U.S. patent application number 12/417505 was filed with the patent office on 2009-09-17 for data processing device and electronic equipment.
This patent application is currently assigned to SEIKO EPSON CORPORATION. Invention is credited to Makoto Kudo.
Application Number | 20090235052 12/417505 |
Document ID | / |
Family ID | 31180314 |
Filed Date | 2009-09-17 |
United States Patent
Application |
20090235052 |
Kind Code |
A1 |
Kudo; Makoto |
September 17, 2009 |
Data Processing Device and Electronic Equipment
Abstract
A data processing device is provided using pipeline architecture
to reduce a time loss due to a branch without causing an increase
in circuit scale. The data processing device uses pipeline control.
The data processing device includes an instruction queue in which a
plurality of instruction codes can be fetched, a fetch address
operation circuit which calculates a fetch address, a fetch circuit
which fetches an instruction code based on the fetch address, and a
branch information setting circuit which decodes a branch setting
instruction, stores a branch address in a branch address storage
register, and stores a branch target address in a branch target
address storage register. The fetch address operation circuit
compares either a previous fetch address or an expected next fetch
address with a value stored in the branch address storage register,
and determines a next fetch address to be output, based on the
comparison result.
Inventors: |
Kudo; Makoto; (Nagano-ken,
JP) |
Correspondence
Address: |
HOGAN & HARTSON L.L.P.
1999 AVENUE OF THE STARS, SUITE 1400
LOS ANGELES
CA
90067
US
|
Assignee: |
SEIKO EPSON CORPORATION
Tokyo
JP
|
Family ID: |
31180314 |
Appl. No.: |
12/417505 |
Filed: |
April 2, 2009 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
10601005 |
Jun 20, 2003 |
|
|
|
12417505 |
|
|
|
|
Current U.S.
Class: |
712/205 ;
712/238; 712/241; 712/E9.016; 712/E9.045; 712/E9.056;
712/E9.058 |
Current CPC
Class: |
G06F 9/325 20130101;
G06F 9/3806 20130101; G06F 9/3842 20130101 |
Class at
Publication: |
712/205 ;
712/238; 712/241; 712/E09.016; 712/E09.045; 712/E09.056;
712/E09.058 |
International
Class: |
G06F 9/30 20060101
G06F009/30 |
Foreign Application Data
Date |
Code |
Application Number |
Jun 25, 2002 |
JP |
2002-184379 |
Claims
1. A data processing device using pipeline control, comprising: an
instruction queue in which a plurality of instruction codes are
fetched; a fetch address operation circuit which calculates a fetch
address used to fetch an instruction code in the instruction queue;
a fetch circuit which fetches an instruction code that is read out
based on the fetch address into the instruction queue; and a branch
information setting circuit which decodes a branch setting
instruction instructing a branch to a branch target address when
the fetch address is a branch address, stores the branch address in
a branch address storage register, and stores the branch target
address in a branch target address storage register, wherein the
fetch address operation circuit includes a circuit which compares
one of a previous fetch address and an expected next fetch address
with a value stored in the branch address storage register, and
then determines whether or not to output a value stored in the
branch target address storage register as a next fetch address,
based on the comparison result.
2. A data processing device using pipeline control, comprising: an
instruction queue in which a plurality of instruction codes are
fetched; a fetch address operation circuit which calculates a fetch
address used to fetch an instruction code in the instruction queue;
a fetch circuit which fetches an instruction code that is read out
based on the fetch address into the instruction queue; and a branch
information setting circuit which decodes a branch setting
instruction instructing a branch to a branch target address when
the fetch address is a branch address, stores the branch address in
a branch address storage register, and stores the branch target
address in a branch target address storage register, wherein the
fetch address operation circuit includes a circuit which compares
an expected next fetch address obtained by incrementing a value in
a fetch program counter by one instruction length with a value
stored in the branch address storage register, and then outputs a
value stored in the branch target address storage register as a
next fetch address when the expected next fetch address coincides
with the value in the branch address storage register, or outputs
the expected next fetch address as a next fetch address when the
expected next fetch address does not coincide with the value in the
branch address storage register.
3. The data processing device as defined in claim 1, wherein: the
branch setting instruction includes a loop instruction which
designates a loop count; the branch information setting circuit
decodes the loop instruction which instructs to repeat a branch to
the branch target address the number of times equal to the loop
count, and stores the loop count designated by the loop
instruction; and the fetch address operation circuit includes a
circuit which outputs a value stored in the branch target address
storage register as a next fetch address until the number of times
the branch to the branch target address repeats reaches the loop
count.
4. The data processing device as defined in claim 2, wherein: the
branch setting instruction includes a loop instruction which
designates a loop count; the branch information setting circuit
decodes the loop instruction which instructs to repeat a branch to
the branch target address the number of times equal to the loop
count, and stores the loop count designated by the loop
instruction; and the fetch address operation circuit includes a
circuit which outputs a value stored in the branch target address
storage register as a next fetch address until the number of times
the branch to the branch target address repeats reaches the loop
count.
