Exception Handling

1)
2)
3)
4)
5)
6)

The 8087 detects six different types of exception

conditions that occur during instruction execution. These
will cause an interrupt if unmasked and interrupts are
enabled.
INVALID OPERATION
OVERFLOW
ZERO DIVISOR
UNDERFLOW
DENORMALIZED OPERAND
INEXACT RESULT

M. Krishna Kumar

MM/M4/LU11/V1/2004

Data Types
Internally, all data operands are converted to the 80-bit
temporary real format.
We have 3 types.
Integer data type
Packed BCD data type
Real data type

M. Krishna Kumar

MM/M4/LU11/V1/2004

Coprocessor data types

Coprocessor Data Types

Integer
Word

M. Krishna Kumar

Short

Packed BCD
Long

Real
Short

MM/M4/LU11/V1/2004

Long

Temporary

Integer Data Type

Word integer
S

2 bytes

Magnitude
0

15

Short integer
S

4 bytes
Magnitude

31

Long integer
S

8 bytes
Magnitude

63
M. Krishna Kumar

0
MM/M4/LU11/V1/2004

Packed BCD
Packed BCD
S 0
79 78

10 bytes

d17

d1 d0

72

M. Krishna Kumar

MM/M4/LU11/V1/2004

Real data type

Short real
S
31 30

4 bytes

8 bytes
E

79

F
0

51

Temporary real
78

M. Krishna Kumar

= 4066540000000000h

10 bytes

= 4332A000h

23

Long real
S
63 62

178.625 decimal

F
63

= 4006B2A0000000000000h

MM/M4/LU11/V1/2004

Example
Converting a decimal number into a Floating-point
number.
1) Converting the decimal number into binary form.
2) Normalize the binary number
3) Calculate the biased exponent.
4) Store the number in the floating-point format.

M. Krishna Kumar

MM/M4/LU11/V1/2004

Example
Step
1
2
3
4

In step 3 the biased

exponent is the exponent a
26 or 110,plus a bias of
01111111(7FH) ,single
100.25
1100100.01 = 1.10010001 * 26 precision no use 7F and
double precision no use
110+01111111=10000101
3FFFH.
Sign = 0
IN step 4 the information
Exponent =10000101
found in prior step is
combined to form the
Significand =
floating point no.
10010001000000000000000

M. Krishna Kumar

Result

MM/M4/LU11/V1/2004

Instruction Set
The 8087 instruction mnemonics begins with the letter F
which stands for Floating point and distinguishes from
8086.
These are grouped into Four functional groups.
The 8087 detects an error condition usually called an
exception when it executing an instruction it will set the bit
in its Status register.

M. Krishna Kumar

MM/M4/LU11/V1/2004

Types
I.
II.
III.
IV.

DATA TRANSFER INSTRUCTIONS.

ARITHMETIC INSTRUCTIONS.
COMPARE INSTRUCTIONS.
TRANSCENDENTAL INSTRUCTIONS.
(Trigonometric and Exponential)

M. Krishna Kumar

MM/M4/LU11/V1/2004

10

I.

Data Transfers Instructions

REAL TRANSFER
FLD
Load real
FST
Store real
FSTP
Store real and pop
FXCH
Exchange registers
INTEGER TRANSFER
FILD
Load integer
FIST
Store integer
FISTP
Store integer and pop

M. Krishna Kumar

MM/M4/LU11/V1/2004

11

I.

Data Transfers Instructions

(contd..)

