PDSP16116 [ZARLINK]
ALU and Barrel Shifter; ALU和桶式移位器型号: | PDSP16116 |
厂家: | ZARLINK SEMICONDUCTOR INC |
描述: | ALU and Barrel Shifter |
文件: | 总16页 (文件大小:121K) |
中文: | 中文翻译 | 下载: | 下载PDF数据表文档文件 |
PDSP1601/PDSP1601A
ALU and Barrel Shifter
DS3705
ISSUE 3.0
November 1998
The PDSP1601 is a high performance 16-bit arithmetic
logic unit with an independent on-chip 16-bit barrel shifter.
The PDSP1601A has two operating modes giving 20MHz or
10MHz register-to-register transfer rates.
The PDSP1601 supports Multicycle multiprecision
operation. This allows a single device to operate at 20MHz for
16-bitfields, 10MHzfor32-bitfieldsand5MHzfor64-bitfields.
ThePDSP1601canalsobecascadedtoproducewiderwords
at the 20MHz rate using the Carry Out and Carry In pins. The
Barrel Shifter is also capable of extension, for example the
PDSP1601 can used to select a 16-bit field from a 32-bit input
in 100ns.
PIN 1A INDEX MARK
ON TOP SURFACE
A
B
C
D
E
F
G
H
J
K
L
11 10
9
8
7
6
5
4
3
2
1
AC84
FEATURES
■
■
■
■
■
16-bit, 32 instruction 20MHz ALU
16-bit, 20MHz Logical, Arithmetic or Barrel Shifter
Independent ALU and Shifter Operation
4 x 16-bit On Chip Scratchpad Registers
Multiprecision Operation; e.g. 200ns 64-bit
Accumulate
GC100
■
Three Port Structure with Three Internal Feedback
Paths Eliminates I/O Bottlenecks
■
■
■
Block Floating Point Support
300mW Maximum Power Dissipation
84-pin Pin Grid Array or 84 Contact LCC Packages
or 100 pin Ceramic Quad Flat Pack
Fig.1 Pin connections - bottom view
APPLICATIONS
ORDERING INFORMATION
■
■
■
■
■
Digital Signal Processing
Array Processing
Graphics
Database Addressing
High Speed Arithmetic Processors
PDSP1601 MC GGCR
10MHz MIL883 Screened -
QFP package
PDSP1601A BO AC
20MHz Industrial - PGA
package
N.B
Further details of the Military grade part are
available in a separate datasheet (DS3763)
ASSOCIATED PRODUCTS
PDSP16112 Complex Multiplier
PDSP16116 16 x 16 Complex Multiplier
PDSP16318 Complex Accumulator
PDSP16330 Pythagoras Processor
PDSP1601/PDSP1601A
PIN DESCRIPTION
AC pin
AC pin
Function
Function
Function
Function AC pin
AC pin
F9
F11
E11
E10
E9
D11
D10
C11
B11
C10
A11
B10
B9
A10
A9
B8
A8
B6
B7
A7
C7
C6
A6
A5
B5
C5
A4
B4
A3
A2
B3
A1
B2
C2
B1
C1
D2
D1
E3
E2
E1
F1
IS0
IS1
IS2
GND
C8
C9
C10
C11
C12
C13
C14
C15
IA4
MSB
MSS
B15
B14
B13
B12
B11
B10
B9
B8
B7
B6
B5
B4
B3
B2
B1
GND
MSA0
MSA1
A15
A14
A13
A12
A11
A10
A9
A8
A7
A6
A5
A4
A3
A2
A1
J6
J7
L7
K7
L6
L8
K8
L9
L10
K9
L11
K10
J10
K11
J11
H10
H11
F10
G10
G11
G9
F3
G3
G1
G2
F1
H1
H2
J1
IS3
SV0
SV1
SV2
SV3
K1
J2
SVOE
RS0
RS1
VCC
RS2
C0
C1
C2
C3
C4
OE
BFP
L1
K2
K3
L2
L3
K4
L4
J5
VCC
CO
RA0
RA1
RA2
CI
IA0
IA1
IA2
IA3
B0
C5
C6
C7
A0
K5
L5
K6
CEB
CLK
CEA
MSC
GC
GC
SIG
GC
SIG
SIG
GC
SIG
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
N/C
N/C
N/C
N/C
VCC
RS2
C0
C1
C2
C3
C4
C5
C6
C7
GND
C8
1
2
3
4
5
6
7
8
N/C
N/C
N/C
N/C
VCC
C0
RA0
RA1
RA2
CI
N/C
N/C
N/C
N/C
B7
B6
B5
B4
B3
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
N/C
N/C
N/C
N/C
A7
A6
A5
A4
A3
A2
A1
A0
CEA
MSC
IS0
IS1
IS2
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
B2
B1
B0
IA0
IA1
IA2
IA3
CEB
CLK
GND
MSA0
MSA1
A15
A14
A13
A12
A11
A10
A9
IA4
MSB
MSS
B15
B14
B13
B12
B11
B10
B9
C9
C10
C11
C12
C13
C14
C15
OE
IS3
SV0
SV1
SV2
SV3
SVOE
RS0
RS1
BFP
B8
A8
N/C = not connected - leave open circuit
All GND and VDD pin must be used
2
PDSP1601/PDSP1601A
PIN DESCRIPTIONS
Symbol
Description
ALU B-input multiplexer select control.1 This input is latched internally on the rising edge
of CLK.
MSB
Shifter Input multiplexer select control.1 This input is latched internally on the rising edge
of CLK.
MSS
B15 - B0
B Port data input. Data presented to this port is latched into the input register on the rising
edge of CLK. B15 is the MSB.
