CPU basic knowledge of the source Web browser 886 Anonymous Posted 08/08/08 CPU Basics. .
CPU is what we call the central processor, the English name CentralProcessingUnit, it is a piece of VLSI chips, internal tens of millions more to a transistor symbols consisting of very complex circuit, including a calculator, registers, controller and the bus (including data, control, address bus). It operates through the instructions and control system, which is the core component of the overall system.
Now use the most CPU are Intel and AMD (AdvanceMicroDevices, Inc..). The original 16-bit processor, starting from 386 to 32-bit processors, and later 32-bit processor can run 16-bit processor running instructions to collectively known as the x86 family of processors. .
Here to introduce you to the CPU specifications and features.
1 word length. .
CPU of generally refers to its data bus width, binary units, that is what we are talking about is how many bits of the processor, the CPU data processing capability as an important indicator that reflects the processor capable of handling the data is width, precision and speed, etc. Now the processor's internal data cable as a 32-bit external data cables for 64-bit, we still call it a 32-bit processor, it is just a 64-bit mode and memory Exchange data. But the latest AMD Athlon64, its internal bus is 64 bits, the 64-bit processor.
2, the main frequency. .
The main CPU frequency refers to the work of the CPU clock frequency, measured in Hz, but the units are CPU GHz (1MHz = 1000Hz, 1GHz = 1000MHz), which is an important indicator of CPU speed, usually marked in CPU models, such as Pentium42. .4GHz, its clock speed to 2. .4GHz. CPU input clock frequency as FSB, often taken as the motherboard system bus frequency, such as the P4 started from 133MHz to 533MHz, up to the current 800MHz, CPU's internal multiplier with × 3 × 3. .5 × 4 × 4. .5 × 5, such as an FSB to 800MHz, multiplier is × 3 of Pentium4CPU, its frequency on the 2. .4 GHz. .
3. the internal Cache.
In order to solve the host of low-speed memory and high-speed CPU's do not match, the memory access speed up the CPU speed, CPU and memory used in the insertion of cache (Cache) The way to start the Cache is installed on the motherboard, and later as the internal CPU L1 and L2Cache using high bandwidth bus, outside the L2Cache much faster than the chip, the efficiency is much higher, and can run simultaneously on the CPU's frequency, it gradually replaced the outside of the Cache, as now mainstream.In Intel's newest Pentium4ExtremeEdition, joined up 2ML3Cache allowing the processor to access the memory speed and reached a new level.
4, Hyper-Threading. .
Hyper-Threading (HyperThreading) technology was developed by Intel in P4 processor technology, it allows a single P4 processor simultaneously perform two threads (or part of a software program). Pentium 4 processor supporting Hyper-Threading technology for operating systems (such as WindowsXPProfessional) can be a physical P4 processor "as" two virtual processors. Idle resources through the use of other, in a multitasking environment with Hyper-Threading technology for P4 processor can make existing software to realize significant performance gains, without modifying the code.
CPU speed upgrade, and update technology to allow us to appreciate the technology has brought us pleasure. AMD introduced 64-bit processor means that the coming era of 64-bit processors, expected to be launched Pentium5, will give us an experience of what it was? . .
What is bus:.
In order to constitute a computer's various features into a working system, they must be organized by some way connected to the bus (BUS) is the computer to transfer information between the various components of public access. .
。.
In the computer, the internal bus and external bus into two categories: internal bus is the connection between the internal CPU; external bus is the CPU and other parts of the connection between. Daily said the bus generally refers to the external bus, according to their different functions, the bus is divided into three types:. .
Data bus (DB), address bus (AB), control bus (CB).
. .
A.-data bus: used to transmit data median, which is generally the same length with the microprocessor, data bus that has a two-way functionality.
B. . Address bus: used to send the address information, which is one-way transmission. .
C.-Control bus: used for transmitting control signals.
. .
。.
List of CPU family. .
Original author: zhaoxin.
Intel co-founder of GordonE. . Moore Dr. 1965: IC chip integration will increase every 3 years, 4 times, while the process feature size reduced 2-fold, as Moore's Law; up only today, time is a leisurely 37 years my Baby ... .... .
Intel company in 1971, November 15, launched the world's first microprocessor-4004, Intel has introduced a 8-bit 8008/8085, of which 79 year launched the first IBM 8088 PC into Intel 8088, also released a 16-bit 8086/80286, 32-bit 80386/80486 (also known as i486) until you are familiar with Pentium (80586), PentiumII PentiumMMX, until today's PentiumIII, Pentium4, 64-bit Itanium, Northwood, Moore's law, and even become a truth claim to solving technical progress and development . .
In the late 80 's, the new global economy ushered in an unprecedented period of high speed and large leap. At this point, the microprocessor areas. Intel to era leader posture leads the microprocessor market. 20 October 1992, in New York at the tenth session of the General Conference, the PC user who was CEO Andrew Intel. Grove announced the fifth generation of Intel processors, named Pentium, but not before 586 people expected, which contrary to many people expected, because the Intel products prior to the 386,486 outstanding performance, but AMD and Cyrix processors produced as the figures named. .
Although the Intel but helpless, because, according to the trademark law, pure digital is not used as trademarks. The new period with Pentium, microprocessor into new round clash near the troubled times. For decades, regardless of market share or core technical or market strategy, Intel has always been in the limelight. AMD, Cyrix, etc. even though Motorola has been at all times their own unique skills and brilliant, but in the end is difficult to shake Intel's leading position in today, Cyrix has been unable to keep up with high-frequency processors, real-time R & D faster pace, while the growing trend micro! Huang Huang Diocese, the capable, with the passage of time natural selection, how many gallant brave man fade away in mediocrity because of flow in rivers and lakes ... ... but the name list of those achievements in the past has not left did not, a section of the old footprint, forged the history of today, whether it is brilliant, mediocrity or failure. .
This paper finishing from 486 times to date, Intel, AMD and Cyrix in desktop PCs, servers, workstations, mobile market previously all CPU types and their descriptions. For your reference.
Intel products. .
Code descriptions.
Desktop CPU. .
486.
P24486DX2. .
P24CDX4 。.
P24T486DX2 with OverDriveprocessor. .
OverDriveprocessor P24CTDX4 used.
Pentium / MMXPentium. .
P50.8 μ m Pentium version.
P54C0. .6 (0. .5) Μm version of Pentium. .
P54CTPentiumOverDriveprocessor 。.
P54CQS early version of 0. .6 Μm version of Pentium. .
P54CS0.35 μ m Pentium version.
P55C0. .35 Μm Edition MMXPentium. .
P54CTBMMXPentiumOverDriveprocessor 。.
Tillamook0. .25 Μm Edition MMXPentium. .
PentiumPro/II/III series.
P6PentiumPro. .
Klamath0.35 μ m PentiumII version.
Deschutes0. .25 Μm Edition PentiumII. .
Katmai0.25 μ m PentiumIII version.
Coppermine0. .18 Μm copper version of the core PentiumIII. .
Tualatin-256k0.13 μ m PentiumIII version.
Celeron. .
Covington0.25 μ m Celeron version.
Coppermine-128k current 0. .18 Μm version of Celeron. .
TimnaGMCH consolidation Celeron version.
Timna + Enhanced Edition Timna. .
Willamette-128kPentium4 core Celeron.
Pentium4. .
Willamette-423423 pins 0.18 μ m version Pentium4.
Willamette-478478 Pin 0. .18 Μm version of Pentium4. .
Northwood0.13 μ m version Pentium4.
Prescott90nm subsequent version Pentium4 CPU. .
Tejas future IA-32.
Server CPU. .