PACKED DECIMAL TRANSFER(BCD)

FBLD
Load BCD
FBSTP
Store BCD and pop

M. Krishna Kumar

MM/M4/LU11/V1/2004

12

Example

FLD Source- Decrements the stack pointer by one and copies

a real number from a stack element or memory location to the
new ST.
FLD
ST(3)
;Copies ST(3) to ST.
FLD
LONG_REAL[BX] ;Number from memory
;copied to ST.
FLD Destination- Copies ST to a specified stack position or
to a specified memory location .
FST
ST(2)
;Copies ST to ST(2),and
;increment stack pointer.
FST
SHORT_REAL[BX] ;Copy ST to a memory at a
;SHORT_REAL[BX]

M. Krishna Kumar

MM/M4/LU11/V1/2004

13

Example (contd..)
FXCH Destination Exchange the contents of ST with
the contents of a specified stack element.
FXCH
ST(5) ;Swap ST and ST(5)
FILD Source Integer load. Convert integer number
from memory to temporary-real format and push on 8087
stack.
FILD
DWORD PTR[BX] ;Short integer from
memory at
; [BX].
FIST Destination- Integer store. Convert number from
ST to integer and copy to memory.
FIST
LONG_INT ;ST to memory locations named
;LONG_INT.
M. Krishna Kumar

MM/M4/LU11/V1/2004

14

Example (contd..)
FISTP Destination-Integer store and pop. Identical to FIST
except that stack pointer is incremented after copy.
FBLD Source- Convert BCD number from memory to
temporary- real format and push on top of 8087 stack.

M. Krishna Kumar

MM/M4/LU11/V1/2004

15

II.

Arithmetic Instructions.

Four basic arithmetic functions:

Addition, Subtraction, Multiplication, and
Division.
Addition
FADD
Add real
FADDP
Add real and pop
FIADD
Add integer

M. Krishna Kumar

MM/M4/LU11/V1/2004

16

II.

Arithmetic Instructions (contd..)

Subtraction
FSUB
FSUBP
FISUB
FSUBR
FSUBRP
FISUBR

M. Krishna Kumar

Subtract real
Subtract real and pop
Subtract integer
Subtract real reversed
Subtract real and pop
Subtract integer reversed

MM/M4/LU11/V1/2004

17

II.

Arithmetic Instructions (contd..)

Multiplication
FMUL
FMULP
FIMUL

M. Krishna Kumar

Multiply real
Multiply real and pop
Multiply integer

MM/M4/LU11/V1/2004

18

II.

Arithmetic Instructions (contd..)

Division
FDIV
FDIVP
FIDIV
FDIVR
FDIVRP
FIDIVR

M. Krishna Kumar

Division real
Division real and pop
Division integer
Division real reversed
Division real reversed and pop
Division integer reversed

MM/M4/LU11/V1/2004

19

II.

Arithmetic Instructions (contd..)

Advanced
FABS
FCHS
FPREM
FPRNDINT
FSCALE
FSQRT
FXTRACT

M. Krishna Kumar

Absolute value
Change sign
Partial remainder
Round to integer
Scale
Square root
Extract exponent and mantissa.

MM/M4/LU11/V1/2004

20

Example
FADD Add real from specified source to specified destination
Source can be a stack or memory location. Destination must be
a stack element. If no source or destination is specified, then ST
is added to ST(1) and stack pointer is incremented so that the
result of addition is at ST.
FADD
ST(3), ST
;Add ST to ST(3), result in ST(3)
FADD
ST,ST(4)
;Add ST(4) to ST, result in ST.
FADD
;ST + ST(1), pop stack result at ST
FADDP
ST(1)
;Add ST(1) to ST. Increment stack
;pointer so ST(1) become ST.
FIADD
Car_Sold
;Integer number from memory + ST
M. Krishna Kumar

MM/M4/LU11/V1/2004

21

Example (contd..)
FSUB - Subtract the real number at the specified source from
the real number at the specified destination and put the result in
the specified destination.
FSUB
ST(2), ST
;ST(2)=ST(2) ST.
FSUB
Rate
;ST=ST real no from memory.
FSUB
;ST=( ST(1) ST)
FSUBP - Subtract ST from specified stack element and put
result in specified stack element .Then increment the pointer by
one.
FSUBP
ST(1)
;ST(1)-ST. ST(1) becomes new ST
FISUB Integer from memory subtracted from ST, result in
ST.
FISUB
Cars_Sold
;ST becomes ST integer from
;memory
M. Krishna Kumar

MM/M4/LU11/V1/2004

22

III.
Comparison
FCOM
FCOMP
FCOMPP
FICOM
FICOMP
FTST
FXAM

M. Krishna Kumar

Compare Instructions
Compare real
Compare real and pop
Compare real and pop twice
Compare integer
Compare integer and pop
Test ST against +0.0
Examine ST

MM/M4/LU11/V1/2004

23

IV.

