U.S. patent application number 09/816120 was filed with the patent office on 2001-08-02 for data transfer control method, and peripheral circuit, data processor and data processing system for the method.
This patent application is currently assigned to Hitachi, Ltd.. Invention is credited to Takeda, Hiroshi.
Application Number | 20010011311 09/816120 |
Document ID | / |
Family ID | 12069881 |
Filed Date | 2001-08-02 |
United States Patent
Application |
20010011311 |
Kind Code |
A1 |
Takeda, Hiroshi |
August 2, 2001 |
Data transfer control method, and peripheral circuit, data
processor and data processing system for the method
Abstract
A memory 1 performs its internal operation in response to access
requests (200, 201 and 202) of a CPU 2 in synchronism with the
oscillated output of a self-excited oscillator 102 incorporated
therein and according to said access requests, and outputs a
response request 103 for said access requests to said CPU in
synchronism with its internal operation. The CPU performs the
access requests for the memory, and fetches data from the outside
or outputs the data to the outside in response to and in
synchronism with the response request 103 from the accessed memory
and according to the kinds of said access requests.
Inventors: |
Takeda, Hiroshi; (Tokyo,
JP) |
Correspondence
Address: |
MATTINGLY, STANGER & MALUR, P.C.
104 E HUME AVE
ALEXANDRIA
VA
22301
US
|
Assignee: |
Hitachi, Ltd.
|
Family ID: |
12069881 |
Appl. No.: |
09/816120 |
Filed: |
March 26, 2001 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
09816120 |
Mar 26, 2001 |
|
|
|
09400999 |
Sep 21, 1999 |
|
|
|
6247073 |
|
|
|
|
09400999 |
Sep 21, 1999 |
|
|
|
08896473 |
Jul 18, 1997 |
|
|
|
5978891 |
|
|
|
|
08896473 |
Jul 18, 1997 |
|
|
|
08369722 |
Jan 6, 1995 |
|
|
|
5673398 |
|
|
|
|
Current U.S.
Class: |
710/61 ; 711/167;
713/400; 713/500; 713/600 |
Current CPC
Class: |
G11C 7/22 20130101; G11C
7/1027 20130101; G06F 13/1689 20130101 |
Class at
Publication: |
710/61 ; 713/400;
713/500; 713/600; 711/167 |
International
Class: |
G06F 013/36 |
Foreign Application Data
Date |
Code |
Application Number |
Jan 21, 1994 |
JP |
6-21969 |
Claims
What is claimed is:
1. A method of controlling a data transfer between a data processor
and its peripheral circuit, comprising the steps of: sending an
access request to a peripheral circuit from said data processor;
allowing said peripheral circuit to perform an internal operation
according to the access request in synchronism with the output from
a self-excited oscillator incorporated therein and in a manner to
understand the content of said access request; sending a response
request to said peripheral circuit from said data processor in
synchronism with its internal operation in response to said access
request; and transferring data to or from said data processor
depending on the type of said access request in synchronism with
said response request.
2. A data transfer controlling method according to claim 1, wherein
said access request contains the information indicating both the
peripheral circuit selected as an object to be accessed and the
transfer direction of the data, and wherein said response request
is a signal varying in synchronism with the internal operation of
said peripheral circuit.
3. A peripheral circuit comprising: a cycle timing generator for
producing an access cycle signal of an internal operation in
response to an access request from a data processor and on the
basis of an output of a self-excited oscillator; an external
terminal for outputting said access cycle signal as a response
request in response to said access request; and an internal timing
generator for producing an internal operation timing signal in
synchronism with said access cycle signal.
4. A peripheral circuit according to claim 3, wherein said
peripheral circuit is constructed as a memory including: a memory
cell array having a plurality of memory cells arranged in a matrix
form; a selector for selecting the memory cells on the basis of an
external address signal; a read circuit for reading the data from
the selected memory cell; and a counter for counting the number of
continuous data words read from said memory cell array on the basis
of a change in said access cycle signal, to stop the oscillation of
said self-excited oscillator when the counted result reaches a
predetermined count value.
5. A peripheral circuit according to claim 4, wherein said counter
has a parameter register for latching said predetermined count
value presettably from the outside.
6. A data processor comprising: an access control circuit for
sending an access request to a peripheral circuit; an input
terminal for receiving a response request from the peripheral
circuit in response to said access request; a buffer memory capable
of writing/reading on the basis of said response request and in
synchronism with an internal operation clock; and an arithmetic
circuit and an input/output buffer circuit connected with said
buffer memory.
7. A data processor according to claim 6, further comprising: a
central processing unit for controlling said access control
circuit, said buffer memory, said arithmetic circuit and said
input/output buffer circuit.
8. A data processor according to claim 7, wherein said buffer
memory includes a counter for detecting the number of continuous
read access requests to said peripheral circuit from said access
control circuit, in terms of the number of changes in said response
request, whereby said central processing unit receives said
detection result as the information meaning the complete of the
data acquisition by said access request.
9. A data processor comprising: an access control circuit for
sending an access request to a peripheral circuit; an input
terminal for receiving a response request from the peripheral
circuit in response to said access request; a buffer memory having
an asynchronous port for writing/reading on the basis of said
response request, and a synchronous port for writing/reading in
synchronism with an internal operation clock; an arithmetic circuit
connected with the synchronous port of said buffer memory; and an
input/output buffer circuit connected with the asynchronous port of
said buffer memory for interfacing with external devices.
10. A data processor according to claim 9, further comprising: a
central processing unit for controlling said access control
circuit, said buffer memory, said arithmetic circuit and said
input/output buffer circuit.
11. A data processor according to claim 10, wherein said buffer
memory includes a counter for detecting the number of continuous
read access requests sent to said peripheral circuit from said
access control circuit, in terms of the number of changes in said
response request, whereby said central processing unit receives
said detection result as the information meaning the complete of
the data acquisition by said access request.
12. A data processing system comprising: a peripheral circuit that
performs an internal operation according to an access request from
a data processor and in synchronism with the output of a
self-excited oscillator incorporated therein, and outputs a
response request to said data processor in synchronism with said
internal operation in response to said access request; a data
processor that sends an access request to a desired peripheral
circuit and receives a response request from the desired peripheral
circuit, thereby transferring data to or from the peripheral
circuit depending on the type of said access request; and a bus
connecting said data processor and said peripheral circuit.