PentiumII/III 。.
DS2PPentiumIIXeon. .
Tanner0.25 μ m PentiumIIIXeon version. KatmaiSlot2 interface.
Cascades0. .18 Μm Edition PentiumIIIXeon. .
Pentium4 。.
Foster0. .18 Μm version of Xeon (Willamette). .
FosterMPHyper-Threading corresponding bulk-Xeon Server Edition.
Gallatin0. .13 Μm version of Xeon. .
Prestonia servers and workstations used 0.13 m version Xeon μ.
Nocona2003 debut in the new CPU. .
IA-64 。.
Merced 1st generation Itanium. .
McKinley0.18 μ m version IA-64 generation 2.
MadisonMcKinley0. .13 Μm version. .
DeerfieldMcKinley0.13 μ m version.
Montecito90nm version of IA-64. .
Mobile CPU.
Banias. .
Banias2003 2q08.
Dothan90nm version of Banias. .
PentiumII/III 。.
Dixon0. .25 Μm mobile PentiumII. .
Dixon (0.18 μ) 0.18 μ m mobile PentiumII.
Coppermine0. .18 Μm mobile PentiumIII. .
Tualatin-512k0.13 μ m mobile PentiumIII.
Celeron. .
Dixon-128k0.25 μ m Edition Mobile Celeron.
Coppermine-128k0. .18 Μm version of mobile Celeron. .
Mobile version of Timna + cancelled as Timna mobile.
Tualatin-256k0. .13 Μm version of mobile Celeron. .
Chipset.
PentiumIII / Celeron. .
CaminoIntel820 。.
Camino2Intel820E. .
SolanoIntel815 。.
Solano2Intel815E. .
WhitneyIntel810 。.
Camino3 canceled support PentiumIII, RDRAM chipset. .
Cancellation support Almador Tulatin, RDRAM chipset.
Pentium4/Celeron. .
TehamaIntel850 。.
Brookdale-SDRAMIntel845, Pentium4 support SDRAM. .
DDRSDRAM Brookdale-DDRIntel845, Pentium4 support.
Brookdale-GIntel845G, Pentium4 integrated graphics chip. .
Brookdale-GLIntel845GL,Pentium4 。.
Tehama-EIntel850E. .
Tulloch had been cancelled Pentium4 support future RDRAM specifications.
Springdale-PNorthwood & Prescott. .
Springdale-GNorthwood&Prescott support future graphics chip integration.
Mobile CPU. .
BanisterIntel440MX 。.
Almador-MIntel830M, Tualatin. .
Greendale has been canceled by the chipset supports RDRAM.
Brookdale-M mobile Pentium4-M corresponding to Intel845 chipset. .
Brookdale-MZ Mobile Celeron corresponding Intel845 chipset.
OdemBanias corresponding mobile. .
The corresponding consolidated graphics chip Montara-GMBanias mobile version.
Montara-GMLPentium4-M/Celeron corresponding integrated graphics chips for mobile. .
Servers and workstations.
CarmelIntel840. .
ColusaIntel860(Foster) 。.
PlumasIntelE7500. .
IntelE7505 PlacerPrestonia corresponds.
GraniteBayNorthwood corresponding IntelE7205. .
LindenhurstNocona 。.
TumwaterNocona corresponding workstation. .
Pentium.
Marcury60/66MHzPentium chipset. .
Neputune75/90/100MHzPentium with chipset.
TritonPentium with the chipset. .
Slot.
Pentium4 Socket-W first slot. .
Socket-NNorthwood slot.
Socket-FFoster slot. .
Other.
YamhillIA-32 64-bit extended version of the corresponding?. .
JacksonTechnologyHyper-Threading 。.
GeyservilleSpeedStep. .
Powered by new generation SpeedStep Geyserville-IIIBanias.
AuburnIntel740 graphics chip. .
PortolaIntel752 。.
IndianBeach has been canceled Intel752 successor. .
Coloma has been cancelled.
Capitola has been canceled. .
Descanso has been cancelled.
AMD products. .
Code descriptions.
5x86. .
X5Am5x86 。.
K6. .
Nx686NexGen K6 processor code groups.
K6K6. .
AMD-K63DK6-2 。.
AMD-K6 +3 D0. .25 Μm version of K6-III. Sharptooth renamed. .
Sharptooth0.25 μ m K6-III version.
SharptoothSC500. .18 Μm version of K6-III. .
Athlon systems.
K7Athlon. .
K750.18 μ m version of the Athlon.
K760. .18 Μm copper core version of Athlon. .
Magnolia1GHz Edition Athlon (k75).
Thunderbird now integrated L2 0. .18 Μm version of Athlon. .
CorvettePalomino renamed.
PalominoAthlonXP (0. .18 Μm new Athlon). .
Thoroughbred0.13 μ m AthlonXP version.
Barton0. .13 Μm version of L2 capacity increase AthlonXP. .
Mustang bulk L2 version Athlon.
Duron system. .
SpitfireDuron 。.
CamaroMorgan renamed. .
Morgan0.18 μ m new Duron.
Appaloosa0. .13 Μm version of Duron. .
Hammer lines.
K8AMD generation CPUHammer 総 said. .
ClawHammer servers and workstations with a Hammer.
ClawHammer-SClawHammer the 90nm version. .
SledgeHammer server and workstation using L2 version Hammer.
Chipset. .
GolemHyperTransport support for server and workstation chipsets corresponding Hammer.
LokarHyperTransport support server and workstation chipsets corresponding Hammer. .
Thor future server and workstation chipsets?.
. .
VIATechnologies description.
. .
Old Cyrix.
CayenneMII extended CPU core. .
Jalapeno next-generation CPU core.
MXiCayenne integrated CPU. .
The CyrixIII JediCayenne. Gobi renamed.
GobiCayenne of CyrixIII. Joshua renamed. .
The CyrixIII JoshuaCayenne.
Mojave on Jalapeno new generation CPU. .
CentaurTechnology 。.
Samuel1 now 0. .18 Μm Edition CyrixIII = C5A. .
Samuel2C3 = 0.15 m version CyrixIII μ = C5B.
C5A is 0. .18 Μm Edition CyrixIII = Samuel1. .
C5BC3 = 0.15 m version CyrixIII μ = Samuel2.
MatthewC5B integration of the core CPU. .
C5C0.13 μ m version CyrixIII = Ezra.
Ezra0. .13 Μm Edition CyrixIII = C5C. .
The mutual exchange of Ezra-TTualatin bus C3 = C5M/C5N.
C5MTualatin bus interchange C3 = Ezra-T. .
The mutual exchange of C5NTualatin bus C3 = Ezra-T, 0.13 μ m.
C5XC5. .
NehemiahC5XL 。.
C5XLC5X high-end low-end version and version. .
Low frequency version C5YLC5XL.
Other Old CyrixCPU. .
M1sc5x86 。.
M16x86. .
M26x86MX/MII 。.
5GX86MediaGX. .
GX86MediaGX 。.
GXmMMX version MediaGX. .
。.
CPU sockets and slots. .
Slots and Sockets and Slocket are used to put the CPU installed on the motherboard. In the early 1980s, IBM PC 8086 fresh out of the oven, is directly welded on the motherboard, then the 286 and 386 are welded on the motherboard, it's not good to remove the user, once you have bought a new computer is basically nothing to upgrade. A 486 processor manufacturers later, began to use the socket or slot to install CPU, believe that computer users are not unfamiliar to us. A variety of CPU currently on the market a wide range of sockets and slots used also vary, this article gave you a brief overview of the various CPU sockets and slots. .
Socket1 。.