5. The data processing device as defined in claim 1, wherein: the
branch setting instruction includes a loop instruction which
designates a loop count; the branch information setting circuit
decodes the loop instruction which instructs to repeat a branch to
the branch target address the number of times equal to the loop
count, and stores the loop count designated by the loop
instruction; and the fetch address operation circuit includes a
circuit which decrements a value set in the loop counter each time
when a branch to the branch target address occurs, and outputs a
value obtained by incrementing the branch address by one
instruction length as a next fetch address when the value of the
loop counter reaches zero.
6. The data processing device as defined in claim 2, wherein: the
branch setting instruction includes a loop instruction which
designates a loop count; the branch information setting circuit
decodes the loop instruction which instructs to repeat a branch to
the branch target address the number of times equal to the loop
count, and stores the loop count designated by the loop
instruction; and the fetch address operation circuit includes a
circuit which decrements a value set in the loop counter each time
when a branch to the branch target address occurs, and outputs a
value obtained by incrementing the branch address by one
instruction length as a next fetch address when the value of the
loop counter reaches zero.
7. The data processing device as defined in claim 3, wherein: the
branch setting instruction includes a loop instruction which
designates a loop count; the branch information setting circuit
decodes the loop instruction which instructs to repeat a branch to
the branch target address the number of times equal to the loop
count, and stores the loop count designated by the loop
instruction; and the fetch address operation circuit includes a
circuit which decrements a value set in the loop counter each time
when a branch to the branch target address occurs, and outputs a
value obtained by incrementing the branch address by one
instruction length as a next fetch address when the value of the
loop counter reaches zero.
8. The data processing device as defined in claim 4, wherein: the
branch setting instruction includes a loop instruction which
designates a loop count; the branch information setting circuit
decodes the loop instruction which instructs to repeat a branch to
the branch target address the number of times equal to the loop
count, and stores the loop count designated by the loop
instruction; and the fetch address operation circuit includes a
circuit which decrements a value set in the loop counter each time
when a branch to the branch target address occurs, and outputs a
value obtained by incrementing the branch address by one
instruction length as a next fetch address when the value of the
loop counter reaches zero.
9. The data processing device as defined in claim 3, wherein: the
loop instruction has the branch target address which is fixed
relative to the loop instruction and also has no branch target
address information in an operand; and the branch information
setting circuit includes a circuit which calculates the value fixed
relative to the loop instruction and stores the calculated value in
the branch target address storage register.
10. The data processing device as defined in claim 4, wherein: the
loop instruction has the branch target address which is fixed
relative to the loop instruction and also has no branch target
address information in an operand; and the branch information
setting circuit includes a circuit which calculates the value fixed
relative to the loop instruction and stores the calculated value in
the branch target address storage register.
11. The data processing device as defined in claim 5, wherein: the
loop instruction has the branch target address which is fixed
relative to the loop instruction and also has no branch target
address information in an operand; and the branch information
setting circuit includes a circuit which calculates the value fixed
relative to the loop instruction and stores the calculated value in
the branch target address storage register.
12. The data processing device as defined in claim 6, wherein: the
loop instruction has the branch target address which is fixed
relative to the loop instruction and also has no branch target
address information in an operand; and the branch information
setting circuit includes a circuit which calculates the value fixed
relative to the loop instruction and stores the calculated value in
the branch target address storage register.
13. Electronic equipment comprising: the data processing device as
defined in claim 1; means for receiving input data; and means for
outputting a result of processing the input data by the data
processing device.
14. Electronic equipment comprising: the data processing device as
defined in claim 2; means for receiving input data; and means for
outputting a result of processing the input data by the data
processing device.
15. Electronic equipment comprising: the data processing device as
defined in claim 3; means for receiving input data; and means for
outputting a result of processing the input data by the data
processing device.
16. Electronic equipment comprising: the data processing device as
defined in claim 4; means for receiving input data; and means for
outputting a result of processing the input data by the data
processing device.
17. Electronic equipment comprising: the data processing device as
defined in claim 5; means for receiving input data; and means for
outputting a result of processing the input data by the data
processing device.
18. Electronic equipment comprising: the data processing device as
defined in claim 6; means for receiving input data; and means for
outputting a result of processing the input data by the data
processing device.
Description
[0001] Japanese Patent Application No. 2002-184379, filed on Jun.
25, 2002, is hereby incorporated by reference in its entirety.
BACKGROUND OF THE INVENTION
[0002] The present invention relates to a data processing device
and electronic equipment.
[0003] In pipeline architecture, if a branch instruction or a loop
instruction occurs, a branch target instruction must be newly
fetched for execution. Therefore, in pipeline architecture
consisting of "Fetch", "Decode", "Execution", and "Write", three
clock cycles are wasted each time a branch (including a branch in a
loop) occurs.
[0004] There may be a case where a branch prediction circuit is
provided for a microcomputer in order to prevent such a waste of
clock cycles. However, since tens of thousands of gates are
necessary for realizing the branch prediction circuit, adding the
branch prediction circuit inevitably increases the circuit scale
and cost.