13. A data processing system according to claim 12, wherein a
plurality of peripheral circuits of different kinds are commonly
connected with said bus such that their output terminals for
response requests share the single input terminal for response
requests of said data processor.
Description
BACKGROUND OF THE INVENTION
[0001] The present invention relates to a method of controlling the
transfer of a variety of information such as instruction
information or data information between a memory or a peripheral
circuit and a data processor, and a peripheral circuit, a data
processor and a data processing system using the method, and, more
particularly, to a technique which is especially effective if
applied to the control technique of the data transfer between the
data processor and a memory. Incidentally, the data processor in
the present Specification will cover the general concept of a CPU
(i.e., Central Processing Unit), a microprocessor, a microcomputer,
a single-chip microcomputer, a digital-signal processor or a direct
memory access controller.
SUMMARY OF THE INVENTION
[0002] Some RISC processors of the prior art include one or more
cache memories in a chip from the point of view of performance,
cost, manufacture process and technical level of the LSI. Such a
CPU is connected with a number of memories and input/output)
circuits on a circit board to construct a system. It is usual to
use an operation clock (or system clock) as a reference to the
operation of the system. Usually, the peripheral circuits such as
the memories and the input/output circuits to constitute the system
are individually given different functions and characteristics to
have individually different operating procedures, response times or
operating speeds. It is needless to say that the CPU interfaces
owned by the memories and the input/output circuits are frequently
different from one another although they have some similarity in
the functions or timings.
[0003] As to the differences in such functions, operating speeds
and interface specifications, memory controllers are used for the
memories, and I/O controllers are used for the input/output
circuits. These controllers have functions, as roughly divided into
the following two points.
[0004] The first function is to inform the memories and the
input/output circuits of which memory or input/output circuit is
selected by the CPU, and initiates a data transfer. This function
can be regarded as the so-called "chip selection" or "chip enable
control". For example, logic operations between the signals
indicating the kinds of addressing and access are carried out to
produce pulses or level signals by using an operation clock or the
like thereby to activate only the memory selected or the signal
connected with the input/output circuit.
[0005] The second function is to count the operation clocks by a
counter thereby to produce a signal demanding the CPU for an
extension of the access period at the unit of the operation clock
for the wait or ready operation. Under to the rule of confirming
the signal for each operation clock by the CPU, the difference in
the timing or the operation speed between the CPU and the memory or
the peripheral circuit is absorbed to realize the data transfer
without fail. This function is the so-called "wait state control
function".
[0006] However, we have revealed that the aforementioned wait state
function by the controller has the following problems.
[0007] (1) Since the duration of the data transfer time to be
extended by the wait state is always determined at the operation
clock unit of the system, it is impossible to sufficiently extract
the performance intrinsic to the memory or peripheral circuit.
Moreover, it is substantially impossible to design the system by
using the performance, which is based upon the design data
submitted by the maker/seller as to the memory or the input/output
circuit, in the limit state. Since a certain operation margin is
considered, a data transfer involves idle time in most cases so
that the data transfer efficiency on the data bus decreases. This
problem applies not only to the case in which the system is
constructed on a circuit board, i.e., in which the connections
between the memories or the input/output circuits and the CPU are
made through the buses on the board, but also to the case in which
the CPU and the memories are formed over a common semiconductor
chip. Specifically, if an optimum design were to made considering
the electric characteristics and the arrangement of circuit
elements, the controllers and the memories could effect the data
transfer efficiently to the operation clock of the controllers. In
the actual circuit design, however, a delicate timing has to be
made in the chip although not easy, while considering the
characteristics of the individual logic circuit blocks.
[0008] (2) The aforementioned wait state control takes serious
troubles because the designer has to design the system for the
individual memories or input/output circuits, if in plurality, due
to the differences in the functions (including the protocol) and
performances.
[0009] (3) The circuit portions required for the wait state control
have to cover the sets of memories and input/output circuits, thus
causing difficulties in the high speed, the small size and the low
prices such as the complicatedness of the system, the increase in
the part number or the increase in the load upon the signal
line.
[0010] (4) As has been described in the aforementioned problem (1),
the wait state control cannot sufficiently extract the performances
intrinsic to the memories and the peripheral circuits so that it
limits the speed-up of the operations. In order to eliminate this
limit, therefore, all or the highly efficient memories or
input/output circuits could be connected without the wait state
control. If, however, the operation clock of the controller is
suppressed according to the characteristics such as the operation
speed of the memories and the input/output circuits, the controller
such as the CPU has a tendency to have its operation clock speeded
up to drop the value of the system. If, on the contrary, a fast
memory or input/output circuit is to be used in conformity to the
operation clock of the controller, an extremely high rise is caused
in the system price.
[0011] Thus, the system of the prior art for producing the timing
of the data transfer between the CPU and the peripheral circuit
from the operation clock of the CPU or the system cannot realize
the data transfer which can sufficiently exploit the intrinsic
performance of the peripheral circuit such as the memory.
Specifically, we have thought it difficult to desire a basic
development to a high speed if the CPU and the peripheral circuit
are connected by the wait state control function which stresses the
reliable operation by returning the wait signal to the CPU at a
timing of integer times as high as that of the operation clock on
the basis of the characteristics of the peripheral circuit.
[0012] An object of the present invention is to provide a technique
which is enabled to effect a data transfer by sufficiently
exhibiting the intrinsic characteristics owned by a peripheral
circuit such as a memory.
[0013] Another object of the present invention is to provide a
peripheral circuit for producing a timing of the data transfer
according to its own characteristics.
[0014] A further object of the present invention is to provide a
data processor capable of transferring data efficiently with such
peripheral circuit.
[0015] A further object of the present invention is to provide a
data processing system capable of transferring data fast with the
data processor by sufficiently exhibiting the intrinsic
characteristics owned by the peripheral circuit such as the
memory.
[0016] The aforementioned and other objects and novel features of
the present invention will become apparent from the following
description to be made with reference to the accompanying
drawings.