Intel developed the most ancient CPU socket for the 486 chip. 169 feet, and voltage of 5V. The multiplier can only support the DX4. .
Socket2 。.
Intel made on the basis of the Socket1 small improvements are Socket2. Socket2 have 238 feet, the voltage remains at 5V. Although it. .
Or 486 Socket, but slightly modified to support the Pentium.
Socket3. .
Socket3 is Socket2 developed on the basis of. It has 237 feet, the voltage is 5V, but can be viewed by jumpers on the motherboard is set to 3. .3V. It supports all of the CPU, Socket2 also supports 5x86. It is finally a 486 Socket.
Socket4. .
Pentium CPU socket from the era of Socket4 begins. It has 273 feet, the working voltage of 5V. It is precisely because of its operating voltage is too high, so it didn't spread was replaced by Socket5. Socket4 can only support the Pentium 60-66MHz.
Socket5. .
Socket5 has 320 feet and a working voltage of 3 .3V for.. It supports from the Pentium 75MHz to 133MHz. Socket5 socket in the early Pentium is very popular in. ..
Socket6. .
At first you might think this is a Pentium socket, but in fact Socket6 is a 486 Socket. It has 235 feet, the working voltage of 3 .3V, Socket3. a little advanced. However, as the popular Pentium 486 will soon no longer be mainstream Socket6 also soon forgotten.
Socket7. .
Socket7 is by far the most popular and most widely used CPU socket. It has 321 feet, operating voltage range of 2.5-3. .3V. It supports all starting from 75MHz Pentium processors, including PentiumMMX, K5, K6, K6-2, K6-3, M2 and M3 6x86,. Published by Intel Socket7 is, in fact has become the industry standard, you can support the IDT, AMD and Cyrix CPU of the sixth generation. However, Intel in the development of their own sixth-generation CPU-PentiumII that has decided to abandon Socket7, also create a situation. .
Socket8 。.
Socket8 is PentiumPro dedicated outlet. It has 387 feet, the working voltage of 3. .1 / 3. .3 V. It also made for dual-processor motherboard with special design. But as the mainstream shifted from PentiumMMX PentiumII, Socket8 soon be forgotten. .
Slot1 。.
Slot1 completely changed the appearance of Intel's CPU socket consistent shape. Intel original CPU is a square, pin the bottom of the chip, install plug in the motherboard CPU socket. The PentiumII is no longer the Quartet, and processor chip welding in a circuit board, then this circuit board and then plug into the motherboard slot, the slot is Slot1. This design processor cores and the communication between the L2 cache faster. Slot1 have 242 feet, the working voltage of 2. .8-3. .3 V.Slot1 is mainly used for P2, and P3 Celeron (Celeron), in addition to the riser Socket8 used to install PentiumPro.
Slot2. .
Slot2 is improvement of Slot1, mainly for the Xeon series processor. There are 330 feet Slot2, Slot1 and the biggest difference between lies in Slot1 CPU L2 cache and only half of the CPU operating frequency to communicate, and Slot2 allows the CPU to CPU L2 cache and the frequency for communication.
Socket370. .
As the name suggests Socket370 outlet has 370 pins. Found in the Intel processor cores and L2 cache is very cheap to do together, its CPU socket returned from the Socket Slot. Socket370 is based on Socket7, however only in the sides of each side of the socket with a row of pins. The first is using Socket370 PPGA package Celeron, followed by the FC-PGA package PentiumIII and CeleronII. Equally, there Socket370 to Slot1 the adapter. The current mainstream Intel CPU are Socket370 type. .
SlotA 。.
As to the Slot1 Intel patent application for a very comprehensive, AMD can not be copied, as usual, Intel's socket, so AMD has independently developed SlotA, SlotA is the AMD CPU sockets with independent intellectual property rights, mainly for Athlon family of processors. Its design and Slot1 similar, but not the same protocol used, it is the EV6 bus protocol with. EV6 bus protocol used, CPU and memory can be achieved between the operating frequency 200MHz.At present, more and more popular as Athlon processor, SlotA motherboard is also increasing.
SocketA. .
When Intel Socket returned from the Slot, AMD also picked up the pace, from the SocketA SlotA back. 0. .18 Micron's Athlon and Duron takes the SocketA outlet, it also supports 200MHz and EV6 bus 266MhHz 's. And Socket370, Socket370CPU can directly use the radiator, Socket7 SocketA radiator to be slightly modified. In addition to the SocketA AMD does not provide SlotA's riser card. SocketA have 462 feet, it is not compatible with Socket370. AMD CPU is the current mainstream SocketA type. .
Slockets 。.
The so-called Slocket is a combination of Slot and Socket, from which can be seen on the spelling. It is essentially a Slot1 to Socket370 the adapter, at different levels and switch between interfaces. Some Slocket can plug two CPU, there's Slocket the locking CPU can be removed to make overclocking easier. .
Above the introduces a bit of various CPU sockets and slots, and hope to be able to raise this aspect of the content. Although buying whole, usually together with CPU and motherboard, CPU loading does not appear does not appear. But when the user during the upgrade, it is necessary to note in this regard, to buy their own Motherboard CPU can support.
CPU microprocessor instruction set. .
Instruction set is the CPU to perform a collection of all the directive is to improve the efficiency of the microprocessor is the most effective tools, from its present mainstream architecture, instruction set can be divided into complex instruction set and reduced instruction set, but from the concrete application, multimedia instruction sets and is one of the most compelling one.
RISC application. .
After the initial invention of computers N decade, along with the computer function is increasingly growing, the increasingly strong performance, increasing the internal components, the instruction set is becoming increasingly complex, overly miscellaneous instructions seriously affects the efficiency of your computer. Later, some scientists has found that, in the computer program only, 80 per cent to 20 per cent of instruction set, based on this discovery, RISC reduced instruction set came, this is a computer system architecture of a profound revolution.
RISC architecture, the basic idea is: to seize the many species of CISC instruction directive, instruction format is not standardized way of addressing many of the shortcomings, by reducing the instruction types, instruction formats and simplify the addressing standard way to facilitate the internal processor parallel processing, VLSI devices to improve efficiency, thus greatly improving processor performance. .
RISC instruction set has many features, the most important of which are: less directive directive type format specification RISC instruction set is usually only one or a few formats. Instruction length single (typically 4 bytes), and aligned on word boundaries. Field locations, especially the location of the opcode is fixed. Addressing ways to simplify the use of almost all the instruction register addressing addressing total is less than five. Other more complex addressing modes, such as the indirection etc from software with simple addressing to synthesis.
Heavily on inter-register operation RISC instruction set to register the majority of operations are register operations, only a simple access memory Load and Store operations. Therefore, each instruction to access the memory address will not be more than one, access memory and arithmetic operations do not mix. .
Simplify processor architecture using RISC instruction set, you can greatly simplify processor controller and other functional unit design, you do not have to use a large number of special-purpose registers, especially allows for hardware lines to implement the directive, without having to use as the CISC processors to enable micro-program. Hence the need for RISC processors CISC processor that sets the microprogram control memory, you can quickly direct implementation of the directive. VLSI technology as easy to use LSI and VLSI technology, the processor (or even more than one processor) can be placed in a single chip.
RISC architecture to the design of single-chip processor can bring many benefits to help improve performance, simplify design and implementation of VLSI chips. Based on VLSI technology, manufacturing, CISC processors, RISC processors, the workload is much less than the cost is much lower. Enhanced RISC processor instruction set parallelism can very effectively for using lines, super lines and superscalar techniques to achieve instruction-level parallel operations to improve processor performance. The commonly used technique is basically parallel operations inside the processor is based on RISC architecture, development and maturing. .