BRIEF SUMMARY OF THE INVENTION
[0005] (1) According to a first aspect of the present invention,
there is provided a data processing device using pipeline control
including an instruction queue in which a plurality of instruction
codes are fetched. A fetch address operation circuit is provided
which calculates a fetch address used to fetch an instruction code
in the instruction queue. A fetch circuit is provided which fetches
an instruction code that is read out based on the fetch address
into the instruction queue. A branch information setting circuit is
provided which decodes a branch setting instruction instructing a
branch to a branch target address when the fetch address is a
branch address, stores the branch address in a branch address
storage register, and stores the branch target address in a branch
target address storage register. The fetch address operation
circuit includes a circuit which compares one of a previous fetch
address and an expected next fetch address with a value stored in
the branch address storage register, and then determines whether or
not to output a value stored in the branch target address storage
register as a next fetch address, based on the comparison
result.
[0006] (2) According to a second aspect of the present invention,
there is provided a data processing device using pipeline control
including an instruction queue in which a plurality of instruction
codes are fetched. A fetch address operation circuit is provided
which calculates a fetch address used to fetch an instruction code
in the instruction queue. A fetch circuit fetches an instruction
code that is read out based on the fetch address into the
instruction queue. A branch information setting circuit is provided
which decodes a branch setting instruction instructing a branch to
a branch target address when the fetch address is a branch address,
stores the branch address in a branch address storage register, and
stores the branch target address in a branch target address storage
register. The fetch address operation circuit includes a circuit
which compares an expected next fetch address obtained by
incrementing a value in a fetch program counter by one instruction
length with a value stored in the branch address storage register,
and then outputs a value stored in the branch target address
storage register as a next fetch address when the expected next
fetch address coincides with the value in the branch address
storage register, or outputs the expected next fetch address as a
next fetch address when the expected next fetch address does not
coincide with the value in the branch address storage register.
[0007] (3) According to a third aspect of the present invention,
there is provided electronic equipment including any of the above
data processing devices.
[0008] Furthermore, a means for receiving input data and a means
for outputting a result of processing the input data by the data
processing device is provided.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING
[0009] FIGS. 1A and 1B are views illustrating a conventional
microcomputer (or a data processing device in a broad sense) using
pipeline control in which a branch occurs.
[0010] FIG. 2 is a functional block diagram showing a microcomputer
(or a data processing device in a broad sense) according to one
embodiment of the present invention.
[0011] FIGS. 3A, 3B, and 3C are diagrams for illustrating a branch
setting instruction in one embodiment of the present invention.
[0012] FIG. 4 is a diagram for illustrating a fetch address
operation circuit.
[0013] FIGS. 5A, 5B, and 5C are diagrams for illustrating a loop
instruction according to one embodiment of the present
invention.
[0014] FIG. 6 is a diagram for illustrating a fetch address
operation circuit which can process a loop instruction according to
one embodiment of the present invention.
[0015] FIGS. 7A, 7B, and 7C are diagrams for illustrating another
loop instruction according to one embodiment of the present
invention. In this loop instruction, a branch target address is an
instruction subsequent to the loop instruction.
[0016] FIGS. 8A, 8B, and 8C are block diagrams showing examples of
the electronic equipment having a microcomputer.
[0017] FIGS. 9A, 9B, and 9C show external views of various types of
electronic equipment.
DETAILED DESCRIPTION OF THE EMBODIMENT
[0018] Embodiments of the present invention are described
below.
[0019] Note that the embodiments described below do not in any way
limit the scope of the invention laid out in the claims herein. In
addition, all elements of the embodiments described below should
not be taken as essential requirements of the present
invention.
[0020] (1) According to one embodiment the present invention, there
is provided a data processing device using pipeline control
including an instruction queue in which a plurality of instruction
codes are fetched. A fetch address operation circuit is provided
which calculates a fetch address used to fetch an instruction code
in the instruction queue. A fetch circuit is provided which fetches
an instruction code that is read out based on the fetch address
into the instruction queue. A branch information setting circuit is
provided which decodes a branch setting instruction instructing a
branch to a branch target address when the fetch address is a
branch address, stores the branch address in a branch address
storage register, and stores the branch target address in a branch
target address storage register. The fetch address operation
circuit includes a circuit which compares one of a previous fetch
address and an expected next fetch address with a value stored in
the branch address storage register, and then determines whether or
not to output a value stored in the branch target address storage
register as a next fetch address, based on the comparison
result.
[0021] The branch setting instruction includes branch address
specifying information and branch target address specifying
information in operands thereof either explicitly or
implicitly.
[0022] The branch address specifying information can be a value
capable of specifying a location at which a branch occurs. The
branch address specifying information may be an address of an
instruction at which a branch occurs. The branch address specifying
information may be a relative address between the branch setting
instruction address and the branch address, or the number of
instructions present between the branch setting instruction address
and the branch address. The branch target address specifying
information may be a number of a general-purpose register in which
one of the above values is stored.
[0023] Similarly, the branch target address specifying information
can be a value capable of specifying the branch target. For
example, the branch target address specifying information may be an
address of a branch target instruction. The branch target address
specifying information may be a relative address between the branch
setting instruction address and the branch target address, or the
number of instructions present between the branch setting
instruction address and the branch target address. The branch
target address specifying information may be a number of a
general-purpose register in which one of these values is
stored.