[0017] A representative of the invention to be disclosed herein
will be briefly described in the following.
[0018] Specifically, as represented in FIG. 1, a memory (1) acting
as a peripheral circuit performs its internal operation in
accordance with access requests (200, 201 and 202) from a CPU (2)
exemplifying the data processor, and in synchronism with the output
signal of a self-excited oscillator (102) incorporated therein; and
outputs a response requests (103) to the data processor in
synchronism with that internal operation in response to the access
requests.
[0019] The data processor sends an access request to a desired
peripheral circuit; and transfers data to or from the peripheral
circuit depending on the type of the access request in synchronism
with a response request received from the peripheral circuit.
[0020] The control of the data transfer between the data processor
and the peripheral circuit comprises the steps of sending an access
request to the peripheral circuit from the data processor; allowing
the peripheral circuit to perform its internal operation according
to the access request in synchronism with the output signal of a
self-excited oscillator incorporated therein; sending a response to
the peripheral circuit from the data processor in synchronism with
its internal operation in response to the access request; and
transferring data to or from the data processor depending on the
type of the access request in synchronism with said response
request.
[0021] In order to realize the aforementioned means with the
minimum number of circuits to be added to the construction of the
existing data processor or peripheral circuit, the access requests
can contain the information (200 and 201) for indicating that the
peripheral circuit is selected as the object to be accessed and the
data transfer direction, and the response request can contain the
signal (103) to be changed in synchronism with the internal
operation of the peripheral circuit.
[0022] In order to construct the peripheral circuit having the
aforementioned functions relatively simply, as representatively
shown in FIG. 5, the peripheral circuit includes a cycle timing
generator (1010) for producing an access cycle signal (1013) of the
internal operation in response to the access request from the data
processor and on the basis of an output signal of the self-excited
oscillator (102); an external terminal (AC) for outputting the
access cycle signal as the response request to the outside; and an
internal timing generator (1011) for producing an internal
operation timing signal in synchronism with the access cycle signal
(103).
[0023] In case such peripheral circuit is constructed as a burst
readable memory (capable of reading a continuous data of a
plurality of words), there may be added a counter (or burst
counter) (105) for counting the number of continuous data read
words from the memory cell array on the basis of the change in the
access cycle signal to stop the oscillations of the self-excited
oscillator when the counted result reaches a predetermined count
value, as representatively shown in FIG. 6. At this time, in order
to set the number of continuous data read words programmably, the
counter is equipped with a parameter register (1051) for latching
the predetermined count value presettably from the outside, as
representatively shown in FIG. 12. This parameter register can be
positioned such that in case the counter has storage stages
corresponding to its bit count, the memory stages are used as a
substantial parameter register in a presettable manner.
[0024] In order to transfer data at different transfer rates
quickly and efficiently between the internal unit and the outside,
the data processor having the aforementioned functions is equipped,
as representatively shown in FIG. 8, with a buffer memory (206)
which includes: a asynchronous port (2064) for writing/reading on
the basis of the response request; and a synchronous port (2065)
for writing/reading in synchronism with the internal operation
clock. The synchronous port of the buffer memory is connected as
the internal unit to an arithmetic unit or register, and the
asynchronous port of the buffer memory is connected with an
input/output buffer circuit (205) to be interfaced with the
outside. At this time, in order that the data transferred from the
peripheral circuit to the buffer memory may be quickly used for the
operation of an internal unit (204), the buffer memory may be
equipped with a counter circuit (2066) for counting the number of
continuous read accesses, which are sent to the peripheral circuit
from the access control circuit, in terms of the number of changes
in the response request, so that the resultant detection result may
be fed as the information meaning the complete of the read data
acquisition by said access request (i.e., the output information of
an AND gate 2063R5, as representatively shown in FIG. 9) to the
central processing unit. The buffer memory should not be limited to
a perfect dual port but may be used as an apparent dual port for
operating a unit-port buffer memory in a time sharing manner.
[0025] When the data processor is interfaced with a plurality of
peripheral circuits of different kinds, one input terminal of the
data processor for response request is connected through an OR gate
or a wired OR so that it may be shared among the output terminals
of each peripheral circuit for the response request, as
representatively shown in FIG. 14.
[0026] In order that mutually identical peripheral circuits having
a multi-bit input/output function of 1/2n bits for the number of
bits of the data bus may be interfaced with the data processor, the
data processor may be provided with a plurality of sets of buffer
memories (206U and 206L) each having: an asynchronous port for
writing/reading in response to the response request; and a
synchronous port for writing/reading in synchronism with the
internal operation clock, as representatively shown in FIG. 13.
[0027] According to the means described above, the peripheral
circuit is operated synchronously with the output signal of the
self-excited oscillator intrinsic thereto but asynchronously from
the operation clock signal of the data processor requesting access
to the peripheral circuit. In this relation, the mutual interface
of data is realized by the mutually equivalent access requests and
the response requests for the former. As a result, the time period
for the series of data transfer, which as been limited to integer
times as high as that of the fundamental operation clock of the
data processor of the prior art, is determined according to the
clock cycle of the response request depending upon the intrinsic
self-excited oscillation frequency which is produced according to
the characteristics such as the operation speed of the peripheral
circuit such as the memory. As a result, the data transfer can be
easily realized for the individual characteristic limit time
periods of the peripheral circuit and the data processor. In other
words, it is possible to reduce the spare time period which has
been established for synchronization with the operation clock of
the data processor, as has been troubled in the prior art.
Moreover, the wait state control circuit for the interface between
the data processor and each peripheral circuit can be dispensed
with to simplify the circuit connecting means.
[0028] The data processor equipped on-chip with the buffer memory
to be interfaced with the peripheral circuit can internally absorb
the difference in the data transfer rate between the internal unit
of the data processor and the outside to require no sequential wait
time for reading/writing the data in response to the access
request.