Clock enable, B Port input register. When low the clock to this register is enabled.
CEB
CLK
Common clock to all internal registered elements. All registers are loaded, and outputs
change on the rising edge of CLK.
ALU A-input multiplexer select control.1 These inputs are latched internally on the rising
MSA0 - MSA1
A15 - A0
edge of CLK.
A Port data input. Data presented to this port is latched into the input register on the rising
edge of CLK. A15 is the MSB.
Clock enable, A Port input register. When low the clock to this register is enabled.
CEA
C-Port multiplexer select control.1 This input is latched internally on the rising edge
of CLK.
MSC
Instruction inputs to Barrel Shifter, IS3 = MSB.1 These inputs are latched internally on the
rising edge of CLK.
IS0 - IS3
SV0 - SV3
Shift Value I/O Port. This port is used as an input when shift values are supplied from
external sources, and as an output when Normalise operations are invoked. The I/O functions
are determined by the IS0 - IS3 instruction inputs, and by the SVOE control.
The shift value is latched internally on the rising edge of CLK.
SV Output enable. When high the SV port can only operate as an input. When low the SV
port can act as an input or as an output, according to the IS0 - IS3 instruction. This pin should
be tied hihg or low, depending upon the application.
SVOE
Instruction inputs to Barrel Shifter registers.1 These inputs are latched internally on the
rising edge of CLK.
RS0, RS1
RS2
C0 - C15
C Port data output. Data output on this port is selected by the C output multiplexer.
C15 is the MSB.
Output enable. The C Port outputs are in high impedance condition when this control is high.
Block Floating Point Flag from ALU, active high.
OE
BFP
CO
Carry out from MSB of ALU.
Instruction inputs to ALU registers.1 These inputs are latched internally on the rising
edge of CLK.
RA0 - RA2
CI
Carry in to LSB of ALU.
Instruction inputs to ALU.1 IA4 = MSB. These inputs are latched internally on the rising
edge of CLK.
IA0 - IA3
IA4
Vcc
+5V supply: Both Vcc pins must be connected.
0V supply: Both GND pins must be connected.
GND
NOTES
1. All instructions are executed in the cycle commencing with the rising edge of the CLK which latches the inputs.
3
PDSP1601/PDSP1601A
A INPUT
16
B INPUT
16
CEA
CEB
MSS
B REG
A REG
MSA0-1
MSB
A MUX
A
2
B MUX
B
S MUX
BFP
CO
IA0-4
5
IS0-3
SHIFT
CONTROL
BARREL SHIFTER
ALU
CI
SV0-3
SVOE
RAD-2
RS0-2
ALU REG FILE
LEFT REG. RIGHT REG.
SHIFTER REG FILE
3
3
LEFT REG.
RIGHT REG.
MSC
C MUX
OE
16
COUT
Fig.2 PDSP1601 block diagram
FUNCTIONAL DESCRIPTION
The PDSP1601 contains four main blocks: the ALU, the
Barrel Shifter and the two Register Files.
activetheALUresultmusthaveoverflowedintothe16th(sign)
bit, (this flag is only valid whilst the most significant 16 bit byte
is being processed). The zero condition is active if the result
from the ALU is equal to zero. For multiprecision operations
the zero flag must be active for all of the 16 bit bytes of an
extended word.
The BFP flag is programmed by executing on of the four
SBFXXinstructions(seeTable1). Duringtheexecutionofany
of these four instructions, the output of the ALU is forced to
zero.
The ALU
The ALU supports 32 instructions as detailed in Table 1.
The inputs to the ALU are selected by the A and B MUXs.
Data will fall through from the selected register through the A
or B input MUXs and the ALU to the ALU output register file in
50ns for the PDSP1601A (100ns for the PDSP1601).
The ALU instructions are latched, such that the instruction
will not start executing until the rising edge of CLK latches the
instruction into the device.
Multicycle/Cascade Operation
The ALU arithmetic instructions contain two or three
options for each arithemtic operation.
The ALU accepts a carry in from the CI input and supplies
a carry out to the CO output. Additionally, at the end of each
cycle, the carry out from the ALU is loaded into an internal 1
bit register, so that it is available as an input to the ALU on the
next cycle. In the manner, multicycle, multiprecision
operations are supported. (See MULTICYCLE CASCADE
OPERATIONS).
The ALU is designed to operate with two's complement
arithmetic, requiring a one to be added to the LSB for all
subtractoperations. Theinstructionssetincludesinstructions
thatwillforceaoneintotheLSB, e.g. MIAX1, AMBX1, BMAX1
(see Table 1).
These instructions are used for the least significant 16 bit
byte of any subtract operation.
BFP Flag
The user has an option of cascading multiple devices, or
multicycling a single device to extend the arithmetic precision.
Should the user cascade multiple devices, then the cascade
arithmetic instructions using the external CI input should be
employedforallbuttheleastsignificant16bitbyte,e.g.MIACI,
APBCI, BMACI (see Table 1).
Should the user multicycle a single device, then the
Multicycle Arithmetic instructions, using the internally
registered CO bit should be employed for all but the least
significant 16 bit byte, e.g. MIACO, APBCO, AMBCO,
BMACO (see Table 1).
The ALU has a user programmable BFP flag. This flag
may be programmed to become active at any one of four
conditions. Two of these conditions are intended to support
Block Floating Point operations, in that they provide flags
indicating that the ALU result is within a factor of two or four of
overflowing the 16 bit number range. For multiprecision
operations the flag is only valid whilst the most significant 16
bit byte is being processed. In this manner the BFP flag may
be used over any extended word width.