[0024] Furthermore, the previous fetch address may be a previous
fetch address in a fetch program counter. The expected next fetch
address is a fetch address calculated based on the previous fetch
address. If either the previous fetch address or the expected next
fetch address coincides with the value stored in the branch address
storage register, a value stored in the branch target address
storage register may be output as a next fetch address. If the
previous fetch address and the expected next fetch address do not
coincide with the value stored in the branch address storage
register, a value obtained by incrementing the value of the fetch
program counter by one instruction length may be output as a next
fetch address.
[0025] If the branch address designated by the branch setting
instruction is stored in the branch address storage register before
one of the previous fetch address in the fetch program counter and
the expected next fetch address calculated based on the previous
fetch address coincides with the value stored in the branch address
storage register, the branch target instruction can be fetched in
pipeline control without a loss when a branch occurs.
[0026] According to this data processing device, architecture
including the branch setting instruction in the instruction set can
be implemented by merely adding the fetch address operation circuit
and the branch information setting circuit, and a high
cost-performance data processing device which employs pipeline
architecture and is capable of reducing a time loss due to
occurrence of a branch without increasing the circuit scale.
[0027] (2) According to another embodiment of the present
invention, there is provided a data processing device using
pipeline control including an instruction queue in which a
plurality of instruction codes are fetched. A fetch address
operation circuit is provided which calculates a fetch address used
to fetch an instruction code in the instruction queue. A fetch
circuit is provided which fetches an instruction code that is read
out based on the fetch address into the instruction queue. A branch
information setting circuit is provided which decodes a branch
setting instruction instructing a branch to a branch target address
when the fetch address is a branch address, stores the branch
address in a branch address storage register, and stores the branch
target address in a branch target address storage register. The
fetch address operation circuit includes a circuit which compares
an expected next fetch address obtained by incrementing a value in
a fetch program counter by one instruction length with a value
stored in the branch address storage register, and then outputs a
value stored in the branch target address storage register as a
next fetch address when the expected next fetch address coincides
with the value in the branch address storage register, or outputs
the expected next fetch address as a next fetch address when the
expected next fetch address does not coincide with the value in the
branch address storage register.
[0028] The branch setting instruction includes branch address
specifying information and branch target address specifying
information in operands thereof either explicitly or
implicitly.
[0029] The branch address specifying information can be a value
capable of specifying a location at which a branch occurs. For
example, branch address specifying information may be an address of
an instruction at which a branch occurs. The branch address
specifying information may be a relative address between the branch
setting instruction address and the branch address, or the number
of instructions present between the branch setting instruction
address and the branch address. The branch target address
specifying information may be a number of a general-purpose
register in which one of the above values is stored.
[0030] Similarly, the branch target address specifying information
can be a value capable of specifying the branch target. The branch
target address specifying information may be an address of a branch
target instruction. The branch target address specifying
information may be a relative address between the branch setting
instruction address and the branch target address, or the number of
instructions present between the branch setting instruction address
and the branch target address. The branch target address specifying
information may be a number of a general-purpose register in which
one of these values is stored.
[0031] If the branch address designated by the branch setting
instruction is stored in the branch address storage register before
the expected next fetch address obtained by incrementing the value
in the fetch program counter by the instruction length coincides
with the value stored in the branch address storage register, the
branch target instruction can be fetched in pipeline control
without a loss when a branch occurs.
[0032] According to this data processing device, architecture
including the branch setting instruction in the instruction set can
be implemented by merely adding the fetch address operation circuit
and the branch information setting circuit, and a high
cost-performance data processing device which employs pipeline
architecture and is capable of reducing a time loss due to
occurrence of a branch can be provided without increasing the
circuit scale.
[0033] (3) In any one of the data processing devices, the branch
setting instruction may include a loop instruction which designates
a loop count. The branch information setting circuit may decode the
loop instruction which instructs to repeat a branch to the branch
target address the number of times equal to the loop count, and
store the loop count designated by the loop instruction. The fetch
address operation circuit may include a circuit which outputs a
value stored in the branch target address storage register as a
next fetch address until the number of times the branch to the
branch target address repeats reaches the loop count.
[0034] The loop count included in the loop instruction may be set
in a loop counter, and the loop counter may be decremented each
time when a branch to the branch target address occurs, for
example. A branch to the branch target address may be terminated
when the value of the loop counter reaches zero.
[0035] According to this data processing device, since the branch
target address can be output as a fetch address after the loop
instruction is executed, the branch target instruction can be
fetched without causing a loss in pipeline control due to a branch
by loop.
[0036] Moreover, since it is unnecessary to use a SUB instruction
for decrementing the loop counter which is necessary for the loop
using no loop instruction, the execution time can be increased.
[0037] According to this data processing device, architecture
including the loop instruction in the instruction set can be
implemented by merely adding the fetch address operation circuit
and the branch information setting circuit, and a high
cost-performance data processing device which employs pipeline
architecture and is capable of reducing a time loss due to
occurrence of a loop and can be provided without increasing the
circuit scale.
