Opteron versus Xeon
If youíre at all a tech enthusiast, you probably know that itís pretty easy to find an Intel Xeon thatís on par with an AMD Opteron, but itís tougher to get a pair of Xeons to compete against a pair of Opterons, and even more difficult to get a quartet of Xeons to match the performance of a quartet of AMD Opterons. This is due to both AMDís HyperTransport technology and the on-core memory controller. HyperTransport allows each of the CPUs to communicate with each other very quickly and the on-core memory controller allows greater bandwidth than available with traditional Northbridge style architectures. One of the misconceptions of a dual core CPU is that all manufacturers implement dual core the same way, resulting in maximum performance. It turns out that this isnít the case.
In AMDís case, the memory controller is tied to each physical processor. The two cores of a multi-core CPU are connected with AMDís ďDirect Connect Technology,Ē marketing talk for having two CPU cores with direct on-die access to the shared on-die memory controller.
In comparison to a dedicated dual Opteron system, a single dual-core Opteron has half the memory bandwidth but less latency. However, thatís the pessimistic way to look at it Ė a single dual core Opteron has the same memory bandwidth as one single-core Opteron, but has twice the computational unitsÖ and while I said a pair of single-core Opterons should outperform a single dual-core Opteron, a pair of dual core Opterons will be even more impressive.
At todayís launch AMD has three platforms, the Opteron x65, the x70, and x75 where the x is 1, 2, or 8 depending on whether or not the chip supports 1, 2, or more CPUs. The x65 is a dual-core 1.8GHz (i.e. two Opteron x44ís), the x70 is a dual-core 2.0 GHz (i.e. two Opteron 246ís), and the x75 is a dual-core 2.2GHz (i.e. two Opteron 248ís). The fastest clocked Opteron is the 252, which runs at 2.6GHz.
Intelís dual-core approach is still built on top of Intelís current design philosophy in which the memory controller is still part of the motherboard rather than the CPU. Likewise, instead of something such as HyperTransport, the interconnect between the two CPUs is also similar to a traditional Xeon architecture meaning that the processors donít have as much bandwidth. Moreover, the current dual core processors from Intel only have a FSB of 800MHz as opposed to the fast 1066MHz FSB of the single-core Pentium 4 Extreme Edition 3.73GHz. Dual-core Pentium processors will require new motherboards and chipsets but in theory, AMD dual-core CPUs should be compatible with todayís chipsets and motherboards.
Talking with system builders however, the biggest difference that is often forgotten is the 16GB memory limit. Both Intel and AMD servers claim to support 16GB of system RAM through the use of 2GB DIMMS. Unfortunately, 2GB DDR-2 DIMMS compatible with Intel servers are exceptionally difficult to find, with few manufacturers even willing to make them. So, the difference in cost for RAM for a 16GB system can be as much as $10,000 per server! Fortunately for Intel, most end-users donít need 16GB of RAM.
For the current dual-core processor launch, AMD is focusing their dual core technology on the high-end with Socket 940 Opterons. From a business perspective this makes sense since early on, it will be the typical Opteron user who will want dual-core CPUs first and many of these new customers will be looking to get two dual-core processors rather than just one. In a way, AMD now has the potential of selling the equivalent of 4 CPU cores per physical system. Not a bad deal for them.
As for the typical consumer, two CPU cores will be enough and so weíll still need to wait for the Athlon64 dual-core processors to hit the market in June. Even then, they will start early on at premium pricing (along the lines of the Athlon FX line) but by the holiday shopping season, dual-core processors should be more accessible to the mainstream performance buyer.