The remaining two conditions detect either an overflow
condition or a zero result. For the overflow condition to be
4
PDSP1601/PDSP1601A
Table 1 ALU instructions
1a. ARITHMETIC INSTRUCTIONS
Mnemonic
Inst
Function
IA4-AI0
Operation
Mode
CLRXX
MIAX1
MIACI
00
01
02
03
04
05
06
07
08
09
0A
0B
0C
0D
0E
0F
00000
00001
00010
00011
00100
00101
00110
00111
01000
01001
01010
01011
01100
01101
01110
01111
RESET
CLEAR ALL REGISTERS
NA Plus 1
NA Plus CI
---------
LSBYTE
MINUS A
MINUS A
MINUS A
A/2
A/2
A/2
CASCADE
MULTICYCLE
MSBYTE
MULTICYCLE
MULTICYCLE
MULTICYCLE
CASCADE
MULTICYCLE
LSBYTE
CASCADE
MULTICYCLE
LSBYTE
CASCADE
MULTICYCLE
MIACO
A2SGN
A2RAL
A2RAR
A2RSX
APBCI
APBCO
AMBX1
AMBCI
AMBCO
BMAX1
BMACI
BMACO
NA Plus CO
A/2 Sign Extend
A/2 with RAL LSB
A/2 with RAR LSB
A/2 with RSX LSB
A Plus B Plus CI
A Plus B Plus CO
A Plus NB Plus 1
A Plus NB Plus CI
A Plus NB Plus CO
NA Plus B Plus 1
NA Plus B Plus CI
NA Plus B Plus CO
A/2
A PLUS B
A PLUS B
A MINUS B
A MINUS B
A MINUS B
B MINUS A
B MINUS A
B MINUS A
1b. LOGICAL INSTRUCTIONS
Mnemonic
Inst IA4-AI0
Operation
Function
ANXAB
ANANB
ANNAB
ORXAB
ORNAB
XORAB
PASXA
PASNA
10
11
12
13
14
15
16
17
10000
10001
10010
10011
10100
10101
10110
10111
A AND B
A AND NB
NA AND B
A OR B
NA OR B
A XOR B
PASS A
A. B
A. NB
NA. B
A + B
NA + B
A XOR B
A
INVERT A
NA
1c. CONTROL INSTRUCTIONS
Mnemonic
Inst IA4-AI0
Operation
SBFOV
SBFU1
SBFU2
SBFZE
OPONE
OPBYT
OPNIB
OPALT
18
19
1A
1B
1C
1D
1E
1F
11000
11001
11010
11011
11100
11101
11110
11111
Set BFP Flag to OVR, Force ALU output to zero
Set BFP Flag to UND 1 Force ALU output to zero
Set BFP Flag to UND 2 Force ALU output to zero
Set BFP Flag to ZERO Force ALU output to zero
Output 0001 Hex
Output 00FF Hex
Output 000F Hex
Output 5555 Hex
KEY
A
B
CI
CO
MNEMONICS
= A input to ALU
= B input to ALU
= External Carry in to ALU
CLRXX
MIAXX
A2XXX
APBXX
AMBXX
BMAXX
ANX-Y
ORX-Y
XORXY
PASXX
SBFXX
OPXXX
Clear All Registers to zero
Minus A,
A Divided by 2, XXX = Source of MSB
A Plus B,
A Minus B,
B Minus A,
AND
XX = Carry in to LSB
XX = Carry in to LSB
XX = Carry in to LSB
XX = Carry in to LSB
= Internally Registered Carry out from ALU
RAL = ALU Register (Left)
RAR = ALU Register (Right)
RSX = Shifter Register (Left or Right)
X
X
X
= Operand 1, Y = Operand 2
= Operand 1, Y = Operand 2
= Operand 1, Y = Operand 2
OR
Exclusive OR
Pass
Set BFP Flag
XX = Operand
XX = Function
Output Constant XXX
5
PDSP1601/PDSP1601A
Divide by Two
The Barrel Shifter
The ALU has four (A2SGN, A2RAL, A2RAR, A2RSX)
instructions used for right shifting (dividing by two) extended
precision words. These words, (up to 64 bits) may be stored
in the two on-chip register files. When the least significant 16
bit word is shifted, the vacant MSB must be filled with the LSB
from the next most significant 16 bit byte. This is achieved via
the A2RAL, A2RAR or A2RSX instructions which indicate the
source of the new MSB (see ALU INSTRUCTION SET).
When the most significant 16 bit byte is right shifted, the
MSB must be filled with a duplicate of the original MSB so as
to maintain the correct sign (Sign Extension). This operation
is achieved via the A2SGN instruction (see Table 1).
The Barrel Shifter supports 16 instructions as detailed in
Table 2. The input to the Barrel Shifter is selected by the S
MUX. Data will fall through from the selected register, through
the S MUX and the Barrel Shifter to the shifter output register
file in 50ns for the PDSP1601A (100ns for the PDSP1601).
The Barrel Shifter instructions are latched, such that the
instructions will not start executing until the rising edge of CLK
latches the instruction into the device.
TheBarrelShifteriscapableofLogicalArithmeticorBarrel
Shifts in either direction.
A. Logical shifts discard bits that exit the 16 bit field and fill
spaces with zeros.
Constants
B. Arithmetic shifts discard bits that exit the 16 bit field and
fill spaces with duplicates of the original MSB.
C. Barrel Shifts rotate the 16 bit fields such that bits tha exit
the 16 bit field to the left or right reappear in the vacant
spaces on the right or left.
The ALU has four instructions (OPONE, OPBYT, OPNIB,
OPALT) that force a constant value onto the ALU output.