[0038] (4) In any one of the data processing devices, the branch
setting instruction may include a loop instruction which designates
a loop count. The branch information setting circuit may decode the
loop instruction which instructs to repeat a branch to the branch
target address the number of times equal to the loop count, and
store the loop count designated by the loop instruction. The fetch
address operation circuit may include a circuit which decrements a
value set in the loop counter each time when a branch to the branch
target address occurs, and output a value obtained by incrementing
the branch address by one instruction length as a next fetch
address when the value of the loop counter reaches zero.
[0039] According to this data processing device, since the branch
target address can be output as a fetch address after the loop
instruction is executed, the branch target instruction can be
fetched without causing a loss in pipeline control due to a branch
by loop.
[0040] Moreover, since it is unnecessary to use a SUB instruction
for decrementing the loop counter which is necessary for the loop
using no loop instruction, the execution time can be increased.
[0041] According to this data processing device, architecture
including the loop instruction in the instruction set can be
implemented by merely adding the fetch address operation circuit
and the branch information setting circuit, and a high
cost-performance data processing device which employs pipeline
architecture capable of reducing a time loss due to occurrence of a
loop without increasing the circuit scale.
[0042] (5) In any one of the data processing devices, the loop
instruction may have the branch target address which is fixed
relative to the loop instruction and also have no branch target
address information in an operand. The branch information setting
circuit may include a circuit which calculates the value fixed
relative to the loop instruction and store the calculated value in
the branch target address storage register.
[0043] The term "fixed relative to the loop instruction" means that
a relative distance from the loop instruction can be determined
uniquely.
[0044] For example, the branch target address may be fixed relative
to an instruction subsequent to the loop instruction.
[0045] According to this data processing device, since information
in operands of the loop instruction can be reduced, the device is
effective in the case where the loop instruction must be written
with a short instruction code length.
[0046] (6) According to still another embodiment of the present
invention, there is provided electronic equipment including any one
of the above data processing devices. A means for receiving input
data is provided and a means for outputting a result of processing
the input data by the data processing device is provided.
[0047] According to the present invention, high-function electronic
equipment can be provided at low cost since the electronic
equipment is equipped with an inexpensive high-speed data
processing device.
[0048] Preferred embodiments of the present invention are described
below in detail with reference to the drawings.
1. Data Processing Device
(1) Configuration of Data Processing Device
[0049] FIGS. 1A and 1B are views illustrating a conventional
example in the case where a branch occurs in a microcomputer (data
processing device in a broad sense) which controls pipeline
processing.
[0050] FIG. 1A shows an example of an instruction list which causes
a branch to occur.
[0051] FIG. 1B shows a timetable in which instructions (i) to (iii)
are processed in a microcomputer which controls five stages of
pipeline processing consisting of Fetch (F), Decode (D), Execute
(E), Access (A), and Write (W).
[0052] A fetch address in the Fetch (F) stage of the pipeline
control is generally given by "present fetch address=previous fetch
address+size of instruction fetched last time". Therefore, in the
case where a branch instruction occurs, a branch target instruction
has not been fetched in the fetch queue.
[0053] The instruction (ii) shown in FIG. 1B is a branch
instruction and the instruction (iii) is a branch target
instruction. Since the branch target instruction must be fetched
after the branch instruction is executed, a delay inevitably occurs
for two clock cycles as indicated by reference numeral 210.
[0054] In the case where loop processing is performed by using a
branch instruction, since the instruction (i) sub is necessary for
counting, one additional clock cycle is needed.
[0055] FIG. 2 is a functional block diagram illustrating a
configuration of a microcomputer (data processing device in a broad
sense) in the present embodiment.
[0056] A microcomputer in the present embodiment includes a CPU
(processing circuit in a broad sense) 10 and a BCU (bus control
unit) 100. The microcomputer may also include various types of
peripheral circuits such as a ROM (Read Only Memory), RAM (Random
Access Memory), MMU (Memory Management Unit), DMAC (Direct Memory
Access Controller), LCD (Liquid Crystal Display) driver, and SIO
(Serial Input Output).
[0057] The CPU 10 controls pipeline processing for instruction
fetching, instruction decoding, operation processing, writing into
a register, and the like according to a program stored in a ROM or
RAM. The CPU 10 handles 32-bit width data, but processes a 16-bit
instruction code.
[0058] The BCU 100 controls various buses such as a 32-bit
instruction data bus, an instruction address bus for instruction
data access, a 32-bit data bus, a data address bus for data access,
and a control bus for control signals (not shown).
[0059] The CPU 10 exchanges signals with the outside through these
various buses.
[0060] The CPU 10 includes a fetch circuit 20, a fetch address
operation circuit 40, a decoder circuit 60, an execution circuit
70, a register file 80 (general-purpose register 82 and special
register 84), and the like.
[0061] The fetch circuit 20 prefetches instruction codes of a
plurality of instructions in instruction queues (prefetch queues,
for example) 30-0, 30-1 . . . .