It is precisely because of the RISC architecture, which has advantages in high-end system has been widely used, and the CISC system is the desktop system dominates. But now, in the desktop area, constant osmotic RISC, expected future, will be a marketing stunt RISC.
Multimedia instruction set to change lives. .
For the CPU, the basic functions, their differences are not too large, the basic instruction set is the same, but many manufacturers in order to enhance a particular aspect of performance the extensions, extensions to define a new data and instructions, to dramatically increase the ability of certain aspects of the data processing, but it is also a precondition that will need to have software support. In multimedia instruction sets, the most famous is Intel's MMX and 3Dnow! of AMD.
. .
First, the MMX instruction set.
MMX (MultiMediaeXtension, Multimedia Extension instruction set) instruction set that Intel launched the company in 1996, a multi-media enhancement technology instruction. MMX instruction set includes 57 multimedia instructions, these instructions can handle more than one data processing results in more than the actual processing time can be properly handled so that the cooperation of the software, you can get higher performance. The benefits of MMX is that the then existing operating system without making any changes for this purpose can easily perform MMX program. .
However, the problem is obvious, that is, the MMX instruction set and x87 floating point instruction will not be able to execute simultaneously, you must do the intensive, staggered over to the normal execution, this situation inevitably run quality system as a whole declined.
2, SSE instruction set. .
SSE (StreamingSIMDExtensions, extension) instruction set is Intel in P e t I n u m Ⅲ processor delivers. In fact, as early as before the formal launch of P Ⅲ, Intel Corporation has announced through a variety of channels across the so-called KNI (KatmaiNewInstruction) instruction set, this command sets the SSE instruction set the name of the earliest and at many media called the MMX instruction set of the next version, i.e. MMX2 instruction set. Research background, the original "KNI" instruction set is the next generation of Intel chips for its company name the first instruction set name, and the so-called "MMX2" is completely hardware and media commentators, and the impression on the seat of your pants "KNI" Evaluation of , Intel has never been officially released the news on MMX2. .
While the final launch of the SSE instruction set called the "Internet" SSE instruction set. SSE instruction set includes 70 instruction that includes improved 3D graphics operations efficiency 50 SIMD (single instruction multiple data) floating point instruction, 12 MMX integer arithmetic enhanced instruction, 8 optimization in continuous memory block directive.
These instructions on the current popular theory, image processing, floating-point, 3D computing, video processing, audio processing and many other intensive multimedia applications play a full role. SSE instructions and 3DNow! Instructions incompatible with each other, but the SSE includes 3DNow! Technology, most of the functions, but only realized in different ways. SSE-compatible MMX instructions, it can be a single clock cycle SIMD and parallel processing of multiple floating-point data to effectively improve the speed of floating-point operations. .
In order to deal with later Intel AMD3DNow! + in SSE has developed on the basis of the SSE2 instructions, increase, making it a P4 processor performance increases. SSE2SIMD extension — a low industrial best method is to let each instruction to perform more work. To the end of the P4 design, Intel added a set of instructions, including 144 new SSE2 instruction set. Like most early SIMD Extensions SSE2 instruction set, involving multiple data target immediately execute a single command (i.e. SIMD). The most important thing is SSE2 capable of handling 128-bit double precision floating-point math and. More precise floating point processing ability to accelerate multimedia program SSE2 become, 3D processing projects and tasks based on the workstation type configuration. It is important that the software is optimized to use it properly. .
。.
Given a DVD MPEG4 video file conversion software "FlasKMPEG", re-compile and SSE2 optimization of the performance before and after comparison. Obviously, after the software compiled in Pentium4 floating-point improvement on performance up to +266%, after re-compiled software in the performance of the Athlon and Pentium3 also been greatly improved. .
Third, 3DNow! instruction set.
Proposed by the AMD company 3DNow! SSE instruction set instruction set should be out before now, and is widely used in the AMD K6-2, K6-3 and Athlon (K7) processor. 3DNow! Instruction set technology is in fact 21 of the expansion of machine code instruction set. .
。.
. .
Support the Duron processor 3DNow!.
Intel's MMX technology and focus on integer arithmetic is different, 3DNow! Instruction set mainly for the three-dimensional modeling, coordinate transformation and the effect of three-dimensional rendering and other applications, in cooperation with the software, can greatly improve the 3D processing performance. .
Later in the development of the Athlon Enhanced3DNow!. These standards of SIMD instructions AMD and Intel's SSE has the same effect, but AMD in persuading the software developer to these instructions work program, experienced a relatively difficult period. Because of the commercial as well as Intel PentiumIII successful impact software in support of SSE, 3DNow! is more prevalent.
Enhanced3DNow! AMD has continued to increase to 52 instructions, including a number of SSE code, thus optimized for SSE software do get better performance. But we can be found in AMD's future Hammer-64-bit processors found, AMD finally followed the footsteps of Intel, it will fully support the SSE2, Intel's Pentium4 the introduction of 144 pairs of floating-point SIMD instructions exactly.While AMD 3DNow! later development can only be regarded as a transitional + programme and cannot really be a generation of technology, in fact, from a technical analysis of the 3DNow! + also has no significance, compared with SSE2 has no advantages.
. .
CPU optimization articles.
CPU in a computer's role is crucial, there is a high speed of the "core" may be what every fans dream. In fact, even if your "core" not strong enough, can be achieved through the optimization is "Pentium" effect, how do not believe? Then a try ah! . .
1. .BIOS optimization.
BIOS that BasicInput / OutputSystem, basic input output system, it is a first step in computer work. Long journey let the BIOS on fully prepared it! . .
Open the CPU reserve forces — CacheCache is very important for the CPU, processor, and memory, data exchange, high-speed buffer zones, you can use the data and instructions for selective retention for later called directly, which will significantly improve system speed. In fact, the importance of the cache for example you can see, why the performance of the P4 now back to the game much Jan 4? are the same core why performance is not the same, very simple difference in level 2 cache, which is L2Cache. So open the cache can let the system faster. The general board of these options are turned on by default in the present, but in some older motherboards are not always open. Inside the BIOS we can choose "AdvancedChipestFeatures" inside open CPUL1/L2Cache. .
If the Cache is damaged, or not to open the open because the system will become unstable. Of course, if you shut down the whole system is very slow!.
2. . System Optimization. .
Although in the BIOS has opened up the L2 cache, and so on, but in the system, the operating system does not necessarily have to be very good recognition L2Cache capacity for AMD CPU which is particularly serious, then what we need in an operating system manually, or by software to modify L2Cache data. Then do?.
In the "Start", select "Run" then type "regedit" If you are under the win2000/xp can also enter "regedt32" and press Enter to open the registry modifier, find the following path "HKEY_LOCAL_MACHINE \ System \ CurrentControlSet \ Control \ sessionmanager \ memoryManagement "and then in the right window find the" SecondlevelDataCache ", then modify the Dword value, and then enter the corresponding hexadecimal value of 16 l2Cache。 Celeron processor, 512 KB L2 cache for 128KB should set its value to 80 (hex, the same below). P Ⅱ, P Ⅲ, P4 are 512KB secondary cache, should be set to 200; P Ⅲ E (EB), P4Willamette only a 256KB secondary cache, should be set to 100; AMDDuron only 64KB secondary cache, should be set to 40; K6-3 have 256KB cache 2; Athlon has 512KB cache 2; AthlonXP 256KB secondary cache has; AthlonXP (Barton core) have a 512KB secondary cache. .
。.