Transcendental Instruction

Transcendental
FPTAN
FPATAN
F2XM1
FYL2X
FYL2XP1

M. Krishna Kumar

Partial tangent
Partial arctangent
2x - 1
Y log2X
Y log2(X+1)

MM/M4/LU11/V1/2004

24

Example
FPTAN Compute the values for a ratio of Y/X for an angle
in ST. The angle must be in radians, and the angle must be in
the range of 0 < angle < /4.
F2XM1 Compute Y=2x-1 for an X value in ST. The result Y
replaces X in ST. X must be in the range 0X0.5.
FYL2X - Calculate Y(LOG2X).X must be in the range of
0 < X < any Y must be in the range -<Y<+.
FYL2XP1 Compute the function Y(LOG2(X+1)).This
instruction is almost identical to FYL2X except that it gives
more accurate results when compute log of a number very
close to one.

M. Krishna Kumar

MM/M4/LU11/V1/2004

25

Constant Instructions
Load Constant Instruction
FLDZ
FLDI
FLDPI
FLDL2T
FLDL2E
FLDLG2
FLDLN2

M. Krishna Kumar

MM/M4/LU11/V1/2004

Load +0.0
Load+1.0
Load
Load log210
Load log2e
Load log102
Load loge2

26

Algorithm
To calculate x to the power of y

Load base, power.

Compute (y )*( log2 x)
Separate integer(i) ,fraction(f) of a real number
Divide fraction (f) by 2
Compute (2 f/2) * ( 2f/2)
xy = (2x) * (2y )

M. Krishna Kumar

MM/M4/LU11/V1/2004

27

Program
Program to calculate x to the power of y
.MODEL SMALL
.DATA
x
Dq
4.567 ;Base
y
Dq
2.759 ;Power
temp
DD
temp1
DD
temp2
DD
;final real result
tempint
DD
tempint1 DD
;final integer result
two
DW
diff
DD
trunc_cw DW 0fffh
M. Krishna Kumar

MM/M4/LU11/V1/2004

28

Program (contd..)
start:

load:
comput:

M. Krishna Kumar

.STACK 100h
.CODE
mov ax,@DATA
mov ds,ax
fld y
fld x
fyl2x
fst temp

;init data segment

;load the power

;load the base
;compute (y * log2(x))
;save the temp result

MM/M4/LU11/V1/2004

29

Program (contd..)
trunc:

fyl2x

fldcw trunc_cw
frndint
fld temp

;set truncation command

;load real number of

fist tempint

getfrac:

M. Krishna Kumar

;save integer after

;truncation
fld temp
;load the real number
fisub tempint ;subtract the integer
fst diff
;store the fraction

MM/M4/LU11/V1/2004

30

Program (contd..)
fracby2:
twopwrx:
result
result

M. Krishna Kumar

fidiv two
f2xm1
fst temp1

;divide the fraction by 2

;calculate the 2 to the
;power fraction
;minus 1 and save the

fld1
fadd

;load1
;add 1 to the previous

fst temp1

;save the result

MM/M4/LU11/V1/2004

31

Program (contd..)
sqfrac:

M. Krishna Kumar

fmul st(0),st(0) ;square the result as fraction

fst temp1
;was halved and save the
;result
fild tempint
;save the integer portion
fxch
;interchange the integer
;and power of fraction.

MM/M4/LU11/V1/2004

32

Program (contd..)
scale:

fscale
fst temp2
fist tempint1

over:

M. Krishna Kumar

mov ax,4c00h
int 21h
end start

;scale the result in real and

;integer
;in st(1) and store
;save the final result in real and
;integer
;exit to dos

MM/M4/LU11/V1/2004

33