[0062] The fetch address operation circuit 40 is a circuit which
calculates a fetch address for fetching the instruction code in the
instruction queue. The fetch address operation circuit 40 includes
a branch address storage register 42 and a branch target address
storage register 44. The fetch address operation circuit 40 may
compare either the previous fetch address or the expected next
fetch address with the value stored in the branch address storage
register, and output the value stored in the branch target address
storage register as the next fetch address based on the comparison
results.
[0063] The fetch address operation circuit 40 may compare the
expected next fetch address obtained by incrementing a value
retained in a fetch program counter by a number corresponding to
one instruction length with the value stored in the branch address
storage register. The fetch address operation circuit 40 may output
the value stored in the branch target address storage register as
the next fetch address if the expected next fetch address coincides
with the value stored in the branch address storage register, and
output a value obtained by incrementing the value retained in the
fetch program counter by a number corresponding to one instruction
length as the next fetch address if the expected next fetch address
does not coincide with the value stored in the branch address
storage register.
[0064] The fetch address operation circuit 40 may output the value
stored in the branch target address storage register 44 as the next
fetch address based on the comparison results until the number of
branches to the branch target address reaches a loop count.
[0065] For example, the fetch address operation circuit 40 may
include a loop counter circuit 46 which sets the loop count set in
the loop instruction in a loop counter (not shown), and decrements
the loop counter each time a branch to the branch target address
occurs.
[0066] The decoder circuit 60 is a circuit which is connected with
the instruction queue through a signal line, and performs
processing which allows the instruction code fetched in the
instruction queue to be input thereto and decodes the input
instruction code. The decoder circuit 60 includes a branch
information setting circuit 62.
[0067] The branch information setting circuit 62 is connected with
the branch address storage register 42 and the branch target
address storage register 44 through a signal line. The branch
information setting circuit 62 decodes a branch setting instruction
which causes a branch to the branch target address to occur at the
branch address, stores the branch address in the branch address
storage register, and stores the branch target address in the
branch target address storage register.
[0068] The branch information setting circuit 62 may be connected
with a loop count setting register (not shown) provided to the loop
counter circuit 46 through a signal line. The branch information
setting circuit 62 may decode a loop instruction which causes a
branch to the branch target address to occur at the branch address
before reaching the loop count, fetch the loop count set in the
loop instruction, and store the loop count in the loop count
setting register.
[0069] The execution circuit 70 executes an instruction based on
operation contents of the instruction decoded by the decoder
circuit 60. The execution circuit 70 includes a data operation
circuit 72 which calculates data and an address operation circuit
74 which calculates an address, and executes the function of the
instruction decoded by the decoder circuit 60 while optionally
accessing the general-purpose register 82 or a memory (RAM or the
like).
(2) Branch Setting Instruction
[0070] FIGS. 3A, 3B, and 3C are views illustrating an example of a
branch setting instruction in the present embodiment.
[0071] FIG. 3A shows an example of an instruction code of the
branch setting instruction. FIG. 3B is a view showing a bit field
of the instruction code of the branch setting instruction. An
instruction code 310 of the branch setting instruction includes an
8-bit operation code (opcode) designation region 312 from the bit
15 to the bit 8, a 4-bit branch address designation region 314 from
the bit 7 to the bit 4, and a 4-bit branch target address
designation region 316 from the bit 3 to the bit 0.
[0072] An opcode which indicates that the instruction is a branch
setting instruction is stored in the opcode designation region
312.
[0073] A value which indicates whether the branch address or the
branch target address is after a certain number of instructions
from the branch setting instruction may be designated in the branch
address designation region 314 or the branch target address
designation region 316. In the case where the value which indicates
whether the branch target address is after a certain number of
instructions from the branch setting instruction is designated by
immediate data "imm4", for example, the branch target address is
given by "PC (program counter)+instruction
length+imm4.times.instruction length". The branch address or the
branch target address may be designated by either a relative
address from the branch setting instruction or an absolute address.
A value for a general-purpose register in which the above value is
stored may be designated in the branch address designation region
314 or the branch target address designation region 316.
[0074] FIG. 3C shows a program list including the branch setting
instruction. In the case where a branch setting instruction "jppr
x, y" is placed at 320, a branch to P2 (branch target address)
after y instructions (see 340) from the branch setting instruction
occurs at P1 (branch address) after x instructions (see 330) from
the branch setting instruction.
[0075] FIG. 4 is a view illustrating an example of a configuration
of the fetch address operation circuit.
[0076] The fetch address operation circuit 40 includes a fetch
program counter (FPC) 41, a branch address register 42, a branch
target address register 44, and the like. The fetch program counter
(FPC) 41 is a program counter in which an instruction address to be
fetched is stored, and is updated to the next fetch address each
time the instruction is fetched.
[0077] A branch address designated in the branch setting
instruction is retained in the branch address storage register
42.
[0078] A branch target address designated in the branch setting
instruction is retained in the branch target address storage
register 44.
[0079] When the instruction code is fetched in the fetch queue
based on the fetch address retained in the fetch program counter
(FPC) 41, an address calculator 43 outputs an increment value
signal 55 of a fetch address counter to a multiplexer 47 and a
comparator 45 based on a fetch address 53 and an instruction code
length 54.