If the use of software changes, now more commonly used are "optimized master" and "Super Rabbit" after installation follow the prompts to modify more than good value. Masters selected to optimize "system performance optimization," then choose "File System Optimization" At this moment there are two cache adjustment, as shown. .
。.
Start the Super Rabbit Super Rabbit Magic Set and then select "hard disk and CD-ROM" into the hard drive option, you can find the second cache of adjustment options. Fig. .
。.
3. . Overclocking your CPU. .
CPU due to workmanship and other differences can generally be overclocked, or less. So how do you overclock your CPU? I briefly, the current CPU frequency is how come I think we all know, frequency = FSB x multiplier. According to this formula may simple to figure out. Also the resulting two overclocked method, that is hyper-FSB and hyper-frequency-doubled.
There are two current mainstream CPU: Intel's and AMD's. .
1, Intel, CPU himself worthy of CPU always holds considerable market.
2, AMD, CPU vendors rising star, also has a considerable market share. .
Currently on the market of the CPU, Intel P4, C4 and C3 series are locked multiplier, for this kind of CPU overclocking approach we have taken only one, that is hyper-FSB. Currently on the market for AMD 0. .13 Micron process the Athlonxp is not the fixed multiplier, but its previous CPU frequency is locked.
Hardware overclocking. .
Upgrade: CPU multiplier.
This method is currently only suitable for K62 and Duron and Tbird the CPU, and the latest 0. .13 Micron process AthlonXP, Duron and Tbird if still use a pencil to break frequency, and many articles, presentations, there will not repeat them. Need to modify the motherboard to support super-harmonic frequency, and buy time to pay great attention to the motherboard. .
CPU FSB: upgrading.
FSB improve system performance can bring substantial upgrade for the PIII, the current FSB 100 are generally only to the 133 or so super, PIV processor has reached the maximum FSB of 200, but 100 and 133 relatively FSB overclocking ability more. In the heat good and also when the voltage increases, or even to more than 150. But in time, need your computer's memory, graphics card can work in such a high frequency under. Therefore, relatively speaking, 100 FSB PIII processor, is ideal ultra FSB CPU.This method to enhance the CPU multiplier method, the together for best results. Of course, this requires that your motherboard supports FSB, have a motherboard that supports the regulated by Siu is dedicated solely to hyper-FSB..
Software overclocking:. .
Software overclocking overclocking software is used, for example, gigabyte motherboard, there may be software overclocking model. Examples of these software overclocking will be in a future article. In addition, some third-party software can also be very good overclocking, but it was formerly known as overclocking software only support modifying the FSB, famouser e.g. SoftFSB. following detailed introduce SoftFSB use. Here the main overclocking software usage:.
l. . SoftFSB. .
First, download and install the latest version of the software, it's Needless to say.
SoftFSB no additional setup steps, as long as the configuration files and program files in the same directory can be run. 13 documents generated in four languages, including documentation, we have the best read before using SoftFSB instructions in order to understand a number of considerations. .
General usage.
Double-click the extract generated SoftFSB. . Exe, will start SoftFSB. Most of the system state (temperature) monitoring software should use SMBus, before running SoftFSB the best out of these software to avoid conflicts. .
SoftFSB's main interface, is a fairly simple dialog box, which is displayed below the current internal CPU frequency. TargetMainboard above is used to select the type of motherboard, SoftFSB1. .7 support nearly 90 models of motherboards, including Asus, Abit, EPoX, Intel, MSI, SOYO, dozens of famous motherboard manufacturer's products. If you know the model of the motherboard, and it is listed in the drop-down list box, select it.
If your motherboard does not appear in the above drop-down list box, we must open the case, found from the motherboard clock generator (usually a 48-foot or 56 foot chip, located close to the CPU or memory) and write down the chip models (such as a piece of the motherboard clock chip is ICS9250BF-08), and then select the SoftFSB the TargetClockgenerator, and from the right side of the drop-down list box to select the correct clock chip type (eg ICS9250BF-08 should be selected ICS9250xx-08).SoftFSB supported clock chip types are also up to 70 more.
Select the correct motherboard good or clock chip, click GetFSB button, SoftFSB it can read the current FSB frequency and the corresponding PCI bus frequency. Dialog box button within the SetFSB and TaskTray also become an effective state. Slider is divided into a number scale, each scale indicated that the board can set a FSB frequency. Some motherboards even support a dozen frequencies, in which a small number of frequency can not be found in the motherboard manual, can not get through the jumper or BIOS settings, which is used SoftFSB windfall. .
Now we can drag the slider to increase the level of FSB frequency, e.g. from 133MHz 100MHz transferred to, and then press the SetFSB, just a short pause, the FSB frequency SoftFSB is set to the values we want, and at the bottom shows the frequency of the current internal CPU. Next we can continue to the FSB frequency 166MHz or even higher. Remember that the best time to increase the level. Let the hardware has a process of adaptation, instead of a higher level. After setting up a successful first run some of the commonly used software, feel quite normal after transfer to higher level . If the number of illegal operations there than usual, then the help system in the current FSB frequency has been less stable, continue to increase in significance as insignificant. .
If you press the SetFSB, see a black screen of death or showing blue white alert information is in the FSB frequency does not work, you only need to restart the machine, the system will return to the original FSB frequency. In rare exceptional cases, if the frequency is too high, system restart may be a black screen, you need to clear the CMOS with information that can have the system back to normal.
System can work to find the highest stable FSB frequency, we can run automatically every time SoftFSB set the frequency, methods, according to TaskTray, in the dialog box shown in the frequency of the target set AutoStart FSB frequency, and select Enablesetting. If you also want SoftFSB frequencies in the exit to switch back to normal again, then set in AutoExit normal, and select Enablesetting.EnableTaskTray SoftFSB is specified in the next time you start does not display the main window, but rather in the taskbar's notification area by adding a small icon, right-click the icon, it will show the menu, we will be able to switch directly to the FSB frequency.
Advanced Options. .
SoftFSB also provides several advanced options, which specifies to use the old OldAccesscontrol access time clock chip. Since 1. .7 Edition introduces a new access method, if some motherboards in version 1. .6 can change the FSB frequency, 1. Edition cannot .7, you should select this option.
Option used to set the clock chip to access some of the specific options, including whether the readable chip, whether the AOpen motherboard, etc.. If we choose to overclock the motherboard is SoftFSB default type, so do not modify the options here. If we choose the clock chip type, and can not get or set the correct FSB frequency, you can try to change the option here. The five options appear there are many combinations, but only 6 in the utility usually possible.Because the ReadBack, etc., to the left of the three check boxes have a certain mutually exclusive, but are usually selected Checkcounter 0 or 1, DatabyteModification fixed to 0, so we only need to test 6 combination.
SoftFSB1. .7 The File menu also provides editorial board type and clock chip type of dialog, the dialog box in which the second relates to the clock chip's many specific parameters, not the average user can understand, so we should not use it. As for the editing board type of dialog box, if your motherboard is not in the list of default among SoftFSB, but SoftFSB support the motherboard clock chip, you can use this dialog to establish a board for your entry, visit parameter set overclocking successful value. .
Some caveats:.
For safe and effective use of SoftFSB to achieve soft overclocking, we should note the following:. .
1) some models of motherboards have a different version, such as ASUS P2B of 1.1.1 .0 and Edition. Because different versions of the motherboard may use different clock chip, so it is best to use SoftFSB before understanding the motherboard model number and version number.
2) Some of my friends use SoftFSB, first choose a motherboard model, then choose a clock chip, and then overclocking. In fact, only two clock chips motherboards and choose one. If the board a list of default in SoftFSB among the best to choose a motherboard type, this time Ruozai election time clock chip, instead it possible to flush SoftFSB default parameter values, resulting in not a normal visit to clock chip. .