These values are primarily intended to be used as masks, or
the seeds for mask generation, for example, the OPONE
instruction will set a single bit in the least significant position.
This bit may be rotated any where in the 16 bit field by the
BarrelShifter, allowingtheANDfunctionoftheALUtoperform
bit-pick operations on input data.
Theamountofshiftappliedisencodedontothe4bitBarrel
Shifter input as illustrated in Table 3. The type of shift and the
amount are determined by the shift control block. The shift
controlblock(seeFig.3)acceptsanddecodesthefourbitISO-
3 instruction. The shift control block contains a priority
encoder and two user programmable 4 bit registers R1 and
R2.
CLR
The ALU instruction CLRXX is used as a Master Reset for
the entire device. This instruction has the effect of:
There are four possible sources of shift value that can be
passed onto the Barrel Shifter, there are:
1. Clearing ALU and Barrel Shifter register files to zero.
2. Clearing A and B port input registers to zero.
1. The Priority Encoder
2. The SV input
3. Clearing the R1 and R2 shift control registers to zero.
4. Clearing the internally registered CO bit to zero.
5. Programming the BFP flag to detect overflow conditions.
3. The R1 register
4. The R2 register
Mnemonic
Inst IS3-IS0
Operation
I/O
LSRSV
LSLSV
BSRSV
BSLSV
LSRR1
LSLR1
LSRR2
LSLR2
LR1SV
LR2SV
ASRSV
ASRR1
ASRR2
NRMXX
NRMR1
NRMR2
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
I
I
I
I
X
X
X
X
I
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
Logical Shift Right by SV
Logical Shift Left by SV
Barrel Shift Right by SV
Barrel Shift Left by SV
Logical Shift Right by R1
Logical Shift Left by R1
Logical Shift Right by R2
Logical Shift Left by R2
Load Register 1 From SV
Load Register 2 From SV
Arithmetic Shift Right by SV
Arithmetic Shift Right by R1
Arithmetic Shift Right by R2
Normalise Output PE
I
I
X
X
O
O
O
Normalise Output PE, Load R1
Normalise Output PE, Load R2
Table 2 Barrel shifter instructions
MNEMONICS
LSXYY Logical Shift,
BSXYY Barrel Shift,
ASXYY Arithmetic Shift, X
LXXYY Load
NRMYY Normalise by PE, Output PE value on SV Port, Load YY Reg
KEY
SV
R1
R2
PE
I
X
X
= Direction YY = Source of Shift Value
= Direction YY = Source of Shift Value
= Direction YY = Source of Shift Value
= Shift Value
= Register 1
= Register 2
XX = Target YY = Source
= Priority Encoder Output
=> SV Port operates as an Input
=> SV Port operates as an Output
=> SV Port in a High Impedance State
O
X
6
PDSP1601/PDSP1601A
(1)
Priority encode the 16 bit input to the Barrel Shifter and
SV3
SV2 SV1 SV0
Shift
place the 4 bit value in either of the R1 or R2 registers and
output the value on the SV port (if enabled by SVOE).
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
0
0
0
0
1
1
1
1
0
0
0
0
1
1
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
No shift
1 place
(2)
Shift the 16 bit input by the amount indicated by the
2 places
3 places
4 places
5 places
6 places
7 places
8 places
9 places
10 places
11 places
12 places
13 places
14 places
15 places
Priority Encoder such that the output from the Barrel Shifter is
a normalised value.
SV Input
If the SV port is selected as the source of the shift value,
thentheinputtotheBarrelShifterisshiftedbythevaluestored
in the internal SV register.
SVOE
The SV port acts as an input or an output depending upon
the IS0-3 instruction. If the user does not wish to use the
normalise instructions, then the SV port mat be forced to be
Table 3 Barrel shifter codes
input only by typing SVOE control high. In this mode the SV
port may be considered an extension of the instruction inputs.
R1 and R2 Registers
Priority Encoder
The R1 and R2 registers may be loaded from the Priority
Encoder (NRMR1 and NRMR2) or from the SV input (LR1SV,
LR2SV).
Whilst the latter two instructions are executing, the Barrel
Shifter will pass its input to the output unshifted.
If the priority encoder is selected as the source of the shift
value (instructions:- NRMXX, NRMR1, MRMRZ), then within
one 100ns cycle or two 50ns cycles for the PDSP1601A (one
200ns or two 100ns cycles for the PDSP1601), the shift
circuitry will:
16
4
INSTRUCTION
IS0-3
DECODE
PRIORITY ENCODER
4
4
MUX
SV
MUX
R1
MUX
SVOE
R2
Fig.3 Shift control block
7
PDSP1601/PDSP1601A
The Register Files
Therearetwoon-chipregisterfiles(ALUandShifter),each
containing two 16 bit registers and each supporting 8
instructions (see Table 4). The instructions for the ALU
register file and the Barrel Shifter Register file are the same.
The Inputs to the register files come from either the ALU or
the Barrel Shifter, and are loaded into the Register files on the
rising edge of CLK.
CLK latches the instruction into the device.
The register file instructions (see Table 4) allow input data
to be loaded into either, neither or both of the registers. Data
is loaded at the end of the cycle in which the instruction is
executing.
The register file instructions allow the output to be sourced
fromeitherofthetworegisters,theselectedoutputwillbevalid
during the cycle in which the instruction is executing.