[0080] The comparator 45 allows the increment value signal 55 of
the fetch address counter and the branch address stored in the
branch address register 42 to be input thereto, and sets a branch
signal 46 to be output to the multiplexer 47 to ON ("H" level, for
example) if the increment value signal 55 coincides with the branch
address.
[0081] The branch target address storage register 44 outputs a
branch target address signal 56 to the multiplexer 47.
[0082] An ALU 48 calculates a branch target address generated by a
normal branch instruction (excluding the branch setting instruction
in the present embodiment) or the like, and outputs a normal branch
address signal 57 to the multiplexer 47. A normal branch signal 49
is a signal which notifies the multiplexer 47 of occurrence of a
normal branch instruction (excluding the branch setting instruction
in the present embodiment) or the like.
[0083] The multiplexer 47 chooses the branch target address signal
56 if the branch signal 46 is ON ("H" level, for example), and
outputs the branch target address signal 56 to the fetch program
counter (FPC) 41 as a next fetch address signal 52. The multiplexer
47 chooses the normal branch address signal 57 if the normal branch
signal 49 is ON ("H" level, for example), and outputs the normal
branch address signal 57 to the fetch program counter (FPC) 41 as
the next fetch address signal 52. The multiplexer 47 chooses the
increment value signal 55 of the fetch address counter if both the
branch signal 46 and the normal branch signal 49 are OFF ("L"
level, for example), and outputs the increment value signal 55 to
the fetch program counter (FPC) 41 as the next fetch address signal
52.
(3) Loop Instruction
[0084] FIGS. 5A, 5B, and 5C are views illustrating an example of a
loop instruction in the present embodiment.
[0085] FIG. 5A shows an example of an instruction code of the loop
instruction. FIG. 5B is a view showing a bit field of the
instruction code of the loop instruction. An instruction code 410
of the loop instruction includes a 7-bit opcode designation region
412 from the bit IS to the bit 9, a 3-bit branch address
designation region 414 from the bit 8 to the bit 6, a 3-bit branch
target address designation region 416 from the bit 5 to the bit 3,
and a 3-bit loop count designation region 418 from the bit 2 to the
bit 0.
[0086] An opcode which indicates that the instruction is a loop
instruction is stored in the opcode designation region 412.
[0087] A value which indicates whether the branch address or the
branch target address is after a certain number of instructions
from the branch setting instruction may be designated in the branch
address designation region 414 or the branch target address
designation region 416. In the case where the value which indicates
whether the branch target address is after a certain number of
instructions from the branch setting instruction is designated by
immediate data "imm3", for example, the branch target address is
given by "PC (program counter)+instruction
length+imm3.times.instruction length". The branch address or the
branch target address may be designated by either a relative
address from the branch setting instruction or an absolute address.
A value for a general-purpose register in which the above value is
stored may be designated in the branch address designation region
414 or the branch target address designation region 416.
[0088] A loop count may be designated in the loop count designation
region 418 by immediate data, or a value for a general-purpose
register in which the loop count is set may be designated in the
loop count designation region 418.
[0089] FIG. 5C shows a program list including the loop instruction.
In the case where a loop instruction "loop x, y, z" is placed at
420, a branch occurs at P1 (branch address) after x instructions
(see 430) from the loop instruction, and processing which causes a
branch to P2 (branch target address) after y instructions (see 440)
from the loop instruction to occur is repeated z times (z=loop
count). After z branches have occurred, an instruction 470
subsequent to the branch address is executed.
[0090] FIG. 6 is a view illustrating an example of a configuration
of a fetch address operation circuit which can process the loop
instruction.
[0091] A fetch address operation circuit 40' shown in FIG. 6
includes a loop counter circuit 60 in addition to the configuration
of the fetch address operation circuit 40 shown in FIG. 4.
[0092] The loop counter circuit 60 includes a loop counter 62, a
calculator 64, and a comparator 66. A loop count designated by the
loop instruction is retained in the loop counter 62.
[0093] The calculator 64 allows the value retained in the loop
counter and the branch signal 46 output from the comparator 45 to
be input thereto, decrements the loop counter (processing which
subtracts "1" from the value of the loop counter) if the branch
signal 46 indicates occurrence of a branch (branch signal 46 is ON,
"H" level, for example), and outputs a decrement value signal 65 to
the loop counter 62 and the comparator 66. This allows the value of
the loop counter 62 to be decremented.
[0094] If the decrement value signal 65 becomes "0", the comparator
66 judges that branches have occurred a number of times
corresponding to the designated loop count, sets a loop end signal
68 to ON, and outputs the loop end signal 68 to the multiplexer 47.
The loop end signal 68 is a signal which notifies the multiplexer
47 that the number of loops designated by the loop instruction have
been completed. The loop end signal 68 may be judged to be ON if
the signal at an "H" level is output, for example.
[0095] If the loop end signal 68 is not ON, the multiplexer 47
performs the operation described with reference to FIG. 4. If the
loop end signal 68 is ON, the multiplexer 47 chooses the increment
value signal 55 of the fetch address counter instead of the branch
target address signal 56 even if the branch signal 46 is ON, and
outputs the increment value signal 55 to the fetch program counter
(FPC) 41 as the next fetch address signal 52.
