said, there is a looming barrier we are fast approaching and odds are we're going to hit it within my lifetime. When we do hit the absolute computing threshold the only way to improve performance with binary computing will be to reduce transistor count and the number of logic pathways in order to increase electrical throughput. Instruction sets will also be in the way and we'll need to move into load-store RISC or even ZISC arches and use real time operating systems for everything.



In the back rooms of foundries and in the basements of colleges these things are already being discussed, but we live in a capitalist system. There is more money in incremental improvements now and now is all most suits or investors care about. It's hard to justify research for something you'll need 40 years, 10 years or even 5 years in the future when the main thing most people care about is the last quarter versus LY.

Done.

#5
Posted 02/03/2012 10:49 PM   
[quote name='D1llw33d' date='03 February 2012 - 05:49 PM' timestamp='1328309389' post='1364883']...In the back rooms of foundries and in the basements of colleges these things are already being discussed, but we live in a capitalist system. There is more money in incremental improvements [i]now[/i] and [i]now[/i] is all most suits or investors care about. It's hard to justify research for something you'll need 40 years, 10 years or even 5 years in the future when the main thing most people care about is the last quarter versus LY.[/quote]
Another story for another thread, but it is my firm belief that the biggest mistake a company can make is to ask for an IPO so the [i][b]investors[/b][/i] can control the company. For many of the reasons outlined so theatrically in the last 30 minutes or so of "[i]Other People's Money[/i]"...
[quote name='D1llw33d' date='03 February 2012 - 05:49 PM' timestamp='1328309389' post='1364883']...In the back rooms of foundries and in the basements of colleges these things are already being discussed, but we live in a capitalist system. There is more money in incremental improvements now and now is all most suits or investors care about. It's hard to justify research for something you'll need 40 years, 10 years or even 5 years in the future when the main thing most people care about is the last quarter versus LY.

Another story for another thread, but it is my firm belief that the biggest mistake a company can make is to ask for an IPO so the investors can control the company. For many of the reasons outlined so theatrically in the last 30 minutes or so of "Other People's Money"...

Intel Siler DX79SI Desktop Extreme | Intel Core i7-3820 Sandy Bridge-Extreme | DangerDen M6 and Koolance MVR-40s w/Black Ice Stealths | 32 GB Mushkin PC3-12800LV | NVIDIA GTX 660 Ti SLI | PNY GTX 470 | 24 GB RAMDisk (C:\Temp\Temp) | 120 GB Intel Cherryville SSDs (OS and UserData)| 530 GB Western Digital VelociRaptor SATA 2 RAID0 (C:\Games\) | 60 GB G2 SSDs (XP Pro and Linux) | 3 TB Western Digital USB-3 MyBook (Archive) | LG BP40NS20 USB ODD | LG IPS236 Monitor | LogiTech X-530 Speakers | Plantronics GameCom 780 Headphones | Cooler Master UCP 1100 | Cooler Master HAF XB | Windows 7 Pro x64 SP1

Stock is Extreme now

#6
Posted 02/04/2012 10:06 AM   
Scroll To Top