Intel Core 2: overview of the Core 2 processor family

The designation Intel Core 2 is independent of the number of processor cores of a processor. The "2" rather stands for the second generation in which the brand name "Core" is used (but not for the second generation of the Core microarchitecture, since the Intel Core is based on a heavily modified version of the P6 architecture and is therefore only a precursor of the new microarchitecture).

The variant for the mass market with two processor cores (dual-core processor) is called Core 2 Duo, the quad-core processor is called Core 2 Quad. The models for the high-end market bear the name Core 2 Extreme in the tradition of the Extreme Edition models of the Pentium 4 and Pentium D. For mobile computers, there is still a frugal variant with only one processor core, marketed as Core 2 Solo. All processors receive a four-digit model number, where the first one classifies the series and the exact model is indicated via the remaining three digits.

Prefix

This model number is preceded by another letter (in rare cases two), which gives details about the power consumption in the mobile market. In the desktop market, the letters classify the type of processor (dual-core or quad-core processor, Extreme models):

In contrast to Intel's previous desktop and server processors based on the Intel NetBurst microarchitecture, the Core 2 relies on Intel's new, largely Pentium M-based Core microarchitecture. This differs from the NetBurst architecture not only in its significantly lower electrical power consumption, but also in its drastically shortened pipeline. Although this reduces the possible maximum frequency, the performance increases at the same clock frequency, so that the new processors perform about twice as well at the same clock frequency.

In addition to the shortened pipeline, however, these performance increases are also due to various improvements to the Core 2's computing units. The improved and partly newly introduced vector operations only need one clock cycle - half as much as before - and by means of Macro-OP Fusion the instruction decoder can generate a single MAC function from addition and multiplication. According to Intel, on average every tenth instruction should benefit from the latter optimization.

Furthermore, power saving features of the predecessors Pentium M and Intel Core have been adopted. Besides the dynamic adjustment of clock and operating voltage, many parts of the CPU can be switched off when not in use in order to save energy. In combination with the generally lower power consumption of the new Core microarchitecture, but also the lower clock rates, this makes the Core 2 one of the most energy-efficient processor developments ever when it was released.

New steppings, including M0, have further reduced idle power consumption.

In late 2007, new processors were introduced in 45 nm fabrication with a somewhat optimized architecture. The shrink to 45 nm itself and the change in transistor construction to high-k dielectric (based on a hafnium compound)/metal gate also make faster transistors with smaller leakage current possible.