The register file instructions are latched such that the
instruction will not start executing until the rising edge of the
ALU REGISTER INSTRUCTIONS
Mnemonic
Inst RA2-RA0
Operation
LLRRR
LRRLR
LLRLR
LRRRR
LBRLR
NOPRR
NOPLR
NOPPS
0
1
2
3
4
5
6
7
000
001
010
011
100
101
110
111
Load Left Reg Output Right Reg
Load Right Reg Output Left Reg
Load Left Register, Output Left Reg
Load Right Register, Output Right Reg
Load Both Registers, Output Left Reg
No Load Operation, Output Right Reg
No Load Operation, Output Left Reg
No Load Operation, Pass ALU Result
SHIFTER REGISTER INSTRUCTIONS
Mnemonic
Inst RA2-RA0
Operation
LLRRR
LRRLR
LLRLR
LRRRR
LBRLR
NOPRR
NOPLR
NOPPS
0
1
2
3
4
5
6
7
000
001
010
011
100
101
110
111
Load Left Reg Output Right Reg
Load Right Reg Output Left Reg
Load Left Register, Output Left Reg
Load Right Register, Output Right Reg
Load Both Registers, Output Left Reg
No Load Operation, Output Right Reg
No Load Operation, Output Left Reg
No Load Operation, Pass Barrel Shifter Result
Table 4 ALU and shift register instructions mnemonics
MNEMONICS
LXXYY Load XX = Target,
YY = Source of Output
LBOXX Load Both Registers, XX = Source of Output
NOPXX No Load Operation, XX = Source of Output
8
PDSP1601/PDSP1601A
Multiplexers
There are four user selectable on-chip multiplexers (A-
MUX, B-MUX, S-MUX and C-MUX).
These four multiplexers support instructions as tabulated
in Table 5.
The MUX instructions are latched such that the instruction
will not start executing until the rising edge of CLK latches the
instruction onto the device.
Output
MSA1
MSA0
ALU REGISTER FILE OUPUT
A-PORT INPUT
B-PORT INPUT
MARAX
MAAPR
MABPR
MARSX
0
0
1
1
A-MUX
0
1
0
1
SHIFTER REGISTER FILE OUTPUT
Output
MSB
B-PORT INPUT
SHIFTER REGISTER FILE OUTPUT
0
1
B-MUX
S-MUX
C-MUX
Output
MSS
B-PORT INPUT
SHIFTER REGISTER FILE OUTPUT
0
1
Output
MSC
ALU REGISTER FILE OUTPUT
SHIFTER REGISTER FILE OUTPUT
0
1
Table 5
9
PDSP1601/PDSP1601A
INSTRUCTION SET
ALU Arithmetic Instructions
Op Code
Function
Mnemonic
<00>
On the rising edge of CLK at the end of the cycle in which this instruction is executing, the
A Port, B Port, ALU, Barrel Shifter, and Shift Control Registers will be loaded with zeros.
The internal registered CO will also be set to zero, and the BFP flag will be set to activate
on overflow conditions.
CLRXX
<01>
<02>
<03>
The A input to the ALU is inverted and a one is added to the LSB.
The A input to the ALU is inverted and the CI input is added to the LSB.
MIAX1
MIAC1
MIACO
The A input to the ALU is inverted and the CO output from the ALU on the previous cycle
is added to the LSB.
<04>
<05>
<06>
<07>
TheAinputtotheALUisrightshiftedonebitposition. TheLSBisdiscarded,andthevacant
MSB is filled by duplicating the original MSB (Sign Extension).
A2SGN
A2RAL
A2RAR
A2RSX
TheAinputtotheALUisrightshiftedonebitposition. TheLSBisdiscarded,andthevacant
MSB is filled with the LSB from the ALU register.
TheAinputtotheALUisrightshiftedonebitposition. TheLSBisdiscarded,andthevacant
MSB is filled with the LSB from the ALU register.
TheAinputtotheALUisrightshiftedonebitposition. TheLSBisdiscarded,andthevacant
MSB is filled with the LSB from the B input to the ALU.
<08>
<09>
The A input to the ALU is added to the B input, and the CI input is added to the LSB.
APBCI
TheAinputtotheALUisaddedtotheBinput, andtheCOoutfromtheALUontheprevious
cycle is added to the LSB.
APBCO
<0A>
<0B>
The A input to the ALU is added to the inverted B input, and a one is added to the LSB.
AMBX1
AMBCI
The A input to the ALU is added to the inverted B input, and the CI input is added to the
LSB.
<0C>
The A input to the ALU is added to the inverted B input, and the CO out from the ALU on
the previous cycle is added to the LSB.
AMBCO
<0D>
<0E>
The inverted A input to the ALU is added to the B input, and a one is added to the LSB.
BMAX1
BMAC1
The inverted A input to the ALU is added to the B input, and the CI input is added to the
LSB.
<0F>
The inverted A input to the ALU is added to the B input, and the CO out from the ALU on
the previous cycle is added to the LSB.
BMACO
ALU Logical Instructions
Op Code
Function
Mnemonic
<10>
The A input to the ALU is logically 'ANDed' with the B input.
The A input to the ALU is logically 'ANDed' with the inverse of the B input.
The inverse of the A input to the ALU is logically 'ANDed' with the B input.
The A input to the ALU is logically 'ORed' with the B input.
The inverse A input to the ALU is logically 'ORed' with the B input.
The A input to the ALU is logically Exclusive-ORed with the B input.
The A input to the ALU is passed to the output.
ANXAB
<11>
ANANB
<12>
ANNAB
<13>
ORXAB
<14>
ORNAB
<15>
XORAB
<16>
PASXA
<17>
The inverse of the A input to the ALU is passed to the output.