3) if the motherboard uses clock chip is not supported, then SoftFSB is most likely the chip will not be able to adjust the frequency by the software, which means you cannot use SoftFSB to overclocking. These boards are usually older models, or is some brand machine OEM motherboard. In the description file SoftFSBe SoftFSB. .htm respectively gave the supported and unsupported clock chip.
4) AOpen motherboard uses a unique control method, SoftFSB not allowed to deliver the current FSB frequency, and overclocking, make sure the Option is selected UsingAOpenM / B. .
5) friends SoftFSB to the Windows Startup group, allowing Windows to run automatically at startup. The author does not recommend this practice because of the frequency change the FSB SoftFSB also changed the PCI bus frequency and Windows at boot time and hard drive read/write operation very often, if you change the frequency, there may be some unexpected problems, so it is best to finish at startup and the hard drive stops after you manually run the read and write SoftFSB.
6) the same way as ordinary overclocking, use SoftFSB over to a higher frequency can depend on the hardware, system stability, with better cooling if the measures of good quality used equipment memory and hard drive, then the probability of a successful overclocking is greatly increased. .
2. the FuzzyLogic Weixing.
MSI seems to be the first to overclocking software applications in their motherboard manufacturer, but now the software can only be used MSI motherboards and 6163-Master i820 motherboard, 6163Pro and 6309 had not even use. FuzzyLogic done a very nice interface, like a steering wheel suspended in the desktop, very avant-garde. All software is the graphical interface, first glance, I feel like can not find the direction of a closer look and found that not many adjustable options. A total of eight keys can be, of which only two and overclocking related.Min key is to be reduced to the keys on the taskbar, the Exit key to exit the L1 key is displayed in L1 L2 cache CPU, the CPU L2 button is displayed in the L2 L2 cache CPU key is displayed, the CPU of the relevant information About the key is to display version information, the remaining Auto-key and the key is to Go and overclocking related. FuzzyLogic is auto overclocking procedure, even settings are unavailable, you can run this program. It automatically detects the (Auto), repeated test out your system overclocking and stable work of the ceiling, and then click the Go button, so will not (not) overclock, but want the computer to run faster in the early scholars will enjoy the thrill of overclocking. Therefore, such a fool as FuzzyLogic overclocking software is suitable for beginners, but it seems the lack of a manual adjustment of the fun. .
Weixing new i815EPro (using the i815E chip) motherboard will also auto overclocking software FuzzyLogic improvements introduced F u z z y Ⅱ, in addition to interface more cool features all the system hardware monitoring functions are also integrated into the favorite overclocking users can make hyper-.-
3. . Gigabyte EasyTune Ⅲ. .
Gigabyte in its latest i815 series chip motherboard launched this overclocking software — E a s y T e u n Ⅲ. E a s y T e u n Ⅲ is open, what is strange, it has two modes, one is EasyMode (simple mode), one is AdvanceMode (Advanced mode). In EasyMode, when you press the Default, E a s y T e u n Ⅲ will automatically detect the most appropriate frequency, set, everything is the fool.
In AdvanceMode, you will have more manual adjustment options, you can customize the FSB, a frequency of a frequency adjustment, and key GO button, Sibu Si plane, depends on your tolerance impatience over the CPU. .
4。.BP6SFB 。.
Nothing else to do, then rolled his line, found a special overclocking software, called BP6SFB, that some foreign fans specifically for the BP6 Abit BP6 motherboard produced special edition of "SoftSFB", quite a mini-based, of course, only the BP6 Abit motherboard support functions are not many, but the others also have a function. .
If you're interested, you can also go online to find your own Board, there is no one specifically for its production of overclocking?.
5, CPUBoost. .
This software is no big SoftSFB reputation, I am also a world in a whole bunch of software to find out, download size too large, 122KB.
Unzip with WinZip, you can directly use, CPUBoost icon is a yellow crab, reminiscent of Realtek's products, they are using a similar small crabs as a trademark. I do not know whether the association between the two. .
But quite unfortunately, this software seems to be a long time are not upgraded to the current version is 1. .03, unable to identify me l technical BP6, weixing 6163Pro motherboard and metekhski SY6BA + clock chip, so it will form a clock chip to my l technical BP6, weixing 6163Pro and metekhski SY6BA +, enter the software, all options are virtual, cannot be set. At this point, CPUBoost do worse than the SoftSFB more, if SoftSFB paragraph does not recognize a clock chip or motherboard, or at least have the opportunity to download updated data packets or a packet of its own, but has little CPUBoost did not give us the opportunity, simply can not use. The software does not Readme or Help file, people are confused. .
Not used, so there's not a very good evaluation of the performance of this software. But from its menu settings, than it is simple, SoftSFB by lever through overclocking, the core and SoftSFB is through the clock chip operation, to achieve the overclocking..
Overclocking can be unsuccessful for a number of auxiliary methods to increase the success rate of OC, such as: increasing the voltage:. .
Increased voltage with a certain risk, it is recommended not to use, if you really need to increase the voltage to increase the stability of overclocking, you want to add a little bit, and monitor the temperature. For Intel CPU, a slightly higher voltage effect is obvious; for AMD CPUs, you can add some voltage. Here to mention that the motherboard to support changing the voltage, or room for overclocking. If you need a riser card, pay attention to the choice or change, you can adjust the voltage of the riser card for the best option.
Of course, after use CPU overclocking, heat will be greatly increased, this time using a good heat sink, is useful for overclocking. .
Generally speaking, overclocked CPU as long as you follow the above steps, you should be able to be overclocked to succeed, as for overclocking, it depends on your machine and the quality of the various accessories, merits attention is: overclocking will shorten the life of CPU, if you want to let the machines can now use a decade or not overclocked. But now it is a computer upgrade, excessive caution is necessary.
4. . Cool. .
Since the CPU overclocked, so temperature is sure to be a headache, how to strengthen the CPU heatsink? of course we can replace a better heatsink, the hardware where I'll give you a brief introduction about software cooling solution.
A lot of cool software commonly used, with the possible can not be too numerous, well-known are: CPUIdle, CPUKiller, Rain, WaterfallPro, softcoolerII. .
CPUIdle 。.
CPUIdle (in the DOS environment, it is called DOSIdle) is a German production of the CPU cooling software, which is published by the HLT instruction during CPU idle CPU cooling to the tools faster version updates, the latest 5. .5 Revised version of the original version of the CPU is not automatically the best optimization of the Bug. Currently popular online CPUIdle There are two versions, one is the official full version, compressed a little more about 1M, one Small version, about 400K or so, a little something to monitor the function, but also be able to cool the CPU.CPUIdle can be placed in the "start" of the group automatically, but takes up 1%-2% of system resources. CPUIdle boot interface is very clear, at first glance, a bit like EasyCD interface, full pattern, giving a warm feeling. After entering the user interface, Windows displays the area has four options, namely About, CPUInfo, Greetings, and Star-tup. Open the CPUInfo option, you can see all of the CPU information and enable the HLT. Startup is set every time you start CPUIdle settings, nothing to say, you will always tick it. .
According to author their own on the Internet, you can reduce CPU usage CPUIdle temperature 5-10 ℃. This is because, in General, regardless of whether the work of the CPU is busy, the power consumption should be basically the same, but it is used in the CPU CPUIdle relatively idle, away from development short adjustment, resulting in reduced power consumption, reduce the heat. However, it also has its own weaknesses, it is due to the use of HLT instruction, if the system is busy, it won't. Therefore, CPUIdle only in relatively idle CPU time will play the role of the more obvious. .