BRIEF DESCRIPTION OF THE DRAWINGS
[0029] FIG. 1 is a system block diagram showing a CPU according to
one embodiment of a data processor of the present invention and a
memory according to one embodiment of a peripheral circuit of the
present invention;
[0030] FIG. 2 is a timing chart showing one example of data
transfer operations in the system of FIG. 1;
[0031] FIG. 3 is a block diagram showing a system making a data
transfer possible through a wait state control unit with reference
to the embodiment of FIG. 1;
[0032] FIG. 4 is a timing chart of the data transferring operations
of FIG. 3;
[0033] FIG. 5 is a block diagram showing one embodiment of the
memory of FIG. 1;
[0034] FIG. 6 is a circuit diagram showing one detailed example of
a cycle timing generator of FIG. 5;
[0035] FIG. 7 is a operation timing chart of one example of the
memory of FIG. 6;
[0036] FIG. 8 is a block diagram of one detailed embodiment of the
CPU of FIG. 1;
[0037] FIG. 9 is a block diagram of one detailed example of a
circuit portion relating to a read buffer in the buffer control
circuit of FIG. 8;
[0038] FIG. 10 is a block diagram of one detailed example of a
circuit portion relating to a write buffer in the buffer control
circuit of FIG. 8;
[0039] FIG. 11 is a block diagram showing one embodiment of a CPU
having a buffer memory of the type sharing a read buffer and a
write buffer;
[0040] FIG. 12 is a block diagram of one embodiment of a memory in
which a burst counter is equipped with a parameter register;
[0041] FIG. 13 is a block diagram of one embodiment when identical
memories having a function to input/output multiple bits of 1/2n
times, for example, as large as the bits of a data but are
interfaced with the CPU;
[0042] FIG. 14 is a block diagram of one embodiment of the case in
which memories having different characteristics/functions are mixed
to construct a system; and
[0043] FIG. 15 is an overall block diagram of one embodiment of a
data processing system.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0044] FIG. 1 shows a CPU according to one embodiment of a data
processor of the present invention and a memory according to one
embodiment of a peripheral circuit of the present invention such
that the CPU and the memory are connected.
[0045] A memory 1, as shown in FIG. 1, is formed on a semiconductor
substrate and includes a memory cell array 100 and an access cycle
control unit 101, as representatively illustrated. In response to
access requests (200, 201 and 202) from a CPU 2, the memory 1
performs read operations or write operations in synchronism with an
oscillation output signal of a self-excited oscillator 102
incorporated therein, and outputs a response request (103)
responding to the access requests to the CPU 2 in synchronism with
its internal operations.
[0046] The CPU 2, as shown in FIG. 1, is formed on a semiconductor
substrate and includes an arithmetic circuit 204; a buffer memory
206 having its one port coupled to the arithmetic circuit 204; an
input/output buffer circuit 205 coupled to the other port of the
buffer memory 206 and an external data bus 211; an access control
circuit 207 for sending access requests to the external memory 1
and other peripheral circuits (not shown); and a central control
unit 208 for controlling the entire operations of the central
processing unit such as an instruction executing sequence control
circuit or an interrupt control circuit. The CPU 2 sends the access
requests (200, 201 and 202) to a desired peripheral circuit such as
the memory 1 and receives the response request (103) of the
peripheral circuit such as the memory 1 to latch the data in the
buffer memory 206 from the outside or to output the data from the
buffer memory 206 to the outside in synchronism with the response
request and according to the kind of said access request. The
memory 1 is operated in synchronism with the output signal of its
own self-excited cillator 102. On the other hand, the CPU 2 is
operated in synchronism with an operation clock 209 of the
system.
[0047] In case the CPU 2 gains access to the memory 1, the start of
this access is transmitted to the memory 1 by the access start
signal 200. This access start signal 200 is deemed as a signal
equivalent to a chip selecting signal by the memory. According to
the present embodiment, the access control circuit 207 has a
function as a chip selecting controller, although not especially
limited thereto. This function can be replaced by a decoder for
decoding several more significant bits of the address signal
outputted from the CPU 2 to the outside, to produce the chip
selecting signal. In any case, reference is made to both an address
assigned to the peripheral circuit to be accessed and an address to
be made in the CPU 2. In this sense, an access request to the
peripheral circuit such as the memory, such as an instruction of
the access start is sent directly or indirectly by a circuit
portion for making the access address, and the access control
circuit is understood to contain such circuit portion.
[0048] The direction of data transfer is instructed by the
read/{overscore (write)} signal 201. The "read" means the data
transfer from the peripheral circuit such as the memory 1 to the
CPU 2, and the "write" means the data transfer from the CPU 2 to
the peripheral circuit such as the memory 1. According to the
present embodiment, the location of the data in the peripheral
circuit, as requested for access, is designated by the address
signal fed to an address bus 210. The number of words for data
transfer is instructed by the instruction signal (i.e.,
single/{overscore (burst)} signal) 202 in a single mode/burst mode.
Without the burst mode or the continuous data transfer mode, the
single/{overscore (burst)} signal 202 can be eliminated.
[0049] Upon detecting the access request in terms of the access
start signal 200, the access cycle control unit 101 produces the
internal access cycle signal on the basis of the output signal of
the self-excited oscillator 102. Inside of the memory 1, the read
or write operations instructed by the read/{overscore (write)}
control signal 201 are performed in synchronism with the access
cycle signal. Outside of the memory 1, moreover, this access cycle
signal is outputted as the access clock signal 103 to the CPU 2.
This access clock signal 103 is a clock signal intrinsic to the
memory 1 and is fed to the CPU 2 as the response request in
response to the access request from the CPU 2.
[0050] FIG. 2 shows the relations between the data output of the
memory 1 in the read operation and the data output timing of the
CPU 2 in the write operation, and the access clock signal 103.
According to FIG. 2, when instructed to perform the read operation,
the memory 1 outputs, the desired data to the data bus 211 at a
timing for ensuring a setup time (Trs)/a hold time (Trh) for the
rising edge of the access clock signal (or the access cycle signal
in the memory). That data is latched by the CPU 2 in the buffer
memory 206 at the rising edge of the access clock signal 103. In
the writing operation, the CPU 2 outputs the data from the buffer
memory 206 to the data bus 211 so as to ensure a setup time
(Tws)/hold time (Twh) for the fall of the access clock signal 103.
The memory 1 latches that data at the falling edge of the access
cycle signal. Incidentally, the write operation can refer to the
rise of the access clock signal 103.
