MC68882EI40A Freescale, MC68882EI40A Datasheet - Page 7
![](/photos/31/37/313790/68pin_plcc_sml.jpg)
MC68882EI40A
Manufacturer Part Number
MC68882EI40A
Description
Manufacturer
Freescale
Datasheet
1.MC68882EI40A.pdf
(26 pages)
Specifications of MC68882EI40A
Operating Temperature (min)
0C
Operating Temperature (max)
70C
Operating Temperature Classification
Commercial
Mounting
Surface Mount
Lead Free Status / RoHS Status
Compliant
Available stocks
Company
Part Number
Manufacturer
Quantity
Price
Company:
Part Number:
MC68882EI40A
Manufacturer:
Freescale Semiconductor
Quantity:
10 000
5%
-’<
:<21
,, .. .
..-Y
~r.:
\
.,-,
... ..,..
t&&~81
MC=
BR~/Rov.
tions of the MC68881, The FMOVE instructions
~&’’~PU
the calc(~lation.
tio;s.
the MC68882
ecute twice as fast as the corresponding
the null state frame re-establishes
the MC68882’S
along wit~~@~~@U, can initiate
extq,~$~~,.,$rmat
ins~s%s~ution
n~xt instruction.
ring delav
binary
when the FSAVE instruction
ferred to memory.
time than it takes to save the larger state in mid-instrucl~c?~~~c, “ &~ integer
tion are allowed
to a minimum.
sizes greatly reduces the average conte”xt switch~~g~tie.
ternal state that was saved earlier and co~~~q.~
eration that was previously
a function
MC68882 PERFORMANCE
instructions
rency is to utilize$~~~~rithmetic
more efficient~,{b~~’@ &creasing
necesWW
the MC68882 through
tions.
and/or a task time out, the MC68020 or MC68030 can force
the MC68882 to stop whatever
at any time and save the MC68882 internal
ory, During the execution
the MC68882
as at the completion
dependent
of state is transferred
ined by the operating
coprocessor
is received,
state, it may be necessary to save the entire internal state
of the machine,
idle state, Thus, the size of the saved internal state is kepl
the logical
the MC68882 is-always
ory management
FSAVE is executed,
processor
bus cycles (to and from CPU space) and the MC68882 is
never a bus master, the MC68882 can be placed on either
agement
memory
MC68030 is supported
Another
The high performanceKo$:~h$iMC68882
When th~;~,t*82
The FRESTORE instruction
The size of the saved internal
Since the coprocessor
The virtual
If the MC68020
real data formats.
,.
.
is busy, the MC68882 CU can proceed
‘
management
can only instruct
unit in an MC68020-based
3
,,,
due to data movement
~erands,
factor
identical
or physical
upon the state of the APU at the time the
only a few words of internal
programmer’s
is in the idle state when the save instruction
concu,$$ent?~gF~he direct
., --------
is the oDtimized
can stop at predetermined
machine
-
of a previous
The ability to utilize several inter~~y, state
abilty $* ~~cute
it
Instructions
When the APU is finally ready to perform
to complete
even though
in obtaining
can dn sn imm~diatelv
unit in an MC68030 system.
If executing
to the MC68w
If the MC68882 is in the reset state
of the instruction.
and convert
the FSAVE and FRESTORE instruc-
to memory,
system to determine
side of the system memory
or MC68030
receives an instruction,
by the coprocessor
on the physical side of the mern~
unit of the MC68030
architecture
These FMOVE instructions
--
interface
of a floating-point
~~@@EMENTS
~:\.,
the main processor
suspende~.
completing
execution
!...,,,.
is received,
model
permits
increased
instruction.
Freescale Semiconductor, Inc.
the APU is busy complet-
an instruction
its idle time.
,\L: ~, ..::,** \ ~\>
the instruction,
FMOVE instructions
operation
state of the MC68882 is
For More Information On This Product,
Processing
d~~b$~@pister
multiple
and preparation
which
them to the internal
,,
is based solely
,.
detects
~rdware
: ~$~.
of the MC68020
is empty,
reloa~i’k~~k%n
system.
result
and then save the ‘I+i,
~~,~&~STORE
execution
FMOVE instruc-
performance
state are trans-
only one word
is the result of
without
may be exam-
points
Although
is in progress
floating-point
Go to: www.freescale.com
state in mem-
intetiace
a page fault
that the co-
instruction,
in the busy
Unit (APU)
of concur-
is on-chi~,
reset,
Since the
to wait if
fetch the
the BIU,
with the
are also
any op-
as well
values,
If the
in less
incur-
,.
func-
upon
man-
and
the
ex-
for
in-
in
of
or
“’” ‘“’’’’’’’tiger is automatically
, $,sd$ported
,,,~~oral. are the standard
The exponent
quire normalized
giving
ante with the IEEE standard,
specify
single and double
MC68882 consists of 96 bits (three long words),
bits are actually
than the equivalent
point formats
involving
mantissa
magnitude
mantissa
fixes
extended
the number
operations
sentation
FLOATING-POINT
bits) and double
IEEE standard.
The capital
source spW~l@~~fhe
lNTE~~:~A$A
used. For example,
(FP3), the following
ogs are designed
subsequent
decrease the effective
potentially
pletely
struction.
interface
has data formats
~~e
The extended
The ability
The floating-point
The MC68882 also has a more optimized
(The Motorola
Double
Extended
Packed Decimal”t~$@~’Real
added .~d}$~:troctions
one extra bit of precision,
t~be
this format
,$ $*,,,
executed
size for single, double, and extended
than the MC68881. If an arithmetic
*“
ofa number,
is implied
real numbers,
precision
in the M68000 Family architecture.
is used in a floating-point
form.
saves user memory
fully concurrent
letter<~$$~%ined
Precision
denote immediate
integer
instructions,
contained
to effectively
.~y.
and should
is biased, and the mantissa
Precisi~~,~<$PIX)
These data formats
precision
FORMATS
numbers,
used; the other
of Single, Double or Extended,
precision
during
Since single
.
precision.
to increase the potential
DATA FORMATS
assembler
to the bit level whereas
to add an integer
floating-point
FADD.W #5,FP3
floating-point
instruction
data formats
as a one and is ndt included,
data formats,
.-
Rga~~~~N’
data format
if representable,
instruction
converted
in ‘floating-point
twos complement
Table 1 lists the exponent
the execution
data format
This overlap
be used for most calculations
the most-significant
use integers
(64-bits),
and, therefore,
The memory
in the assembly
in parentheses
but the standard
syntax “#”
and double
addressing.)
can be used:
(P)
since an integer
by the MC68882 to an
execution
16 bits are for future
representation.
(byte, word,
to be used.
number
are the main floating-
single
are defined
is also in conform-
constant
,,
. .. ... ..
operation,
is usually smaller
of a previous
can significantly
in floating-point
data “register
format
is used to
precision
is in sign and
precision
data formats
before
overlap
can be com-
,,.
coprocessor
time.
,,
denote
it does for
MOTOROU
.
instruction
Whenever
precision,
does not
language
and long
of five to
bit of the
Only 80
the dial-
on the
by the
the in-
repre-
being
thus
with
(32-
and
suf-
re-
in-
7
3
,,.
.