AMD Opteron 6100 Series review
We thought Intel was bad with its codenames, but AMD now throws a flurry of them into the ring. The Opteron 6100 “Magny-Cours” is the first in AMD’s “Maranello” 6000 platform series. The platform and Socket G34 are fully forward-compatible with the next-generation “Interlagos” 12- and 16-core processors.
The next generation is due to be launched in 2011 and will introduce AMD’s “Bulldozer” technology. This is expected to deliver AMD’s answer to Intel’s Hyper-Threading, but it may also implement hardware encryption instructions to counter Intel’s AES-NI, which is already present in the Xeon 5600 family. The soon-to-be-launched 4000 platform series (San Merino and Adelaide) will introduce the “Lisbon” quad- and six-core processors. These target low-cost 1P and 2P server applications and will also be forward-compatible with the next-generation “Valencia” six- and eight-core processors.
AMD’s future plans look impressive, but if it’s going to make any serious headway against Intel it needs to get a lot of server vendors on the bandwagon. In the volume markets it has Dell, HP, Supermicro and Acer, with the latter aiming to offer Opteron 6100 servers under the Gateway brand it acquired a couple of years ago. We’ve always found our requests for AMD server samples to be met by indifference, with most vendors preferring to focus on getting Intel servers into our lab. Supermicro is the one exception and it tied in with the 6100 launch by announcing a whole slew of dual- and quad-socket Opteron 6100 motherboards and systems.
Boston, a UK distributor of Supermicro products, is clearly impressed by the versatility of being able use the same processors across 1P, 2P and 4P servers. It cites the 7U 480-core TwinBlade virtualised server, based on Opteron 6100 processors, as an example.
However, the real question is this: how fast is the Opteron 6100?
We benchmarked the new Opteron 6100 in the lab alongside systems equipped with Xeon X6550, Xeon X7550, Xeon X5650 and Xeon E5630 processors. We used FlamMap, POV-Ray and CineBench, and also used SiSoftware Sandra’s Cryptography test to highlight the benefits of Intel’s AES-NI instruction set implemented in the Xeon 5600. Note that although the graphs below name the processors, they were in different rigs and thus used different amounts of memory. This information is shown at the foot of each graph.
FlamMap is a fire-behaviour mapping and analysis program that computes potential fire behaviour. It demonstrates the benefits of multicore technology and is very compute-intensive, with a heavy number of floating point operations.
The Opteron 6174 came fourth in this test and was notably beaten by the X5650 rig, which was over 36% faster. The four Xeon X7550 processors trounced everything else, with their 48 logical cores giving a speed increase of 45% over the 6174.
CineBench is a dedicated benchmark based on the 3D software Cinema 4D, which performs CPU-intensive rendering operations using multiple CPUs. We used the latest version, 11.5, which can measure systems with up to 64 processor threads. The scores are all relative, with the basic message being that bigger is better and, as with FlamMap, involves a huge amount of computation including many floating point operations.
No surprises that the X7550 came top of the class here. Although some way behind, the 6174 was a very creditable second. Raw processor speed as well as core count is the success story with CineBench, so it was no surprise to see the X5650 close on AMD’s heels.