LSI and Microprogramming

Microprogramming concepts were well known before LSI, however, with LSI, the impact of microprogramming techniques on computer architecture will increaes. Microprogramming was not used much until the advent of relatively low-cost ROS. It is already used extensively in many systems, including System/360 and System/370 for control store and emulation purposes.

大規(guī)模集成電路和微程序設(shè)計

雖然在大規(guī)模集成電路出現(xiàn)之前，微程序設(shè)計的概念已經(jīng)是眾所周知的了，但是，有了大規(guī)模集成電路，微程序設(shè)計技術(shù)對于計算機結(jié)構(gòu)格式的影響將與日俱增。直到出現(xiàn)了成本比較低的只讀存儲器（ROS)之后微程序設(shè)計才用得多起來。現(xiàn)在它已廣泛應用于許多計算機系統(tǒng)中，其中包括用于控制存儲和仿真的360和370系統(tǒng)。

Well publicized attributes of microprogramming include ease of system design and change, greater flexibility and versatility, wider functional capability and improved performance in certain important areas. In addition, a system organization using microprogramming usually results in a more regular and repetitive hardware design (more than just the storage arrays) which is more amenable to LSI implementation.

LSI projections offer ROS at costs far below any here-to-date --- but more important, these projections can offer highspeed writable control stores (WCS) via RAM semiconductor memory at sufficiently low costs, as well. 

廣為宜傳的微程序設(shè)計的屬性包括以下幾點：容易進行系統(tǒng)設(shè)計與變換、靈活性與通用性更強、功能更齊全以及在某些重要的領(lǐng)域性能有所改善等。加之，使用微程序設(shè)計的系統(tǒng)結(jié)構(gòu)通常會導致硬件設(shè)計更為整齊劃一、規(guī)格一致（不僅僅是存儲陣列)，這樣就能更適合于實現(xiàn)大規(guī)模集成電路。

利用大規(guī)模集成電路的工程設(shè)計能以大大低于至今任何一種存儲器的成本造出只讀存儲器（ROS)——更為重要的是，利用這些工程設(shè)計還能以足夠低的成本通過半導體隨機存儲器（RAM)造出高速可寫控制存儲器(wCs)。

It is the LSI-afforded, dynamically writable control store which portends the most significant impact on future architecture --- and accounts for much of the growing current interest in "dynamic microprogramming". Advocates of microprogramming point to excessively complex software, the logical primitiveness of present computer hardware, and the large gap between machine and programming languages in present-day systems. The employment of microprogramming-enhanced design principles, problem-oriented (macro-oriented) instruction sets, and more complex machine language primitives, respectively, are suggested for the above problems. There are examples where factors of 20 and greater savings were achieved using microprogrammed problem-oriented macros, as compared with a typical general-purpose processor.

正是由大規(guī)模集成電路所提供的動態(tài)可寫控制存儲器才預示著它對未來的計算機格式會產(chǎn)生最重要的影響，也才說明了目前對于“動態(tài)微程序設(shè)計”的興趣日益增加的主要原因.提倡微程序設(shè)計的人認為：軟件過于復雜：目前計算機硬件在邏輯上較為原始；在現(xiàn)代系統(tǒng)中機器語盲與程序語言之間存在著很大的鴻溝。針對上述問題.他們提出分別使用由微程序設(shè)計所啟發(fā)的設(shè)計原理、面向問題的（宏面向的）指令組以及更復雜的機器語言的原語等來解決。有一些使用了微程序設(shè)計的面向問題的宏指令的例子，它們同典型的通用處理機比較起來，成本就只有原來的二十分之一或更低。

Microprogramming techniques, exploiting the less sensitive cost-capacity relationships of semiconductor memory, facilitate conisideration of decentralized control. This allows a more compressed microcode and its execution in paralleled or pipelined fashion. Further, it is usually easier to design, analyze, and maintain smaller, decentralized programs. 

Opler projected the notion of a microprogramming region between software and hardware, dubbed "firmware," for fourthgeneration computers many years ago. Microprogramming techniques are presently being used in the industry in the areas of memory and file management, resource allocation, interrupt handling, recovery, and communication between processors in multiprocessing configurations. 

微程序設(shè)計技術(shù)由于利用了對成本-容量關(guān)系不太敏感的半導體存儲器，所以便利了對局部控制的考慮。這就可能創(chuàng)造更為壓縮的微程序代碼并有可能以并行或流水線的方式來執(zhí)行這種微程序代碼。此外，設(shè)計、分析和維護比較短的局部程序通常也都比較容易。

許多年前，奧普勒就為第四代計算機提出了關(guān)于在軟件與硬件之間建立叫做“固件”的微程序設(shè)計區(qū)域的概念。現(xiàn)在，微程序設(shè)計技術(shù)正在工業(yè)上用于如下幾個領(lǐng)域：存儲器管理與文件管理、資源分配、中斷處理、復原以及在多邊處理配置中處理機之間的通信等。

責任編輯：admin