PASNA
10
PDSP1601/PDSP1601A
ALU Control Instructions
Op Code
Function
Mnemonic
<18>
The BFP flag is programmed to activate when an ALU operation causes an overflow of the
16 bit number range. This flag is logically the exclusive-or of the carry into and out of the
MSB of the ALU. For the most significant Byte this flag indicates that the result of an
arithmetic two's complement operation has overflowed into the sign bit. The output of the
ALU is forced to zero for the duration of this instruction.
SBFOV
<19>
The BFP flag is programmed to activate when an ALU operation comes within a factor of
two of causing an overflow of the 16 bit number range. For the most significant Byte this
flag indicates that the result of an arithmetic two's complement operation is within a factor
oftwoof overflowingintothesignbit. TheoutputoftheALUisforcedtozerofortheduration
of this instruction.
SBFU1
<1A>
The BFP flag is programmed to activate when an ALU operation comes within a factor of
four of causing an overflow of the 16 bit number range. For the most significant Byte this
flag indicates that the result of an arithmetic two's complement operation is within a factor
offourofoverflowingintothesignbit. TheoutputoftheALUisforcedtozerofortheduration
of this instruction.
SBFU2
<1B>
The BFP flag is programmed to activate when an ALU operation causes a result of zero.
The output of the ALU is forced to zero for the duration of this instruction. During the
execution of this instruction the BFP flag will become active.
SBFZE
<1C>
The ALU will output the binary value 0000000000000001, the MSB on the left.
The ALU will output the binary value 0000000011111111, the MSB on the left.
The ALU will output the binary value 0000000000001111, the MSB on the left.
The ALU will output the binary value 0101010101010101, the MSB on the left.
OPONE
<1D>
OPBYT
<1E>
OPNIB
<1F>
OPALT
Barrel Shifter Instructions
Op Code
Function
Mnemonic
<0>
The 16 bit input to the Barrel Shifter is right shifted by the number of places indicated by
the magnitude of the four bit number present in the SV register. The LSBs are dicarded,
and the vacant MSBs are filled with zeros.
LSRSV
<1>
The 16 bit input to the Barrel Shifter is left shifted by the number of places indicated by the
magnitude of the four bit number present in the SV register. The LSBs are dicarded, and
the vacant MSBs are filled with zeros.
LSLSV
<2>
The 16 bit input to the Barrel Shifter is rotated to the right by the number of places indicated
by the magnitude of the four bit number present in the SV register. The LSBs that exit the
16 bit field to the right, reappear in the vacant MSBs on the left.
BSRSV
<3>
The 16 bit input to the Barrel Shifter is rotated to the left by the number of places indicated
by the magnitude of the four bit number present in the SV register. The LSBs that exit the
16 bit field to the right, reappear in the vacant MSBs on the right.
BSLSV
<4>
The 16 bit input to the Barrel Shifter is right shifted by the number of places indicated by
the magnitude of the four bit number resident within the R1 register. The LSBs are
discarded, and the vacant MSBs are filled with zeros.
LSRR1
<5>
The 16 bit input to the Barrel Shifter is left shifted by the number of places indicated by the
magnitude of the four bit number resident within the R1 register. The LSBs are discarded,
and the vacant LSBs are filled with zeros.
LSLR1
<6>
The 16 bit input to the Barrel Shifter is right shifted by the number of places indicated by
the magnitude of the four bit number resident within the R2 register. The LSBs are
discarded, and the vacant MSBs are filled with zeros.
LSRR2
<7>
The 16 bit input to the Barrel Shifter is left shifted by the number of places indicated by the
magnitude of the four bit number resident within the R2 register. The LSBs are discarded,
and the vacant LSBs are filled with zeros.
LSLR2
11
PDSP1601/PDSP1601A
Op Code
Function
Mnemonic
<8>
<9>
<A>
<B>
On the rising edge of CLK at the end of the cycle in which this instruction is executing, the
R1 register will be loaded with the data present on the SV port. The input to the Barrel
Shifter will be passed onto the output unshifted.
LR1SV
LR2SV
ASRSV
ASRR1
On the rising edge of CLK at the end of the cycle in which this instruction is executing, the
R2 register will be loaded with the data present on the SV port. The input to the Barrel
Shifter will be passed onto the output unshifted.
The 16 bit input to the Barrel Shifter is right shifted by the number of places indicated by
the magnitude of the four bit number present in the SV register. The LSBs are discarded,
and the vacant MSBs are filled with duplicates of the original MSB. (Sign Extension).
The 16 bit input to the Barrel Shifter is right shifted by the number of places indicated by
the magnitude of the four bit number resident within the R1 register. The LSBs are
discarded, and the vacant MSBs are filled with duplicates of the original MSB.
(Sign Extension).
<C>
<D>
The 16 bit input to the Barrel Shifter is right shifted by the number of places indicated by
the magnitude of the four bit number resident within the R2 register. The LSBs are
discarded, and the vacant MSBs are filled with duplicates of the original MSB.
(Sign Extension).
ASRR2
NRMXX
The 16 bit input to the Barrel Shifter is left shifted by the number of places indicated by the
magnitudeofthefourbitnumberoutputfromthePriorityEncoder. Thisvalueisalsooutput
on the SV port (provided SVOE is low).
The effect of this operation is to left shift the input by the necessary amount
(max 15 places) to result in the MSB and the next most significant bit being different. This
has the effect of eliminating unnecessary Sign Bits, and hence Normalising the input data.
The MSBs shifted out to the left are discarded, and the vacant LSBs on the right are filled
with zeros.
<E>
The 16 bit input to the Barrel Shifter is left shifted by the number of places indicated by the
magnitudeofthefourbitnumberoutputfromthePriorityEncoder. Thisvalueisalsoloaded
NRMR1
NRMR2
into the R1 register at the end of the cycle, and is output on the SV port (provided SVOE
is low).
