Renesas Single-Chip Microcomputer M34519T-MCU Spécifications Page 15

  • Télécharger
  • Ajouter à mon manuel
  • Imprimer
  • Page
    / 51
  • Table des matières
  • MARQUE LIVRES
  • Noté. / 5. Basé sur avis des utilisateurs
Vue de la page 14
The increased amount of address calculations in the CPU in the case of complex addressing causes disturbances in the pipeline. As
a result, address calculation becomes a bottleneck for pipeline processing making it difcult to raise the frequency and increase the
performance. The V850 microcontrollers avoid this problem by supporting only simple addressing.
Simple addressing
The V850 microcontrollers employ a 2-byte instruction code to perform basic processing to enable compact program development
equivalent to 16-bit CISC microcontrollers.
1.03
1.02
1.48
1.00
Object Code Size Comparison
(Dhrystone 1.1/Large model)
16-bitV (CISC)
78K4 (CISC)
V850 (RISC)
V
R
TM
/MIPS
TM
32 (RISC)
• Improved object efciency through ROMization programming
Application of 2-byte instructions to all basic processing, consisting of load, store,
arithmetic/logic operations, and branching.
• To realize ease of use, restrictions on 16-bit xed-length instructions are
partially removed through the incorporation of 32-bit instructions.
• Bit manipulation instructions, etc., are available.
2-byte basic instruction set
Operand addresses
reg1
031
disp16
016 1531
Signed extension
Memory subject to
manipulation
reg1
031
disp16
0
16 15
31
Signed extension
Memory subject to
manipulation
• Register addressing
Addressing that accesses the general-purpose register specied by the general-
purpose specication eld or a system register as an operand.
• Immediate addressing
Addressing of 5-bit data or 16-bit data for manipulation in the instruction code.
• Based addressing
Addressing that
accesses memory, with
the sum of the contents
of the general-purpose
register (reg1) and
16-bit displacement
(disp16) used as the
operand address.
• Bit addressing
Addressing that
accesses 1 bit of 1 byte
of the memory space,
with the sum of the
contents of the general-
purpose register (reg1)
and 16-bit displacement
(disp16) that has been
sign extended to word
length used as the
operand address.
Instruction addresses
PC
26 2531
0
022 2131
Signed extension
disp22
PC
026 2531
0
Memory subject to
manipulation
0
reg1
PC
026 2531
026 2531
0
Memory subject to
manipulation
• Register addressing (register indirect)
Transfer the contents of the general-purpose register specied by the instruction (reg1) to
the program counter (PC).
• Relative addressing (PC dependent)
Add 9 signed bits or 22 signed bits of data of the instruction code to the program counter.
Addressing modes
Processing time Processing time
All processing is standardized and efcient
Pipeline processing sequence
Pipeline Processing Time and CPU Operating Frequency
Operating
frequency held back
by slow processing
In case of excessive addressing
In case of simple addressing
Address calculation
Execution
Memory access
Writeback
Instruction fetch
The V850 microcontrollers support bit manipulation instructions ideal for manipulating the ags in I/O registers, which play a large
role in embedding control.
Object size
Execution time
Coding example
Item
4 bytes
4 clock cycles 4 clock cycles 8 clock cycles
When Used When Not Used
set1 6, ASIM00[r0] ld.b ASIM00[r0], r20
ori 0x0040, r20, r20
st.b r20, ASIM00[r0]
Save r20
Restore r20
add -4, sp
st.w r20, 0[sp]
ld.b ASIM00[r0], r20
ori 0x0040, r20, r20
st.b r20, ASIM00[r0]
ld.w 0[sp], r20
add 4, sp
12 bytes 24 bytes
Bit Manipulation
Instruction
• Improvement of operability of memory mapped
I/O devices for control applications
• Manipulation of any 1 bit of byte data in the
memory space
• Provision of test (tst1)/set (set1)/clear (clr1)/
invert (not1) instructions
• Effective for reducing object size and execution
time since ags can be manipulated in 1-bit
units using 1 instruction
Example: Setting bit 6 of ASIM00 register to 1
CISC-like instructions for embedding (bit manipulation instructions)
In the V850 microcontrollers, calculation results are reected in registers as status ags. As a result, the delay branching that can
occur in the RISC microcontrollers of other manufacturers does not occur and programs can be coded with the same feel as CISC
microcontrollers.
CISC Microcontroller V850
Other Manufacturers’ RISC Microcontroller
cmp/eq #0, r10
bt ZERO
cmp/pl r10
bt PLUS
bra MINUS
nop ;
For delay branching
cmp 0, r10
bz ZERO
/jointfilesconvert/436525/bgt PLUS
br MINUS
cmp ax, 0
jz ZERO
jgt PLUS
jmp MINUS
• Easy coding using an assembler
• Improved object efciency and execution speed
Multi-status flags
The V850 microcontrollers can realize bit manipulations frequently used during signed data and image data processing using 1
instruction per clock cycle.
Shifting of any number of bits (0 to 31) executable in 1 instruction per clock
cycle
Improved execution speed and object efciency
Effective for extracting arbitrary bit lengths of image data and signed data
(extracting code during MH/MR/MMR encoding, etc.)
Number of instructions4 1
V850
Number of execution clock cycles
4 1
Example: 27-bit logical right shift
SHR 27, Rn
Other manufacturers'
RISC microcontroller
SHR16 Rn
SHR8 Rn
SHR2 Rn
SHR Rn
Processing sequence
32-bit barrel shifter
The V850 microcontrollers provide a DSP function for executing high-speed multiplication and product-sum operations
indispensable for digital signal processing such as image and speech processing.
• Direct data handling via general-purpose registers
• Realization of digital signal processing through general-
purpose CPU
• High-speed 16-bit (V850, V850ES CPU) and 32-bit (V850E1
CPU) multiply/sum-of-products operations
(Multiply: 1 to 2 clocks, sum-of-products: 3 clocks)
• Effective for lter operations and matrix operations for
feedback calculations in speed, position, and other servo
control.
V850
General-purpose register
CPU
DSP
Memory
MUL
INT
CPU
ALU
CPU + DSP
MUL
SAT
ag
ALU
DSP function
ZERO: Zero processing
PLUS: Positive processing
MINUS: Negative processing
Example: Program that branches to positive/negative/zero according to register contents
28 29
Vue de la page 14
1 2 ... 10 11 12 13 14 15 16 17 18 19 20 ... 50 51

Commentaires sur ces manuels

Pas de commentaire