AMD Demos 6-core Opteron "Istanbul"

[iml3g3nd]

AMD provided a live demonstration of their 6-core Opteron processor code-named “Istanbul.” It is basically a 6-core version of 45nm “Shanghai” Opteron and is pin compatible with the current Socket F boards. AMD seems to be right on track for launch in H2 2009.
As Tech Report has reported, AMD has provided a few demonstrations on Istanbul silicon in action. The first demo was a simple showing of Task Manager on the Windows Server 2008 desktop showing 24 cores in a 4P system. The second demo was on a 2P system with 12 cores running virtualization. The third demo was a Stream benchmark test comparing Shanghai vs Istanbul. The 16-core Shanghai system produced throughput numbers in the range of 25,000 MB/s while the 24-core Istanbul box attained about 42,000 MB/s.

HARDSPELL.COM-AMD Demos 6-core Opteron “Istanbul”

[hatter]

AMD’s top exec talks about its upcoming six-core server CPU.

[

With Nehalem-based Xeons expected to crush everything that AMD’s Opterons offer right now](AnandTech Forums: Technology, Hardware, Software, and Deals), AMD is banking heavily on Istanbul to save it from a complete annihilation. So far this CPU is shaping up decently. AnandTech (AnandTech) and TechReport (AMD puts on the ritz with six-core Opteron demo - The Tech Report - Page 1) have been impressed so far.

Istanbul may or may not be able to make a mark. But there is no doubt that as far little nuggets of information go, this one has some flair:

Last week we had the pleasure of hosting a select number of reporters to brief them on our upcoming “Istanbul” processor and show them some demos as well.

As always, when you are doing a demo on an unreleased product, you’re always on your toes. It’s like teaching a dog a trick; you worry that it won’t want to do it when your friends are over.

Well Istanbul can fetch, sit up, shake your hand, and apparently, play the piano. Quite a feat. The new 6-core version of the AMD Opteron™ processor is scheduled to be available in the second half of this year, and it is everything we had hoped for – and more.

As a city, Istanbul is the only city that sits on two continents, Asia and Europe. As a processor, Istanbul also bridges two worlds, the socket 1207 that has been such a strong platform in the past, and a 6-core Direct Connect architecture, with 12, 24 or 48 cores per server for the future.

Despite putting more cores in the processor, we managed to keep it in the same power and thermal ranges as our existing “Shanghai” processors. And since it fits into the same socket, our OEM customers should be able to bring products to the market quickly. End users will be able to quickly qualify and deploy these servers because the overall platform is the same as what they are using. In today’s challenging economic times, that’s music to the ears of IT departments both near, and as far away as Turkey.

So what did we show? We showed a platform being easily upgraded from Shanghai to Istanbul, some amazing memory throughput courtesy of the new HT Assist feature, and a 4-socket server with all 24 cores being stressed by one of our development programs.

We were impressed by how well the silicon was progressing, when one of the reporters noted “well it should be doing great, it has the same DNA as Shanghai.”

Things are heating up for this year, and we’re excited about where it will be taking us.

AMD at Work

[abbY]

The Istanbul cores are the same as those that can be found in the AMD’s latest Shanghai CPU. But the “uncore” part of Istanbul is more interesting. By now, you have probably heard about AMD’s “HT-assist” technology, a probe or snoop filter. Every time a new cacheline is brought into the L3-cache of for example CPU 1 on the current Shanghai Platform, a broadcast message is sent to all L3-caches of all CPUs, and CPU 1 has to wait until those CPUs answer.
In the case of Istanbul, the CPU will simply check it’s snoop filter in it’s own L3-cache, and if none of the other CPUs have that certain cacheline, it can go ahead. This lowers the latency of bringing in a new cacheline and raises the effective bandwidth.
To better understand this, we combined our own stream benchmarking with the one that AMD presented. All AMD systems are using DDR-2 800.
Stream Triad benchmark
As each Stream thread works on its own data, there is no reason to send out coherency synchronization requests. These requests slow the process of getting new cachelines in the L3 and hence lower effective memory bandwidth. What is interesting is that this will not only benefit the applications that use the HT interconnects a lot for coherency traffic, but also applications like stream which do not need the HT interconnects. Also notice that HT 3.0 does not improve memory bandwidth, as Stream will try to keep its thread data local. Our testing used SUSE SLES 10 SP2 and AMD used Windows 2008. Both OSs are well optimized and NUMA aware.
This means that especially HPC applications, with many threads all working on their own data, will benefit from the higher effective bandwidth. Besides HT assist, AMD has now confirmed to us that the memory controller has been tuned quite a bit. This higher amount of bandwidth will allow the quad Istanbul to stay out of the reach of the dual Nehalem EP Xeons in many HPC applications.
HT assist might also improve the SAP and OLTP scores quite a bit, but for a different reason. SAP and OLTP applications perform a lot of cache coherency syncronization requests, so the snoop filter will substantially lower the average latency of such requests as in some cases:

* the CPU will only wait on one other CPU (instead of waiting for all responses to come back)

* the CPU won't have to wait at all, as the other CPUs don't have this line.

Secondly, this will also lower memory latency, which is a bonus for almost every multi-threaded application.
Lower memory latency, higher bandwidth, lower “cache coherency” latency and more interconnect bandwidth: the improved “uncore” of Istanbul will be vital to close the gap with Nehalem. Much will depend on how quickly Intel introduces its own hexacore 32 nm Xeons, but that probably won’t happen before 2010. Istanbul is shaping up to be a really good alternative for Intel’s quadcore Nehalem. We might see a good fight after all…
Don’t forget to check it.anandtech.com (IT portal) often, as many of our blogposts (for example the VMworld 2009 coverage) are not published on the frontpage of Anandtech.com.

Link->AnandTech