CPUKiller2. .0. .
This is a shared software, unpack the zip package volume only slighter 592KB, is a lightweight player. Strictly speaking, CPUKiller2. 0 is not a real CPU cooling software, why? Because it is not the use of process control CPU cooling, but to slow down and your CPU! If your CPU has been upgraded to a faster Pentium Ⅱ MMX or above, while Daozhi as CPU Sudu too fast and Shide other number Jiaolao version of Cheng Xu or the game is not available, you can use Gai program or game before the first run Ci procedures, Dang Ran, spin-down benefits is CPU temperature dropped. .
The use of CPUKiller2. .0 method is simple, start CPUKiller, you will see a P P Ⅱ, Ⅲ look logo. CPUKiller's main interface is not complicated, you might see at a glance on the right side of the screen to start the "START" button, Yes, press it. In the "START" in the bottom there is a small red arrow in the middle is a horizontal telescopic adjustment box. Pull regulator box that you want to adjust the CPU speed (now you can compare the obvious system speed slower). Click the arrow, where edges can "Add" or "Remove" to add or remove programs entries, select OK they can go, and very visual, others would not like to say instead. .
Rain 。.
Waterfall you know? Do not know, does not matter, in introducing it before you introduce the author of a Waterfall with the introduction of another type of software, although some additional functions for the time being inferior to Waterfall, but the operation is extremely simple and consumes system resources are also less, but the cooling effect is unambiguous. Do not believe? Extracted only 261KB, it small enough. After installation is complete, start the interface is relatively simple to set tends to "fool" and you just add it at the installation boot menu it is OK.Each subsequent at, the system will automatically load the Rain, it can automatically optimize your CPU temperature, cooling were set up for anything. However, say, feel at ease, so I installed WaterfallPro.
WaterfallPro. .
Waterfall but a veteran of CPU cooling software, a year ago, as a pioneer in cooling the software it is already in the public media. Today we are going to introduce to you is the next version of the Waterfall WaterfallPro, as one of the Waterfall of the series of works, in addition to providing the convenience of CPU cooling, even joined the system effectiveness monitoring, and additional functionality. I sunshine WaterfallPro is the latest version 2. .1, all installed only 1M more powerful but dapper . Start after the "waterfall" sign reminiscent of memorable, narrow place on the start bar icon like the real "waterfall" as. The new version of WaterfallPro provides three main functions: CPU temperature management, performance monitoring functions and system optimization. Unique CPUThrottle (thermostat) in the CPU temperature management, and the old crack addition to the same level of cooling efficiency Rain outside, through the unique CPUThrottle design, players can set high and low temperature warning, the duration of such additional conditions value.Preset these benefits is that once the CPU temperature reaches warning value or a burden to the CPU are a whole lot of pressure, the CPU performance began to WaterfallPro automatically "forced" to reduce the burden on the system, so the CPU temperature is reduced.
Here we look at its specific use. For such a small instrument, the usage is simple. Start WaterfallPro, the program will automatically reside background. Double click on the Waterfall Systray icon will appear WaterfallPro the program window, if the resources to monitor the figures appear more clearly explain WaterfallPro has been beating the entry into force of the machine.Open the Option option to launch the "Setup" you can make settings in the Setup project, you should, according to the steps prompted Select CPU type, motherboard type, system standards, fully determine and confirm that you want. This will restart WaterfallPro, and in accordance with the set temperature lowering CPU and crazy. According to author your own, you can eliminate the CPU WaterfallPro from 30%-70% of the heat. The program now fully supports all the popular market, CPU (Cyrix chips in the latest version can also be supported), taking up very little CPU resources, and similar well-known software CPUIdle comparison, has a small size, more powerful features. .
SoftCooler 。.
MSI (MSI) company produces the CPU cooling software, now introduced the second generation of software, not only to identify AMD, Cyrix, Intel, etc. of different brands of CPU, the CPU can also be optimized for different temperature, so that CPU time Office In the Normal state, no fever. .
。.
But more important is that it is a real fool-based software, as long as you install it without too much care, all of the tasks entrusted to it to deal with, so that your CPU time in normal operation, no longer "fever." .
S o f C o o l t e r Ⅱ all no 700K after installation. Icon for a very creative: humanization of CPU. The operation interface is very simple and does not need to do to adjust.
Starts, SoftCooler Ⅱ automatically reduced to the background and cool the CPU. After running the system will find the icon bar at the bottom right corner there is a small green icon, indicating that SoftCoolerII already running in the background of the. Is a visual display interface, the implementation of SoftCooler Ⅱ, in the status bar will appear a Icon. Right-click on the Icon that the Taskbar, right click on Properties item, the dialog box that pops up ProceeorInformation.At the same time in the following three message boxes: Currentidle (CPU currently free), Totalidle (always free), TotalidleTime [h:m: s] (for a total of spare time), and so on at a glance. The bottom-most in the interface, you can see the jump more than green words.
From here you can know the appropriate model of the motherboard and CPU, CPU the manufacturer and serial number, and the CPU core voltage, external clock, running on the CPU frequency and other details of parameters, while the lower part also provides the percentage of CPU idle. .
Through software for CPU cooling method is to compare the behavior of mercy, in a sense, even better than hardware cooling. We are interested to try it!.
Evaluation of your CPU. .
Look s models.
Different CPU, its appearance certainly different, even if some processor somewhat similar appearance, the text will also have significant identity to a distinction between them. .
IntelPentiumIII： 。.
PentiumIII into Katami core, Coppermine Tualatin core and the core of three.Which are the core of PentiumIII Katami adopts Slot1 structure, Coppermin core is by far the most common PentiumIII, the vast majority of the FC-PGA package is used, only very few use Slot1 structure, and Tualatin core PentiumIII-pin FC-pga2 package is used, with enhanced thermal cover IHS, some similar appearance and Pentium4, name also called PentiumIII-S, frequency currently only 1. .13GHz and two 1. .26GHz. .
. .
To distinguish three PentiumIII, apart from the shape actually number resolution is the most simple and easy to use. If the figure is the core piece of D0 PentiumIIICoppermine core 1GHz, the top number: 000/256/133/1 .75V. The first number is the frequency, the second is the third L2Cache capacity, is the FSB, the fourth is the default core voltage. Understanding numbers meaning and a variety of core features, can easily determine what your own PentiumIII cores product.
. .
Tidbits: Coppermine core PentiumIII500, 533, 550 and 600 four products and core have duplicate Katami, Intel also deliberately uses such as 600EB/256/133/1 .75V mark, such as the frequency increases the E and b two suffix, which together with the representatives of the Coppermine core E, and B represents the added 133MHz FSB. Each separate use of two suffixes, which may occur 600B this logo, that is, the core of 133MHz FSB processors Katami. .
IntelPentium4： 。.
Pentium4 all use the FC-PGA2 package, all IHS lid, but it also has two core, one is to have 256KBL2Cache the Willamette core, the other is owned 512KBL2Cache the Northwood core.Since the former has two Socket478 Socket423 and interface specifications, while the latter only Socket478 specification, so the appearance of the simple judgement by two core products, but we can still easily troubleshoot the processor number.
Willamette Northwood core as the core and the biggest difference is L2Cache, therefore, simply view the frequency of capacity after the L2Cache can tell the difference between them. In addition, although the frequency of Willamette and Northwood considerable overlap range, from 1. .6 GHz to 2GHz have, but the core of the Pentium4 Northwood total of between only 1. .6 GHz, 1. .8 GHz and 2. .0GHz three, but all frequency overlapping ranges in the Northwood core products are increased in frequency after A suffix (such as 1. .6A) to show the difference with the Willamette core.
