The CPU (Central Processing Unit defined by Smartercomputing), also called a processor or microprocessor, is the “heart” of the computer. Its task is to carry out the instructions of the memory-resident programs. The choice of CPU is perhaps the one that most affects the overall performance of your PC, than the other individual components present (except for 3D game systems, where it can be the video card, or VGA, the so called “bottle neck”).
But first to identify differences between the multiple CPUs on the market, you need to familiarize yourself with some technical terms, which identify the specific properties of each processor.
Note: on the market it’s easy to find both Retail and OEM, for each different CPU commercialized. As a general rule, you must keep in mind that the Retail versions, with a greater market price, feature package (complete package), which will present its heat sink (Heat Sink Fan), instruction manual and warranty, usually 3 years. OEM versions, intended for the market of assemblers, against a lesser price, will be without a pack, heat sink and booklets, while the 1 year warranty nominal duration, will be provided directly by the retailer.
Clock speed (speed or clock frequency)
The clock speed is the number of context switches between the two logical levels “0” and “1”, that the internal logic circuits to the CPU can perform per unit of time (1 second), and is expressed in cycles per second, or hertz, and its multiples: gigahertz (GHz) or megahertz (MHz, where 1 GHz = 1000 MHz); normally to execute a statement or a simple sum requires multiple clock cycles.
Usually a higher frequency value indicates a faster processor, although consider only this value, for the purpose of performance, may lead to erroneous valuations. In fact not all CPUs perform the same amount of work for each clock cycle, whereby two different models, although with the same operating frequency, can potentially have very different overall performance. In any case the frequency should never be used to compare two CPUS that belong to different architectures (AMD Phenom compared to Intel i7, for example), where only the comparison made by the comparison of specific benchmarks (set of software tests which measure the overall performance of the PC), can lead to objective assessments.
IPC (Instructions Per Cycle)
The CPI indicates the number of instructions executed by the CPU in a clock cycle, and directly depends on its efficiency in carrying out the instructions, apart from its clock rate.
Fsb (Front Side Bus)
On the Front Side Bus is a bus that carries data between the CPU and the northbridge. Maximum theoretical throughput (bandwidth) of the FSB is determined by multiplying the bytes of the word (words) processor clock frequency (cycles per second) by the number of data transfer in each cycle. For example, a system with 32-bit processor (4 bytes), FSB to 100 MHz and 4 transfers per cycle has a bandwidth of 4 (b) x 100 (FSB) x 4 (tc) = 1600 megabytes per second (MB/s).
The CPU will be inserted into the appropriate slot on the motherboard. It is essential that the size of the CPU corresponds exactly with that of the socket (connector) on the motherboard.
Under no circumstances attempt to use a CPU on a different socket, as it will be possible damage to both sides, and its invalidation of warranty!
Bit width (bit width)
Depending on the processor architecture, this value reflects the size of the internal registers of the CPU. In the personal computer market, starting from the year 2006, the 32-bit architecture was progressively replaced by a 64-bit, first used exclusively in server scope, i386, EM64T and PowerPC 970 architectures.
For our purposes we will use an approximation: 64-bit CPUs will be considered when they are capable of running a 64-bit operating system. If in doubt, we recommend that you carefully check the documentation for your particular operating system.
The Hyper-threading technology is launched by Intel, which allows a single-core processor to simulate the presence of two cores, resulting in increased performance when running multiple applications simultaneously. It was actively used by Intel between 2003 and 2006, only to find new life in Intel Atom UMPC systems designed for project and those ultra cheap. The idea behind the Hyper-threading technology was to duplicate some Pu within the microprocessors in order to be able to perform some operations, thanks to multithreading techniques. It was, in a sense, an attempt to create a transition between the traditional single-core processor and the next dual core, in a period when, because of the 130 nm fabrication process, did not allow “physically” the presence of 2 cores within a single package.
The CPU cache is the cache implemented directly within the CPU, which is used to reduce the average time to access memory. The cache is a small but very fast memory, which keeps copies of data to which access is more frequently, avoiding the main memory access through the northbridge, and thereby reducing the average latency of memory accesses. Generally, the bigger the size of the cache, the faster the system. The cache is normally of three different types: L1, L2, and L3. The L1 is the smallest and fastest, while the L3 is the largest and the slowest of the three. Normally the value of L2 cache that is advertised, while the L3 only rarely is used in the design of processors.