[0051] According to the embodiment of FIG. 1, the access cycle
control unit 101 outputs a cycle complete signal 104 for informing
the CPU 2 of the complete of the continuous data transfer in the
burst mode. In the access control unit 101, the number for words
transferred is counted by a burst counter 105 in terms of the
access cycle signal equivalent to the access clock signal 103 so
that the count-up state is outputted as the cycle complete signal
104. This cycle complete signal 104 may be replaced by an identical
function at the side of the CPU 2. Specifically, a burst counter
for counting the access clock signal 103 may be disposed at the
side of the CPU 2.
[0052] FIG. 3 is a block diagram showing a system making a data
transfer possible through a wait state control unit, in comparison
with the foregoing embodiment, and FIG. 4 shows a data transfer
timing of the system.
[0053] In case a CPU 400 is to transfer data to an external memory
401, as shown in FIG. 3, the start of this data transfer is noticed
in terms of an access start signal 403 to the memory 401 and a wait
state control unit 402. Upon reception of the access start signal
403, a read or write operation is started according to a
read/{overscore (write)} signal 405 by a read/write control circuit
404. In synchronism with this, the wait state control unit 402 also
interprets the access start signal 403 and the read/{overscore
(write)} signal 405 and starts the counting of a wait counter 408
to produce a wait signal 407 for indicating an access complete on
the basis of an operation clock 406 identical to that of the CPU
400. In the reading operation, the memory 401 can output the data
to be read to a data bus 409 after a time period guaranteed by its
maker/seller has elapsed. In the write operation, the memory 401
can latch the data of the data bus 409 outputted by the CPU 400,
after the time period guaranteed by its maker/seller has elapsed.
The complete of the read operation or write operation after lapse
of the time period guaranteed by the maker/seller is usually
transmitted to the CPU 400 in terms of a change of the wait signal
407 to a False in synchronism with the operation clock 406 of the
CPU 400 (or the wait signal is confirmed at the CPU side in
synchronism with the operation clock in case it is an asynchronous
signal). If the wait signal is set in the read operation to the
false (or a low level) at a time t1, as shown in FIG. 4, the CPU
reads the data from the data bus. If the wait signal is set in the
write operation to the false (or the low level) at a time t2, the
CPU confirms that the data to be written has been latched in the
memory, and stops the outputting of the write data.
[0054] It is apparent from the timings of FIG. 4 that the locations
(or timings) of setting the wait signal to the false are usually
different between the read cycle and the write cycle. In the burst
mode, moreover, the wait signal should naturally be cyclically
produced in series of transfer words, in which the interval of
occurrence of a first word is different from those of second and
later words. As a result, the CPU 400 completes a series of read or
write cycles, when it confirms the change in the wait signal 407,
to keep the access control circuit 410 in the standby state till
the next cycle starts. Moreover, a changing time, as indicated at
Tdis in FIG. 4, is required when the read cycle and the write cycle
in the identical operation mode are to be interchanged. This is
because the wait signal is confirmed in synchronism with the clock.
Thus, the data transfer making use of the wait signal has to take a
complicated control and an unnecessary time period.
[0055] The following operational effects can be obtained according
to the embodiment thus far described.
[0056] (1) In the present embodiment shown in FIGS. 1 and 2, the
locations of starting the production of the access cycle by the
peripheral circuit such as the memory and the timings for updating
the changes are usually different between the read cycle and the
write cycle, but the CPU may be dedicated to the
inputting/outputting of data in accordance with the change in the
access clock signal 103 without considering any of those
complicated timings. In short, the data transfer at the complicated
timings can be realized without the wait state control unit which
has been indispensable in the prior art. This realization naturally
covers both the single transfer and the burst transfer.
[0057] (2) Without the wait state control unit, the data transfer
can be achieved by the access clock signal outputted from the
peripheral circuit such as the memory, to reduce the access cycle
time period and improve the bus using efficiency. Specifically, the
peripheral circuit such as the memory is operated in synchronism
with the output signal oscillated by its own self-excited
oscillator 102 and asynchronously of the operation clock signal 209
of the CPU which sends an access request to that peripheral
circuit, so that the interface between their data is realized the
mutually equivalent access requests and the response requests to
the access requests. As a result, a series of data transfer time
periods, which have been limited in the prior art to an integer
times as high as the fundamental operation clock of the data
processor such as the CPU, can be determined according to the clock
cycle of the response request depending upon the intrinsic
self-excited oscillation frequency which is produced according to
the characteristics such as the operating speed of the peripheral
circuit such as the memory. As a result, it is easy to realize the
data transfer for the limit time periods of the individual
characteristics of the peripheral circuit and the CPU. In other
words, it is possible to reduce the unnecessary time period which
has been consumed for synchronization with the operation clock of
the CPU, as has been troubled in the prior art.
[0058] (3) The CPU 2 is equipped on the common chip with the buffer
memory 206 to be interfaced with the peripheral circuit so that the
difference in the data transfer rate between the internal unit 204
of the CPU and the outside can be internally absorbed to prevent
any sequential standby between the time periods for processing the
read data and the write data by the access request.
[0059] (4) The data transfer type according to the foregoing
embodiment can be widely interpreted to give a bus right to the
memory if locally considered. Specifically, it is thought that the
system operates with the operation clock 209 of the CPU 2 at the
start of the data transfer but with the operation clock 103 of the
memory during the data transfer, and the but right seems to be
dynamically transferred. This concept is thought to have an
effective meaning especially when the degree of integration of the
LSI is improved in the future so that the logic function merges
into the memory.
[0060] FIG. 5 is a block diagram showing one embodiment of the
aforementioned memory. The memory 1, as shown in FIG. 5, is formed
as a static random access memory (i.e., SRAM) over one
semiconductor substrate made of single crystalline silicon by the
well-known semiconductor integrated circuit manufacturing
technique, although not especially limited thereto.