The effect of this operation is to left shift the input by the necessary amount
(max 15 places) to result in the MSB and the next most significant bit being different. This
has the effect of eliminating unnecessary Sign Bits, and hence Normalising the input data.
The MSBs shifted out to the left are discarded, and the vacant LSBs on the right are filled
with zeros.
<F>
The 16 bit input to the Barrel Shifter is left shifted by the number of places indicated by the
magnitudeofthefourbitnumberoutputfromthePriorityEncoder. Thisvalueisalsoloaded
into the R2 register at the end of the cycle, and is output on the SV port (provided SVOE
is low).
The effect of this operation is to left shift the input by the necessary amount
(max 15 places) to result in the MSB and the next most significant bit being different. This
has the effect of eliminating unnecessary Sign Bits, and hence Normalising the input data.
The MSBs shifted out to the left are discarded, and the vacant LSBs on the right are filled
with zeros.
12
PDSP1601/PDSP1601A
Barrel Shifter or ALU Register Instructions
Op Code
Function
Mnemonic
<0>
<1>
<2>
<3>
<4>
<5>
<6>
<7>
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Right register will appear on the output. On the rising edge
of CLK at the end of the cycle, and the data on the register inputs will be loaded into the
Left Register.
LLRRR
LRRLR
LLRLR
LRRRR
LBRLR
NOPRR
NOPLR
NOPPS
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Left register will appear on the output. On the rising edge
of CLK at the end of the cycle, the data on the register inputs will be loaded into the Right
Register.
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Left register will appear on the output. On the rising edge
of CLK at the end of the cycle, the data on the register inputs will be loaded into the Left
Register.
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Right register will appear on the output. On the rising edge
of CLK at the end of the cycle, the data on the register inputs will be loaded into the Right
Register.
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Left register will appear on the output. On the rising edge
of CLK at the end of the cycle, and the data on the register inputs will be loaded into both
Left and Right Register.
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Right register will appear on the output. On the rising edge
of CLK at the end of the cycle no load operation will occur, the register contents will remain
unchanged.
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the contents of the Left register will appear on the output. On the rising edge
of CLK at the end of the cycle no load operation will occur, the register contents will remain
unchanged.
After the rising edge of CLK at the beginning of the cycle in which this instruction is
executed, the input to the registers will appear on the output. On the rising edge
of CLK at the end of the cycle no load operation will occur, the register contents will remain
unchanged.
13
PDSP1601/PDSP1601A
TYPICAL APPLICATION
(2)
The LS byte is logically right shifted, n-places, the
Select a 16 bit field from each word in a block of 32 bit
words with a 10MHz throughput.
The 16 bit field indicated is to be selected from each 32 bit
word.
LSBs being discarded and the MSBs being filled with zeros.
This shifted data is loaded into the shifter register file left
register.
During this cycle the previous contents of this register are
passed through the ALU to the ALU register file left register.
MS Byte
LS Byte
MS Bit
(3)
While the MS byte of the next 32 bit word is shifted in
8
8
8
8
the Barrel Shifter, the two previous results, resident within the
left registers of the ALU and Shifter Register files are 'ORed'
by the ALU, the result being the desired 16 bit field is loaded
intotheALUregisterfilerightregisterreadytobeoutputonthe
next cycle.
16 bits
nbits
The 32 bit words are fed into the B port of the PDSP1601
in two cycles, MS byte first.
The PDSP1601 shift control is initiated by programming
the R1 and R2 registers with n and 16-n respectively.
The shift operation is implemented in three steps:-
The instructions from initialisation are given in Table 6.
(1)
The MS byte is logically left shifted (16-n) places, the
MSBs being discarded and the LSB spaces being filled with
zeros. This shifted data is loaded into the shifter register file
left register.
CLK CEB
MSA
MSB MSS MSC
IA
IS
SV
Comment
RA
RS
1
1
0
0
0
0
0
0
0
0
0
0
CLRXX
X
X
1/
2/
3/
4/
5/
6/
MARSX
MARSX
MARSX
MARSX
MARSX
MARAX
1
1
1
1
1
1
0
0
0
0
0
0
NOPLR
NOPLR
NOPLR
NOPLR
LLRRR
LRRLR
Clear
NOPLR
NOPLR
NOPLR
LLRLR
LLRLR
LLRLR
PASXA LR1SV
PASXA LR2SV
PASXA LSLR2
PASXA LSRR1
ORXAB LSLR2
n
(16-n)
X
X
X
Load R1 with n
Load R2 with (16-n)
Shift 1st MS byte
Shift 1st LS byte
OR 1st bytes and
shift 2nd MS byte
Shift 2nd LS byte
and output first result
Shift 3rd LS byte
PASXA LSRR1
ORXAB LSLR2
7/
8/
0
0
MARSX
MARAX
1
1
0
0
0
0
X
X
LLRRR
LRRLR
LLRLR
LLRLR
Repeat instruction pair 5/ and 6/ until all 16 bit fields have been selected.
Table 6
ABSOLUTE MAXIMUM RATINGS (Note 1)
NOTES
Supply voltage Vcc
Input voltage VIN
Output voltage VOUT
-0.5V to 7.0V
-0.9 to Vcc + 0.9V
-0.9 to Vcc + 0.9V
1. Exceeding these ratings may cause permanent damage.
Functional operation under these conditions is not implied.
2. Maximum dissipation or 1 second should not be exceeded, only
one output to be tested at any one time.
