并行計算機(jī)架構(gòu)

并行處理 (Parallel Processing)

Parallel processing is processing of the data concurrently. We process the data concurrently to fulfill the demands of the increasingly high performance so that to achieve better throughput instead of processing each instruction sequentially as in a conventional computer.

并行處理是同時處理數(shù)據(jù)。 我們同時處理數(shù)據(jù)以滿足日益增長的高性能的需求，從而實現(xiàn)更好的吞吐量，而不是像傳統(tǒng)計算機(jī)那樣順序處理每個指令。

We can achieve parallelism in two ways:

我們可以通過兩種方式實現(xiàn)并行性：

1)多功能單元 (1) Multiple functional units)

These systems have two or more ALUs so two or more instruction can be executed at the same time.

這些系統(tǒng)具有兩個或多個ALU，因此可以同時執(zhí)行兩個或多個指令。

2)多個處理器 (2) Multiple processors)

These systems have two or more processors operating concurrently.

這些系統(tǒng)具有兩個或多個同時運(yùn)行的處理器。

The earlier processors had only one arithmetic and logic unit in its CPU. Furthermore, the ALU could only perform one function at a time, so that the process for executing a long sequence of arithmetic and logic instruction is quite slow. But nowadays the processors are available with multiple functional units, these multiple functional units used to execute the function of ALU which can be distributed and operate in parallel so that the speed of the processor's increases.

較早的處理器在其CPU中只有一個算術(shù)和邏輯單元。 此外，ALU一次只能執(zhí)行一個功能，因此執(zhí)行較長的算術(shù)和邏輯指令序列的過程非常緩慢。 但是如今，處理器具有多個功能單元，這些多個功能單元用于執(zhí)行ALU的功能，這些功能單元可以分布并并行運(yùn)行，從而提高了處理器的速度。

Micheal J.Flynn made an informal and widely used classification of parallelism processor in 1966, which is based on the number of simultaneous instruction and data streams seen by the processor during program execution.

Micheal J.Flynn在1966年對并行處理器進(jìn)行了非正式且廣泛使用的分類，該分類基于處理器在程序執(zhí)行期間看到的同時指令和數(shù)據(jù)流的數(shù)量。

并行處理的分類 (Classifications of Parallel processing)

The classification divides computers into four major groups:

分類將計算機(jī)分為四個主要組：

Single instruction stream - single data stream

單指令流-單數(shù)據(jù)流

Single instruction stream - multiple data streams

單指令流-多個數(shù)據(jù)流

Multiple instruction streams - single data stream

多個指令流-單個數(shù)據(jù)流

Multiple instruction streams - multiple data streams

多個指令流-多個數(shù)據(jù)流

1)單個數(shù)據(jù)流-單個數(shù)據(jù)流(SISD) (1) Single data stream - single data stream (SISD))

Most conventional machines with one CPU containing a single arithmetic-logic unit capable of only scalar arithmetic fall into this category. They may have more than one functional unit, but all functional units are controlled by a single control unit.

具有一個CPU且僅包含一個僅能進(jìn)行標(biāo)量算術(shù)的單個算術(shù)邏輯單元的大多數(shù)傳統(tǒng)機(jī)器就屬于此類。 它們可能具有多個功能單元，但是所有功能單元都由一個控制單元控制。

2)單指令流–多個數(shù)據(jù)流(SIMD) (2) Single instruction stream – multiple data stream (SIMD))

SIMD has multiple processing units and one control unit. They correspond to array processors. In SIMD, all the processing elements received the same instruction from the control unit but operate on different data sets from distinct data streams.

SIMD具有多個處理單元和一個控制單元。 它們對應(yīng)于陣列處理器。 在SIMD中，所有處理元件都從控制單元接收到相同的指令，但是對來自不同數(shù)據(jù)流的不同數(shù)據(jù)集進(jìn)行操作。

3)多指令流單數(shù)據(jù)流(MISD) (3) Multiple instruction stream single data stream (MISD))

In MISD, there are n processor units each receiving distinct instruction operating over the same data stream and its derivatives. This means one processor becomes the input of the next processor.

在MISD中，有n個處理器單元，每個處理器單元都接收對同一數(shù)據(jù)流及其派生的不同指令。 這意味著一個處理器成為下一處理器的輸入。

4)多指令流多數(shù)據(jù)流(MIMD) (4) Multiple instruction stream multiple data stream (MIMD))

In MIMD, there is more than one processor unit having the ability to execute several programs simultaneously. This category involves most of the multiprocessors system and multiple computers systems.

在MIMD中，有不止一個處理器單元能夠同時執(zhí)行多個程序。 此類別涉及大多數(shù)多處理器系統(tǒng)和多計算機(jī)系統(tǒng)。

翻譯自: https://www.includehelp.com/cso/parallel-processing.aspx

并行計算機(jī)架構(gòu)

總結(jié)

以上是生活随笔為你收集整理的并行计算机架构_计算机科学组织| 并行处理的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。