The core is the “processing” of a microprocessor. The CPU is actually composed of 2 main components: the core and the package that contains it. In some types of CPU core is visible, a small rectangle slightly protruding in the center of the package. This contains all the transistors (hundreds of millions), which determine the functioning and processing capability. Everything that’s around the core is the package — the “hoof”, made of plastic, ceramic or glass fiber, electrically connected to the contact pin that will engage in the socket.
Often several cores are marketed under the same name, for which you need to see exactly what we’re about to buy. It is in fact a mistake based in choosing only on the clock frequency measured in gigahertz. This number, although easy to understand, partly explains the actual capabilities of the processor. The number of physical cores are present, and how they communicate with each other, they have a big effect on the final performance of the system in which the CPU will be installed, more than the mere commercial name of the CPU could do portend.
For some years both Intel and AMD have reached a point where it is not practically possible raise the operating frequencies of their processors, and have decided to stake everything on multi core architectures, or two or more cores mounted on the same package. In this way, exploiting the parallelism, you get typical advantages of a multiprocessor system. Starting from the year 2010 Intel Core i7 technology, has switched to producing CPUS that Mount 8 cores on the same package.
- Although it is a relatively new technology, multi core processors have become virtually de facto standard of both major CPU manufacturers, AMD and Intel.
- Multitasking: each processor has a number of “Elaborative nuclei” corresponding to the number of cores, for an increase of 2 x (dual core), 4 x (quad-core), 6 x (hexa core) and 8 x (third core) associated with multitasking. The most obvious performance gains can be seen when running multiple heavy tasks, such as video encoding, virus scan, 3D processing.
- Newer applications are developed using the multi-threading, to reap the benefits of parallel processing of instructions.
- Multi core processors have power saving capabilities, able to selectively “off” one or more cores, when it is not required to use them, with clear benefits in terms of energy.
- In the case of older applications (certainly those released up to 2006, where it was not support multi-threading), the performance of a core multi can be definitely lower than running on a single core system but with a higher operating frequency.
- A quad core is not twice as faster than a dual core processor. When running a single application, which requires high computational resources (such as 3D games), the benefits of multicore architecture are negligible compared with those related to “simple” operating frequency.
Intel Corporation is a multinational semiconductor (microprocessors, memory devices, telecommunications and computer applications support circuits) based in Santa Clara, California. Founded in 1968, is a leader in its market segment, and is known for being for many years ranked first among semiconductor manufacturers in terms of revenue per year.
- Pentium range: Pentium, Pentium Pro, Pentium II, Celeron, Pentium III, Pentium 4, Pentium III-M, Pentium 4-M, Pentium 4, Mobile Pentium 4 EE, Celeron D, Pentium D, Pentium EE
- Mobile Processors: Pentium M, Celeron M, Core Duo, Core Solo
- Core: Core 2 Duo, Core range 2 Quad, Core 2 Extreme
- Range Nehalem: Core i3, Core i5, Core i7, Core i7 Extreme
- Server: Xeon processors
AMD (Advanced Micro Devices, Inc.) is an American multinational semiconductor manufacturer headquartered in Sunnyvale, California. It is one of the world’s leading manufacturers of microprocessors for consumer, workstations and servers, integrated and discrete graphics chips after the merger with ATI as well as chipsets.
- Athlon 64, Athlon 64 X 2 deskyop: sector, Athlon Neo X 2, Athlon 64 FX, Sempron 64, Sempron X 2
- Mobile industry: Mobile Sempron, Sempron for Ultrathin Notebooks, Mobile K8 Athlon XP-M, Mobile Athlon 64, Athlon 64 Mobile Athlon 64 x 2, Athlon 64 X 2 for Notebooks for Notebooks, Athlon Neo for Ultrathin Notebooks, Athlon Neo X 2 for Ultrathin Notebooks, Mobile, Turion 64 X 2 Mobile Turion 64, Turion Neo X 2, Turion X 2 Ultra Dual-Core
- Single-core Opteron Server sector: Opteron, Dual-core Opteron, Second Generation Opteron
- Desktop: Sempron, Athlon X 2, Athlon II X 2, Athlon II X 3, Athlon II X 4, Phenom X 3, Phenom X 4, Phenom II X 3, Phenom II X 4, Phenom II X 6
- Mobile industry: Mobile Sempron, Athlon II Dual-Core Mobile, Turion II Dual-Core Mobile, Turion II Ultra Dual-Core Mobile
- Server sector: Third Generation Opteron, Six-Core Opteron, Dodeka-Core Opteron