Clamp diode current per pin Ik(see note 2)
±18mA
Static discharge voltage (HMB)
Storage temperature TS
Ambient temperature with
power applied Tamb
Military
500V
-65°C to +150°C
-40°C to +125°C
-40°C to +85°C
THERMAL CHARACTERISTICS
Industrial
Package power dissipation PTOT
Package type
ΘJC °C/W
ΘJA °C/W
AC
GC
12
12
36
35
AC
LC
1000mw
1000mw
14
PDSP1601/PDSP1601A
ELECTRICAL CHARACTERISTICS
Operating Conditions (unless otherwise stated)
TAMB (Commercial) = 0°C to +70°C, VCC = 5.0V±5%, Ground = 0V
TAMB (Industrial) = -40°C to +85°C, VCC = 5.0V±10%, Ground = 0V
TAMB (Military) = -55°C to +125°C, VCC = 5.0V±10%, Ground = 0V
Static Characteristics
Symbol
Value
Typ.
Characteristic
Units
Conditions
Min.
2.4
Max.
VOH
VOL
VIH
VIL
IIL
ICC
IOZ
ISC
CIN
Output high voltage
Output low voltage
Input high voltage
Input low voltage
Input leakage current
Vcc current
Output leakage current
Output S/C current
Input capacitance
V
V
V
IOH = 8mA
IOL = -8mA
0.4
3.5
0.5
+10
60
+50
80
V
-10
µA
mA
µA
mA
pF
GND < VIN < VCC
Tamb = -40°C to +85°C
GND < VOUT < VCC
VCC = Max
-50
12
5
Switching Characteristics
Value
PDSP1601A
PDSP1601
Min. Max.
Characteristic
Conditions
Units
Min.
Max.
5
5
5
5
5
5
25
50
50
2 x LSTTL + 20pF
1 x LSTTL + 5pF
1 x LSTTL + 5pF
40
100
100
CLK rising edge to C-PORT
CLK rising edge to CO
CLK rising edge to BFP
ns
ns
ns
ns
ns
ns
ns
ns
15
30
Setup CEA or CEB to CLK rising edge
0
0
0
0
Hold CEA or CEB after CLK rising edge
Setup A or B port inputs to CLK rising edge
Hold A or B port inputs after CLK rising edge
40
40
20
20
Setup MSA0-1, MSB, MSS, MSC, RA2-0, RS0-2, IA0-4,
IS0-3, to CLK rising edge
Hold RS0-2, IA0-4 after CLK rising edge
Hold IS0-3 after CLK rising edge
Hold MSA0-1, MSB, MSS, MSC, RA0-2 after CLK rising edge
Setup SV to CLK rising edge
Hold SV after CLK rising edge
0
3
0
0
3
0
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
ns
40
5
20
5
Input mode
Input mode
3
3
100
40
40
40
40
50
25
25
25
25
20pF load, SV O P mode
2 x LSTTL + 20pF
2 x LSTTL + 20pF
2 x LSTTL + 20pF
2 x LSTTL + 20pF
CLK rising edge to SV
OE
OE
OE
OE
C-PORT
C-PORT
C-PORT Z
C-PORT Z
Z
Z
100
50
20
200
100
40
Clock period (ALU & Barrel Shifter, serial mode)
Clock period (ALU & Barrel Shifter, parallel mode)
Clock high time
40
20
Clock low time
NOTES
1. LSTTL is equivalent to IOH at 20µA IOL of -0.4mA
2. Current is defined as negative into the device.
15
For more information about all Zarlink products
visit our Web Site at
www.zarlink.com
Information relating to products and services furnished herein by Zarlink Semiconductor Inc. or its subsidiaries (collectively “Zarlink”) is believed to be reliable.
However, Zarlink assumes no liability for errors that may appear in this publication, or for liability otherwise arising from the application or use of any such
information, product or service or for any infringement of patents or other intellectual property rights owned by third parties which may result from such application or
use. Neither the supply of such information or purchase of product or service conveys any license, either express or implied, under patents or other intellectual
property rights owned by Zarlink or licensed from third parties by Zarlink, whatsoever. Purchasers of products are also hereby notified that the use of product in
certain ways or in combination with Zarlink, or non-Zarlink furnished goods or services may infringe patents or other intellectual property rights owned by Zarlink.
This publication is issued to provide information only and (unless agreed by Zarlink in writing) may not be used, applied or reproduced for any purpose nor form part
of any order or contract nor to be regarded as a representation relating to the products or services concerned. The products, their specifications, services and other
information appearing in this publication are subject to change by Zarlink without notice. No warranty or guarantee express or implied is made regarding the
capability, performance or suitability of any product or service. Information concerning possible methods of use is provided as a guide only and does not constitute
any guarantee that such methods of use will be satisfactory in a specific piece of equipment. It is the user’s responsibility to fully determine the performance and
suitability of any equipment using such information and to ensure that any publication or data used is up to date and has not been superseded. Manufacturing does
not necessarily include testing of all functions or parameters. These products are not suitable for use in any medical products whose failure to perform may result in
significant injury or death to the user. All products and materials are sold and services provided subject to Zarlink’s conditions of sale which are available on request.
Purchase of Zarlink’s I2C components conveys a licence under the Philips I2C Patent rights to use these components in and I2C System, provided that the system
conforms to the I2C Standard Specification as defined by Philips.
Zarlink, ZL and the Zarlink Semiconductor logo are trademarks of Zarlink Semiconductor Inc.
Copyright Zarlink Semiconductor Inc. All Rights Reserved.
TECHNICAL DOCUMENTATION - NOT FOR RESALE
相关型号:
©2020 ICPDF网 联系我们和版权申明