[0061] The shown memory 1 includes input terminals ARO to ARm of a
row address signal; input terminals ACO to ACn of a column address
signal; data input/output terminals I/OO to I/Op; a chip select
signal input terminal CS; an output enable signal input terminal
OE; a write enable signal input terminal WE; an access cycle signal
output terminal AC; a burst/single change signal input terminal
B/S; and a supply terminal (not shown). In connection with FIG. 1:
the chip select signal input terminal CS is fed with the access
start signal 200; the output enable signal input terminal OE is fed
with a read signal composing the read/{overscore (write)} signal
201; the write enable signal input terminal WE is fed with a write
signal composing the read/{overscore (write)} signal 201; and the
access cycle signal 103 is outputted from the access cycle signal
output terminal AC.
[0062] The memory cell array 100 includes static memory cells
arranged in a matrix, whose select terminals are connected through
word lines to the outputs of a row address decoder 110. This row
address decoder 110 receives and decodes the internal complementary
address signal, which is converted from the row address signal fed
from the out side and are outputted from a row address buffer 111,
to drive one word corresponding to the row address signal, to a
select level. The bit lines connected to the data input/output
terminals of the memory cells are commonly connected with a common
data line 113 through a column switch circuit 112. The selection of
the bit line to be connected to the common data line 113 is
accomplished by the column switch circuit 112 in response to the
output of a column address decoder 114. This column address decoder
114 receives and decodes the internal complementary address signal,
which is converted from the column address signal fed from the
outside and is outputted from a column address buffer 115, to
select the bit line by the column switch circuit 112. Reference
numeral 116 designates a sense amplifier and an output buffer
circuit for amplifying the data read from the memory cells to the
common data line 113 and outputting the amplified data to the
outside. The sense amplifier and output buffer circuit have their
input connected to the common data line 113 and their output
connected to the data input/output terminals I/OO to I/Op. Numeral
117 designates an input buffer for inputting the write data fed to
the data input/output terminals I/OO to I/O, and the input buffer
117 has its output connected to the common data line 113. Numeral
118 designates either latch circuit or a data control circuit for
equalizing or precharging the common data line.
[0063] The access control unit 101 is equipped with a cycle timing
generator 1010 and an internal timing generator 1011. This internal
timing generator 1011 is connected to the input terminals CS, OE,
WE and B/S to determine the internal operation mode by detecting
the access start, deciding the read/write operation and deciding
the burst mode/single mode, to produce an internal operation timing
signal corresponding to the operation mode in synchronism with the
access cycle signal fed from the cycle timing generator 1010. On
the basis of the instruction of the access start signal fed from
the CS terminal, the cycle timing generator 1010 produces a cycle
timing signal 1013 and the access clock signal 103 in synchronism
with the signal fed from the internal timing generator 1011 on the
basis of the output signal from the self-excited oscillator 102.
Incidentally, a delay circuit 1014 is used to adjust the phase of
the self-excited oscillating output, and a delay circuit 1015 is
used to match the phases of the access clock signal 103 outputted
from the outside and the cycle timing signal 1013.
[0064] FIG. 6 shows one detailed example of the cycle timing
generator 1010. The self-excited oscillator 102 has a feedback loop
composed of a two-input AND gate 1020 and an inverter amplifier
1021 for feeding back the output of the AND gate 1020 to one of the
inputs of the same AND gate 1020, although not especially limited
thereto, and a trigger for controlling the oscillation and its stop
is connected with the other input of the AND gate 1020. The trigger
circuit is equipped with an AND gate 1024, to which is inputted the
output of a selector 1022 to be set to a high level at an initial
state and to which is fed back the output of an OR gate 1023. This
OR gate 1023 receives both the output of the AND gate 1024 and a
trigger signal 1025 such as a one-shot pulse to be fed in
synchronism with the read or write operation from the internal
timing generator 1011, and feeds its output to the AND gate 1020.
Incidentally, reference numerals 1026 to 1028 designate waveform
shaping elements (or delay elements). This self-excited oscillator
102 outputs a low level in the initial state. When the trigger
signal 1025 is changed in this state by the one-shot pulse,
oscillations are established in the feedback loop which is composed
of the AND gate 1020 and the inverter amplifier 1021. This
oscillating state is continued till the output of the selector 1022
is changed into pulses at the low level to set the output of the OR
gate 1023 to the low level.
[0065] In the construction of FIG. 6, the burst counter 105 and the
selector 1022 are used to stop the oscillator. The selector 1022 is
fed with either the B/S signal or an equivalent internal signal to
select the output of the waveform shaping element 1027 in the
single mode. Thus, in the single mode, the self-excited oscillator
102 stops the oscillating operation by changing the access clock
signal 103 and the cycle timing signal 1013 by one cycle. In the
burst mode, the output of the burst counter 105 is selected. This
burst counter 105 counts the number of data words read continuously
from the memory cell array on the basis of the change in the output
pulses of the waveform shaping element 1027, and outputs a one-
shot pulse which will change from the high level to the low level
such that the counted result reaches a predetermined value (i.e., a
target number of words for burst transfer). As a result, the
oscillating operation of the self-excited oscillator 102 is stopped
when the access cycle corresponding to the target number of words
in the burst mode is established.
[0066] FIG. 7 shows one example of the operation timing chart of
the memory of FIG. 6. As shown, in the read cycle, the access cycle
signal output terminal AC is changed in synchronism with the timing
at which the read data is outputted. In the write cycle, on the
other hand, the write data is fed from the CPU in synchronism with
the timing at which the access cycle signal output terminal AC is
changed.
[0067] FIG. 8 is a block diagram showing one detailed embodiment of
the CPU 2. The CPU 2, as shown, is formed on a semiconductor
substrate made of single-crystal silicon, by the well-known
semiconductor integrated circuit manufacturing technique, although
not limited thereto. The same circuit blocks as those of FIG. 1 are
designated at the same reference characters, and their detailed
description will be omitted. Here, the buffer memory 206 will be
described in detail.