. .
IntelCeleron： 。.
We are talking here mainly refers to the current Celeron popular CeleronII and CeleronIII. The former Coppermine-128, the main use of the FC-PGA package, which is Tualatin-256 core, all using the FC-PGA2 package. However, it is noteworthy that a small amount of FC-PGA2 package CeleronII are on the market, so we still have to check on the Celeron based number. .
。.
Pentium4 Celeron identification methods and identification methods have the same purpose, the simplest view the number of L2Cache capacity as long as the difference can be. In addition, with the frequency of different core Celeron, Tualatin-256 product number in the frequency of A after the suffix to distinguish, for example 1000A/256/100/1. .475 1GHz said the Tualatin-256Celeron, regardless of whether the use of FC-PGA2 package, numbered 1000/128/100/1. .Coppermine-128Celeron 7V is certainly.
. .
AMDAthlon： 。.
Compared in terms of Intel processors, AMD processors, model identification method of the simple lot. AthlonThunderbird use of ceramic packages, while the AthlonXP, use a similar package FC-PGA's OPGA, heavy ceramic package package than OPGA much weigh can be detected by hand. In addition, AthlonXP identification numbers starting from 1500 (meaning AthlonXP1500 +, clocked at 1..33GHz), and the identification of the AthlonThunderbird is the clock speed of up to 1400, look at the number you can easily identify.
AMDDuron:. .
I'm afraid all processors are the most recognisable is AMDDuron. Duron is divided into two Spitfire and Morgan, former maximum heart rate is the lowest frequency 950MHz, which precisely for 1GHz convergence, there are no duplicates, so as long as you can access code identification.
. .
Software search specifications.
. .
Depending on the processor model number to view the processor specification is naturally the barrel, you can search on the Internet, but the vast majority of cases, CPU installed in a computer is being heavily radiator cover, remove the trouble, but like Athlon, Duron, PentiumIII and Celeron such core exposed the fragility of the processor, it's all too easy to create core collapse, corners etc, cause CPU completely scrapped, in order to check the CPU specifications clearly outweigh the damage it. Can we Take it then there is no way yet? No, software engineers designed our multi-processor specifications in the detection process, so we opened them one by one the veil. .
IntelProcessorFrequencyIDUtility4. .0 SIGGRAPH.
Intel introduced naturally bear the brunt of the authority of the software: IntelProcessorFrequencyIDUtility. Currently, it's the latest version is 4. .0. IntelProcessorFrequencyIDUtility determination algorithm uses a frequency (speed test), to determine what internal speed the processor to run.Then, check the processor ProcessorFrequencyIDUtility i.e. internal data and this data and the operating frequency detected by comparing the final State of the system as a result of the comparison to notify the user. It includes "FrequencyTest" and "CPUID function" two independent label.
. .
。.
The software can detect CPU brand, is now the default frequency and frequency, and to inform whether the overclocking, and even mobile processor that can detect the information. CPUID data in part, is to find CPU type, family, model, stepping, and a series of amendments to internal information, you can also find out the L1 and L2 cache capacity. As the Intel processors produced their own testing software, so its authority is unquestionable, but also the Chinese version, can be regarded as a fool-type software. .
。.
Although this software and authority, but its function is still not strong enough, and is limited to Intel processor-testing work, we must also use other common software. .
WCPUIDVersion3。.0f 。.
WCPUID is the most prominent of several processor specification testing tool, it does not need to install and easy to use, is a truly green software. As in most cases it is the English version, so here a brief introduction about us:. .
Processor: processor model name, for example AMDAthlon IntelPentium4, etc..
Platform: platform, interface type, such as Socket370 (PGA370Socket), Socket478 (mPGA478Socket). .
VendorString: brand identifiers, such as GeniuneIntel, AMD, etc..
CPUType: CPU type, with the retail version and OEM version of the distinction between domestic and more OriginalOEMProcessor. .
Family: Celeron CPU family, is 6, PentiumIII, Athlon 7, Pentium4 belonging to 15.
Model and SteppingID: within the same small size CPU model. .
The name NameString: Pentium4 processors, and AthlonXP has a complete processor name.
InternalClock: processor, operating frequency, referred to as the frequency, the diagram of the processor is currently running at 1. .3 GHz. .
SystemClock: processor frequency, hereinafter referred to as the FSB, the processor clock speed by FSB multiplied by the multiplier.
SystemBus: system bus frequency, is the processor to memory data transfer between the frequency. .
Multiplier: processor frequency doubling, the FSB and the multiplier multiplies get processor clock speed.
L1I-Cache: an instruction cache access instruction code shown here is its capacity, speed and frequency the same. .
L1D-Cache: level data cache, the data access code, here is its capacity, speed and frequency.
L2Cache: 2 cache capacity. .
L2Speed: L2 cache running at full speed, frequency, Half Full representative on behalf of the half-speed, below the actual frequency is displayed.
. .
。.
Here are some processor features and then whether to support the list, mainly SSE, MMX and 3DNow! In addition, there are some standard features test button at the top of the StandardFeatureFlags within the meaning of complex, tedious, without much practical significance, the author describes is not much. .
。.
. .
。.
. .
There is also a relatively important system information detection is CacheInfo. Click on the button, we can see the system cache information. Figure, we can see, Fengyun Tualatin core Celeron is not only a large capacity 256KB cache, also has 8-waysetassociate, the 8-channel, and PentiumIII at the same level, while the old Celeron that only 4-waysetassociate, less a full half, which is TualatinCeleron performance than the old Celeron the main reason for a significant increase. .
Other detection software.
. .
In addition to the above two paragraphs, we can find many different kinds of processors detection software, there are more good:.
CPU-Z: the latest version is 1. .10 Version, features comparable to WCPUID comparable, but also can detect the type of processor packages, updated versions of its drawbacks is not fast enough, I use the Tualatin core Celeron and the Northwood core Pentium4, it can not correctly identified. .
MyCPU: currently the latest version of 1. .13f, function is a relatively simple design software and version is too old, the new processor is unable to respond.
TestCPU: the latest version still is 0. .96, Version too old, not properly recognize the new processor, but with some simple performance testing capabilities, but also very complete history of PC processor description - images and a very characteristic. .
Application performance measurement.
To get a feeling on the CPU only has to look appearance, you can listen to introduction, but really want to rationally understand the processor, it must actually be used to test its performance, so we know it can be doing, compatibility How to speed slow. .
。.
To conduct a simple test of the processor, the best software than SiSoftSandra Series software, and its latest version 2002. I use a SiSoftSandra2002Professional, is the professional version, as commercial software, but the Standard version is free software, free download, and test some of the CPU without any restrictions. .
。.
. .
We want to test content is shown two test projects. The former is a processor performance test in mathematics, which is the processor multimedia performance testing. SiSoftSandra design very simple calculation depending on the processor's background State, no matter whether you use what chipset, memory, video card, hard drive, processor performance negligible effect can not be more than 1% of the error, it is purely a processor performance testing. At the end of the exam, SiSsoftSandra2002 also provides some other processor reference scores on the test results for comparison.
. .
However, after all, SiSoftSandra test is a very idealistic state, and it not only eliminates the memory on the processor exclude significant influence, or even a CPU processor internal cache, especially in the cache size is not reflected in the test results, the more obvious is the test score can Celeron and PentiumIII very close, even Celeron800 scores to more than reasonable PentiumIII750, apparently.
How to look at the system cpu, memory and other utilization of the available vmstat to see the current memory, buffering and so on. Can also be used to find the current top memory or CPU resources used most. .
