APPLICATION EXAMPLES

Any memory reads using addresses 1030H- 17FFH will simultaneously cause a write to the 8276 row buffers (Fig. 4-9).

In this way, the iAPX 88 emulates DMA by addressing both the 8185s and 8276, directly transferring data from the screen memory to the 8276 row buffers. Other accesses of screen memory, such as inputting a character from the keyboard, are done using addresses between 30H and 7FFH.

Another demonstration of the power of the iAPX 88 is the routine which recognizes escape characters (Fig. 4-10).

Using the iAPX 88's Translate (XLAT) instruction and flexible addressing, this routine takes only 9 lines and 22 bytes of code. It executes in 6.6 microseconds. This same routine written for the 8085A-2 takes 20

lines, 61 bytes, and 31 microseconds. The iAPX 88 uses fewer than half the lines and bytes of code, while executing 4.7 times faster!

iAPX 88 MULTIPROCESSING SYSTEMS

Using multiple processors in medium-to- large systems offers several significant advan- tages over the centralized approach that relies on a single CPU and extremely fast memory:

1)System tasks may be allocated to special- purpose processors whose designs are opti- mized to perform specific tasks simply and efficiently.

2)Very high levels of performance can be attained when processors can execute simul- taneously (parallel/ distributed processing).

DATA BUS

DATA FLOW

8088

RDr-----------------------~~

1/2

'--74LS139

Yr-~~------------~~

ADkr. ...D_E_C_O_D_E_R...

BUS

, G 112

74LS139 o--s

0 ---

DECODER o--s

A SINGLE 8088 STRING INSTRUCTION

MOVES DATA BYTES FROM THE 8185 L----o

RAM TO THE 8276 ROW BUFFER. THE 8088 "THINKS" IT IS LOADING THE AX REGISTER.

1

BS WR

CS CRT CONTROLLER

8276

CS RD

SCREEN MEMORY

8185

Figure 4-9. 8276 Row Buffer Loading

4-16

Page 251
Image 251
Intel 210200-002 manual 8088