AnandTech posted their analysis of AMD’s 2010/2011 roadmap and competitive position last week; it’s worth a read
Currently, AMD’s six-core Opteron can match the performance of Intel’s quadcore Xeon 5500 at the same clockspeed in some important server applications: OLAP databases, virtualization and web applications. Intel’s best Xeon wins with a significant margin in OLTP, ERP and rendering. A large part of the HPC market is a lost cause: a quadcore Intel Xeon 5570 at 2.93 GHz is about twice as fast as a AMD Opteron 2389 at 2.9 GHz. The fact that we could not find any Opteron 2435 results in LS-Dyna is another indication of what to expect: the 10-20% higher performance in HPC applications will not be a large step forward.
…Intel is going to increase performance by 20-30% per CPU (50% more cores), while AMD’s CPUs will see only marginal increases. So basically, Intel’s performance advantage is going to grow by 20 to 30%, except in HPC workloads where it is already running circles around the competition. Not an enviable position to be in for AMD.
I did have a quibble with this assessment from the start of the article
Compared to the mobile and desktop market, AMD is doing relatively well in the server and HPC market. The early delivery of the six-core Opteron (codenamed Istanbul) enabled Cray to build the fastest supercomputer in the world (at least for Q4 2009). It’s called the the Cray XT5-HE “Jaguar” with 224162 cores, good for almost 1.76 million GFlops.
Yes, AMD has two of the best marquees in the world of HPC: the number 1 and 3 spots on the Top500 list. These machines are both Crays, the result of an alliance formed at the start of the XT line of supers when AMD was the player to beat. That player is now Intel, of course, and it is reasonable to anticipate that Cray, being an über-pragmatic company, is going to incorporate today’s dominant chips in its next-gen supers.
To AnandTech’s credit, they follow up that paragraph with this line
Nevertheless, AMD’s position in the server and HPC market is seriously threatened.
…and the rest of the article has more of that tone anyway. It is helpful to review the Top500 on this point: Intel’s share of the Top500 has grown to just over 80% of the list (by machine count). By contrast, AMD’s share has dropped to 8.4% from 22.6% in November of 2006. It’s too early to count them out, but AMD’s presence on the Top500 is starting to look like a temporary excursion caused by technological missteps which Intel has since corrected.
Johan did feel more optimistic about some of AMD’s 2010 offerings in 2010 on 2-socket servers
When it comes to expensive 2-socket servers, AMD positioning is cunning. In the midrange we will find servers with sixteen Opteron cores (2-socket x 2-quad-core die per socket) offering 8 memory channels and 24 DIMM slots. Performance will probably be “close enough” to the Westmere EP servers , which can only offer six memory channels and 18 DIMM slots. The extra amount of memory bandwidth might make a dual Opteron 6100 attractive to the HPC folks, while the higher amounts of DIMM slots together with a competitive price may very well convince the virtualization market.
Hat tip to Rich Brueckner for the pointer.
For what it is worth, the #2 machine (Roadrunner) contains around 6K dual-core Opterons in addition to the PowerXCell 8i chips. They may not provide the bulk of the flops, but you couldn’t run applications or do I/O without them.
The Intel advantage lies is stride-1 memory bandwidth per core and the performance of many HPC applications is strongly correlated with stride-1 memory bandwidth. AMD will correct that with Magny-Cours.
Paul – thanks for reminding us of that. I should have included that in my comments.