处理器、CPU、核的区别
2013-09-05 08:37
176 查看
Reprint:http://www.tomshardware.com/forum/342869-28-difference-processor-core
The terms 'Processor', 'Core', and 'CPU' are all poorly defined and have undergone many changes in meaning over the year as computer architectures
have evolved and changed.
Under modern terminology, 'Processor' and 'CPU' mean more or less exactly the same thing. It would be more accurate to refer to the 'processor package' because there's
no real standard to what a the package contains. Very old CPUs from 15-20 years ago contained little more than the bare minimum to execute tasks.
They contained the ALUs, fetch and decode hardware, Instruction pipeline, Interrupt handling hardware, some IO control hardware and that's about it. After this, cache
memory was added to the CPU to improve execution
Then, the execution parts of the processor were duplicated. The ALUs, fetch and decode, instruction pipline, and some cache memory were organized into what we now
call "cores". Each core is capable of functioning on its own and contain all the resources necessary to perform computational tasks that do not involve interacting with components outside the CPU. IO Control, interrupt handling, etc... were all shared between
all the cores.
More recently the memory controller itself has been moved into the CPU package. It sits along side the CPU cores but it is not part of them. Thus it is part of the
package, or part of the processor/CPU but it is not part of the 'cores'. Intel used to specifically refer to this as the 'un-core'.
This gets even more complicated when we talk about systems which have multiple physical processor packages installed. Many server and workstation platforms can have
2, 4, or even more processor packages installed. Each processor package contains the same hardware.
Thus, the number of 'cores' in a machine can be computed by taking the number of cores per package and multiplying it by the number of packages in the system. A computer
which has two quad core processors has the same number of cores as a computer which has a single octal core processor.
Here is an image of a motherboard that has 4 processor sockets
http://i1-news.softpedia-static.com/images/news2/Tyan-I...
Each one of those processor sockets (the flat ones with all the pins and metal coverings) can hold an AMD Opteron 6100 series processor. Each Opteron 6100 processor
has either 8 or 12 cores inside which means that particular system can have between 8 (a single 8-core processor installed) and 48 cores present (4 12-core processors installed).
The terms 'Processor', 'Core', and 'CPU' are all poorly defined and have undergone many changes in meaning over the year as computer architectures
have evolved and changed.
Under modern terminology, 'Processor' and 'CPU' mean more or less exactly the same thing. It would be more accurate to refer to the 'processor package' because there's
no real standard to what a the package contains. Very old CPUs from 15-20 years ago contained little more than the bare minimum to execute tasks.
They contained the ALUs, fetch and decode hardware, Instruction pipeline, Interrupt handling hardware, some IO control hardware and that's about it. After this, cache
memory was added to the CPU to improve execution
Then, the execution parts of the processor were duplicated. The ALUs, fetch and decode, instruction pipline, and some cache memory were organized into what we now
call "cores". Each core is capable of functioning on its own and contain all the resources necessary to perform computational tasks that do not involve interacting with components outside the CPU. IO Control, interrupt handling, etc... were all shared between
all the cores.
More recently the memory controller itself has been moved into the CPU package. It sits along side the CPU cores but it is not part of them. Thus it is part of the
package, or part of the processor/CPU but it is not part of the 'cores'. Intel used to specifically refer to this as the 'un-core'.
This gets even more complicated when we talk about systems which have multiple physical processor packages installed. Many server and workstation platforms can have
2, 4, or even more processor packages installed. Each processor package contains the same hardware.
Thus, the number of 'cores' in a machine can be computed by taking the number of cores per package and multiplying it by the number of packages in the system. A computer
which has two quad core processors has the same number of cores as a computer which has a single octal core processor.
Here is an image of a motherboard that has 4 processor sockets
http://i1-news.softpedia-static.com/images/news2/Tyan-I...
Each one of those processor sockets (the flat ones with all the pins and metal coverings) can hold an AMD Opteron 6100 series processor. Each Opteron 6100 processor
has either 8 or 12 cores inside which means that particular system can have between 8 (a single 8-core processor installed) and 48 cores present (4 12-core processors installed).
相关文章推荐
- 什么是低电压处理器 笔记本CPU标准与低电压区别对比
- 处理器、CPU、核的区别
- 处理器(CPU)发展简史
- CPU、Processor、Core的区别
- 微型计算机的中央处理器<CPU的发展史>
- 墙上时钟时间 ,用户cpu时间 ,系统cpu时间内,核态和用户态的区别
- 中央处理器CPU
- CPU MPU MCU SOC SOPC关系及区别
- 全虚拟化和半虚拟化的区别 cpu的ring0 ring1又是什么概念?
- ios获得设备处理器cpu类型
- 谈GPU的作用、原理及与CPU、DSP的区别
- CPU和GPU的设计区别
- 全虚拟化和半虚拟化的区别 cpu的ring0 ring1又是什么概念?
- 内核的驱动和在cpu的大小端无关,内核提供一套函数来处理大小端的区别
- 多核与多个CPU啥区别?
- ps 和 top cpu 占用区别
- 超线程 多核 多处理器区别
- 处理器(CPU)调度问题
- 05.SOC和CPU的区别
- 深入理解计算机系统之旅(四)处理器(CPU)的体系结构