[0068] The buffer memory 206 is composed of a read buffer 2061, a
write buffer 2062 and a buffer control circuit 2063 of the FIFO
(i.e., First-IN First-Out) type. The read buffer 2061 is dedicated
to the data transfer in the read direction by the CPU, and the
write buffer is dedicated to the data transfer in the write
direction by the CPU. Both buffers 2061 and 2062 are equipped with
an asymmetric port 2064 to be controlled on the basis of a request
request fed from the memory 1 in response to the access clock
signal 103; and a synchronous port 2065 to be controlled in
synchronism with the internal operation clock 209. The buffer
control circuit 2063 is equipped with an asynchronous control unit
2063A for controlling the asynchronous port 2064; and a synchronous
control unit 2063B for controlling the synchronous port 2065. The
asynchronous port 2064 is connected to the input/output buffer
circuit 205, and the synchronous port 2065 can be interfaced with
the register group, the cache memory and so on contained in the
arithmetic circuit 204.
[0069] The asynchronous control unit 2063A feeds the write buffer
2062 with the asynchronous read signal for instructing the read
operation and the asynchronous read address (or pointer) of that
time, in synchronism with the aforementioned change in the access
clock signal 103. In synchronism with this change in the access
clock signal 103, moreover, the read buffer 2061 is fed with the
asynchronous write signal for instructing the write signal and the
asynchronous write address (or pointer) of that time. A choice
between the read buffer 2061 and the write buffer 2062 to be
accessed in synchronism with the change in the access clock signal
103 depends on the information from the central control unit 208
indicating whether the CPU 2 requests reading or writing
corresponding to the access clock signal 103.
[0070] The synchronous control unit 2063B is operated as a part of
the instruction execution control in the central control unit 208.
When the memory read operation is required according to the
execution of a data transfer instruction such as a load
instruction, a store instruction or a move instruction, the read
buffer 2061 is fed with the synchronous read signal for instructing
the read operation and the synchronous read address (or pointer) of
that time, in synchronism with the operation clock 209. When the
memory write operation is required according to the execution of
the data transfer instruction, on the other hand, the write buffer
2062 is fed with the synchronous write signal instructing the write
operation and the synchronous write address (or pointer) of that
time, in synchronism with the operation clock 209. A choice between
the read buffer 2061 and the write buffer 2062 to be accessed
depends on the instruction decoding signal outputted from the
central control unit 208 in accordance with the instruction
execution.
[0071] The memory 1 of the example of FIG. 8 does not have the
function of producing the cycle complete signal 104. A similar
function is executed by a burst counter 2066, in the asynchronous
control unit 2063A to feed the burst transfer cycle complete to the
access control circuit 207. In the CPU 2 of the present embodiment,
the count-up signal of the burst counter 2066 is also used for
informing the central control unit 208 of the write complete in the
read buffer 2061 and the read complete from the write buffer 2062.
This operation will be described with reference to FIG. 9.
[0072] FIG. 9 is a detailed block diagram of one example of a
circuit portion relating to the read buffer 2061 in the buffer
control circuit 2063. The synchronous read address of the read
buffer 2061 is produced by an upcounter 2063R1, and the
asynchronous write address of the read buffer 2061 is produced by
an upcounter 2063R2. The upcount operation of the upcounter 2063R2
is performed in synchronism with the timing at which the access
clock signal 103 is changed to the high level and at which the read
buffer write signal from the central control unit 208 is activated.
The up count operation of the upcounter 2063R1 is performed in
synchronism with the operation clock 209 when the read buffer read
signal from the central control unit 208 is activated. Both the
upcounters 2063R1 and 2063R2 are cleared to 0 by the high-level
output of the AND gate 2063R3. It is the clearing timing when the
outputs of both the upcounters 2063R1 and 2063R2 is detected by a
coincidence detector 2063R6 in case the output value of the
upcounter 2063R1 is not 0. It is detected by a 0 detector 2063R4
that the output value of the upcounter 2063R1 is 0. In case the
output value of the upcounter 2063R1 is 0, the 0 detection result
by the 0 detector 2063R4 means the vacancy of the read buffer 2061
and is fed to the central control unit 208. If this state is
detected, the central control unit 208 can confirm that the read
data from the memory 1 has been wholly transferred to the
arithmetic circuit 204. The bust counter 2066, as shown in FIG. 8,
detects whether or not the given number of continuous data words
transferred is reached. When this number of words is reached, the
output of the burst counter 2066 is changed to the high level for a
predetermined time period. In the read operation of the memory 1,
the change of the output from the burst counter 2066 is fed as a
signal meaning the read complete to an AND gate 2063R5. This AND
gate 2063R5 detects the read complete of the read buffer 2061, when
it receives the signal meaning the read complete if it is detected
by the 0 detector 2063R4 that the output of the upcounter 2063R1 is
not 0, and transfers the detected read complete to the central
control unit 208. When the read complete of the read buffer 2061 is
detected, the central control unit 208 can confirm that the read
data from the memory 1 is wholly latched in the read buffer 2061 so
that the central control unit 208 can read the read data from the
read buffer 2061 to start the internal arithmetic operation
instantly.
[0073] FIG. 10 is a detailed block diagram of one example of a
circuit portion relating to the write buffer 2062 in the buffer
control circuit 2063. The synchronous write address of the write
buffer 2062 is produced by an upcounter 2063W2, and the
asynchronous write address of the write buffer 2062 is produced by
an upcounter 2063W1. The upcount operation of the upcounter 2063W1
is performed in synchronism with the timing at which the access
clock signal 103 is changed to the high level and at which the
write buffer write signal from the central control unit 208 is
activated. The upcount operation of the upcounter 2063W2 is
performed in synchronism with the operation clock 209 when the
write buffer write signal from the central control unit 208 is
activated. Both the upcounters 2063W1 and 2063W2 are cleared to 0
by the high-level output of the AND gate 2063W3. It is the clearing
timing when the outputs of both the upcounters 2063W1 and 2063W2 is
detected by a coincidence detector 2063W6 in case the output value
of the upcounter 2063W1 is not 0. It is detected by a 0 detector
2063W4 that the output value of the upcounter 2063W1 is 0. In case
the output value of the upcounter 2063W1 is 0, the 0 detection
result by the 0 detector 2063W4 means the vacancy of the write
buffer 2062 so that the central control unit 208 recognizes the
vacant state of the write buffer 2062. In the write operation of
the memory 1, the change of the burst counter 2066 to the high
level is fed as a signal meaning the write complete to an AND gate
2063W5. This AND gate 2063W5 detects the write complete of the
write buffer 2062, when it receives the signal meaning the write
complete if it is detected by the 0 detector 2063W4 that the output
of the upcounter 2063W1 is not 0, and transfers the detected write
complete to the central control unit 208. When the write complete
of the write buffer 2062 is detected, the central control unit 208
can confirm that the write data to the memory 1 in response to the
response request from the memory is wholly outputted from the write
buffer 2062.
