Dual Core Processor

Dual Core is a technology that inserts 2 identical processors in one chip. Similar technology used in network servers (Application Server) witch used 2 processors separate (or more) on a single motherboard.

This technique emerged in 2005 for microprocessors AMD and INTEL can theoretically double the processing speed. But to take advantage of this speed increase, the application (software) must be written in multi-threaded version, broken down into pieces, which allows both processors included in the electronic circuit to work simultaneously. It's the same problem with Hyper-Treading Intel. These are generally specific software for use on servers or workstations. In normal use, the advantage of two coupled processors will improve performance when the simultaneous use of two standard applications. The OS will allocate each application on a separate processor. A small flat anyway, memory bus and devices are shared between 2 cores, thus slowing the effective speed of assembly.

Within the processor, interfacing techniques are identical to those of assemblies containing multiple processors on one motherboard (SMP and Numa). The internal architecture of the AMD and Intel today is markedly different. In the case of Intel (Pentium IV, Xeon, Itanium), the memory controller is always included in the chipset. By cons, in current AMD processors (Athlon 64, Sempron 64, Opteron, Phenom), the microprocessor that runs directly from memory. The Dual-Core architecture will indeed be different depending on the market 2.

• Intel couple's 2 processors via an internal bus connected to the chipset. The architecture of the MMP. The L2 cache is directly shared between 2 processors. EE Pentium IV Extreme Edition, Pentium IV D and Celeron D (without the hyper-treading), Xeon and Itanium II came out in Bi-core. Processors for laptops are also available in Centrino.
  Motherboards for these processors require a specific chipset.
  Intel has modified this diagram using a bus by heart in the series Xeon Tigerton quad-core.
  
  

• AMD integrates the memory controller directly into the processor. If this feature requires an architecture NUMA multi-processors, there is nothing in this technology. The memory controller is also shared by both cores. If the L2 caches are managed separately, the memory controller and the HyperTransport (External bus) are operated simultaneously by the microprocessor 2 via a special circuit. AMD does not require a specific motherboard. We find the Athlon 64 and Opteron in this configuration.
  

Dual-Core or Dual processor? In terms of performance, implementation of 2 separate processors on a motherboard has some advantages, particularly in terms of memory buses and devices that are less divided. Thus, 2 separate processors is a little more efficient. The current tests do, however, show that differences of the order of a few percent. By cons, this solution will increase the price for the motherboard.

The other big difference is the price of licenses for operating systems XP and 2003. Microsoft consider a dual-core (including with the hypertreading) as a single entity and allows the use of a standard license (including Windows XP Home). Otherwise, a dual-processor requires at least one Windows 2000 or XP Pro. Vista also accepts these processors, all versions.

Intel released the second quarter of 2006 Xeon Quad-core. These processors are compatible with the motherboards made for the Xeon 5100 series (dual-cores).

Hyperthreading Intel is a software version that handles only 2 logical processors for one physical microprocessor.

Related: 64 bits technologies