[0096] FIGS. 7A, 7B, and 7C are views illustrating another loop
instruction in the present embodiment. This other loop instruction
is a loop instruction of the type in which the branch target
address is an instruction subsequent to the loop instruction.
[0097] FIG. 7A shows an example of an instruction code of another
loop instruction (loop instruction 2). FIG. 7B is a view showing a
bit field of the instruction code of the loop instruction 2. An
instruction code 510 of the loop instruction 2 includes an 8-bit
opcode designation region 512 from the bit 15 to the bit 8, a 4-bit
branch address designation region 514 from the bit 7 to the bit 4,
and a 4-bit loop count designation region 518 from the bit 3 to the
bit 0. Since the branch target is fixed in the loop instruction 2,
it is unnecessary to designate the branch target address in the
operand.
[0098] An opcode which indicates that the instruction is a loop
instruction of the type in which the branch target address is an
instruction subsequent to the loop instruction is stored in the
opcode designation region 512.
[0099] A value which indicates whether the branch address is after
a certain number of instructions from the branch setting
instruction may be designated in the branch address designation
region 514. In the case where the value which indicates whether the
branch address is after a certain number of instructions from the
branch setting instruction is designated by immediate data "imm4",
for example, the branch address is given by "PC (program
counter)+instruction length+imm4.times.instruction length". The
address may be designated by either a relative address from the
branch setting instruction or an absolute address. A value for a
general-purpose register in which the above value is stored may be
designated in the branch address designation region 514.
[0100] The loop count may be designated in the loop count
designation region 518 by immediate data, or a value for a
general-purpose register in which the loop count is set may be
designated in the loop count designation region 518.
[0101] FIG. 7C shows a program list including the loop instruction.
In the case where a loop instruction 2 "loop x, z" is placed at
520, a branch occurs at P1 (branch address) after x instructions
(see 530) from the loop instruction 2, and processing which causes
a branch to P2 (branch target address) subsequent to the loop
instruction 2 to occur is repeated z times (loop count). After z
branches have occurred, an instruction 570 subsequent to the branch
address is executed.
[0102] In the case where the branch target is fixed in this manner,
the fixed branch target address may be set in the branch target
address storage register 44 shown in FIG. 6.
[0103] The above description is given for a case where the branch
target is affixed to the instruction subsequent to the loop
instruction. However, the present invention is not limited thereto.
For example, the branch target may be affixed to another
location.
2. Electronic Equipment
[0104] Electronic equipment including the above-described data
processing device (microcomputer, for example) is described
below.
[0105] FIG. 8A is an internal block diagram of a car navigation
system which is an example of electronic equipment. FIG. 9A is an
external view of the car navigation system. The car navigation
system is operated by using a remote controller 710. A position
detection section 720 detects the position of a vehicle based on
information from the GPS (global positioning system) or a
gyroscope. Information such as a map is stored in a CD-ROM 730
(information storage medium). A memory 740 becomes a work area
during image processing or voice processing. An image generated is
displayed for the driver by using an image output section 750.
Guide voice for car navigation is generated and output for the
driver by using a sound output section 735. A microcomputer 700
inputs information thereto from information sources such as the
remote controller 710, the position detection section 720, and the
CD-ROM 730, performs various types of processing, and outputs the
processed information by using output devices such as the image
output section 750 and the sound output section 735.
[0106] FIG. 8B shows an internal block diagram of a game device
which is another example of electronic equipment. FIG. 9B is an
external view of the game device. The game device generates a game
image and game sound by using a memory 790 as a work area based on
operation information of the player from a game controller 760, a
game program from a CD-ROM 770, player information from an IC card
780, and the like, and outputs the game image and game sound by
using an image output section 810 and a sound output section
800.
[0107] FIG. 8C shows an internal block diagram of a printer which
is still another example of electronic equipment. FIG. 9C is an
external view of the printer. This printer generates a print image
by using a bitmap memory 850 as a work area based on operation
information from an operating panel 820 and character information
from a code memory 830 and a font memory 840, and outputs the image
by using a print output section 860. The printer informs the user
of the state and the mode of the printer by using a display panel
870.
[0108] As other examples of electronic equipment to which the
microcomputer can be applied are a cellular phone, PHS, pager,
portable information terminal, digital camera, hard disk device,
optical disc (CD and DVD) device, magneto-optical disc (MO) device,
audio equipment, electronic notebook, electronic calculator, POS
terminal, device equipped with a touch panel, projector, word
processor, personal computer, TV, and view finder type or direct
viewfinder type video tape recorder can be given.
[0109] The present invention is not limited to the above
embodiments. Various modifications and variations are possible
within the spirit and scope of the present invention.
[0110] The configuration of the electronic equipment of the present
invention is not limited to the configurations described with
reference to FIGS. 8A to 8C and FIGS. 9A to 9C. Various
modifications and variations are possible within the spirit and
scope of the present invention.
* * * * *