[0074] FIG. 11 shows a buffer memory different from the buffer
memory 206 shown in FIG. 8. The buffer memory 206, as shown, is
equipped with a read/{overscore (write)} buffer 2067 which is
shared between the read buffer 2061 and the write buffer 2062. The
buffer control circuit 2063 is equipped with a read/{overscore
(write)} buffer enable flag 2068 which is set with the information
concerning which of the read buffer or the write buffer the
read/{overscore (write)} buffer 2067 is to be operated as. The
operation of the read/{overscore (write)} enable flag 2068 is
controlled according to the instruction coming from the central
control unit 208. The remaining points are similar to those of FIG.
8, and their detail description will be omitted by designating the
same circuit blocks at the same reference characters. This
construction contributes to a reduction of the chip area.
[0075] FIG. 12 shows an essential portion of an embodiment having a
control parameter register in place of the memory of FIG. 5.
Specifically, there is provided a parameter register 1051 for
latching in a preset table manner the target number of the
continuous data words to be counted by the burst counter 105 of
FIG. 6. This parameter register 1051 is controlled by the central
control unit 208 of the CPU 2 to transfer a desired parameter
(i.e., the information for specifying the number of words for a
burst transfer) in a programmable manner. The remaining
constructions are similar to those of FIGS. 5 and 6, and the
detailed description will be omitted by designating the same
circuit blocks at the same reference characters. This raises the
degree of freedom of data transfer and the flexibility of the
control. Incidentally, the parameter register 1051 can be
exemplified by constructing memory stages in a presettable manner
in case the memory stages correspond to the number of counted
bits.
[0076] FIG. 13 shows an embodiment which has a multi-bit
input/output function of 1/2n bits, for example, for the bit number
of the data bus to establish an interface between the same memory
and the CPU 2. In this embodiment, the CPU 2 is equipped with a
plurality of sets of buffer memories 206 and input/output buffer
circuits 205. In case the data bus 211 has 32 bits whereas the
memory 1 has 16 parallel input/output bits, the more significant
data bus 211U of 16 bits is coupled through an input/output buffer
circuit 205U to one memory 1U, whereas the less significant data
bus 211L is coupled through an input/output buffer 205L to the
other memory IL. The access start signal 200, the read/{overscore
(write)} signal 201, the single/{overscore (burst)} signal 202 and
the address bus 210 are commonly connected with the memories 1U and
1L. An access clock signal 103U is connected with a buffer memory
206U whereas an access clock signal 103L is connected with a buffer
memory 206L. Cycle complete signals 104U and 104L are respectively
outputted from the memories 1U and 1L and fed to a cycle complete
control circuit 2069 to transmit the memory access completes of the
two to the access control circuit 207.
[0077] The number of parallel input/output bits of the actual
memory may be .times.4, .times.8, .times.9, .times.16 or .times.18,
and the number of parallel data input/output bits of the CPU may be
.times.16, .times.32, .times.36, .times.64 or .times.72. In order
to interface the memory and the CPU in a corresponding manner, it
is necessary and important to provide every bits with the buffer
memory, as shown in FIG. 13.
[0078] FIG. 14 shows an embodiment of the case in which memories
having different characteristics/functions are mixed to construct
the system. In this case, the data transfer can be achieved
basically in conformity to an access clock if fine terminal
functions or connection conditions are ignored. Thus, an access
clock signal 103-1 of a memory 1-1 and an access clock signal 103-2
of a memory 1-2 are coupled through an OR gate 300 to the buffer
control circuit 2063 outside the CPU 2. Likewise, an access
complete signal 104-1 of the memory 1-1 and an access complete
signal 104-2 of the memory 1-2 are coupled through an OR gate 301
to the access control circuit 207 outside the CPU 2. The remaining
access start signal 200, read/{overscore (write)} signal 201,
single/{overscore (burst)} signal 202, address bus 210 and data bus
211 are commonly connected with the memories 1-1 and 1-2. As a
result, the peripheral circuits such as the memories having the
different characteristics/functions can be mixed to construct the
system.
[0079] FIG. 15 shows one overall embodiment of a data processing
system using the CPU 2 and the memory 1, as described in the
foregoing embodiments. As the peripheral circuits enabled to
perform the data transfer by a protocol similar to that of the
memory (e.g., RAM) 1 of the foregoing embodiments, as shown in FIG.
15, there are provided a memory (e.g., ROM) 3; a file control unit
4 to be interfaced with a hard disk drive 41 and a floppy disk
drive 42; a display control unit 5 for controlling the drawing of a
frame buffer 51 and the display of the drawn data on a monitor 52;
a parallel/serial port 6 to be interfaced with a printer 61 and a
keyboard 62; and a communication unit 10 for transferring data
through an antenna and a cable. These peripheral circuits have
their own self-excited oscillators 102 according to their operating
characteristics so that the data transfer is realized by returning
response requests in response to the access requests from the CPU 2
as in the memory. In FIG. 15, reference numeral 9 designates a
system monitor unit for monitoring a system malfunction and the
status of the supply voltage by means of a watchdog timer. A fast
data transfer unit 8 is a circuit such as a direct memory access
controller, and the bus arbitration with the CPU 2 is accomplished
by the bus arbitrator 7. The fast data transfer unit 8 performs the
data transfer control like that of the CPU 2. Numeral 21 designates
an external cache memory intrinsic to the CPU 2 to act as a
secondary cache memory for a cache memory 22 in the CPU 2. The data
processing system of FIG. 15 is constructed over a packaging board
which is formed with address/data buses 11 and a control bus
12.
[0080] Since the data processing system of FIG. 15 requires no wait
state control for the memories or input/output circuits, neither a
memory controller nor an input/output controller is formed on the
substrate.
[0081] Although our invention has been specifically described in
connection with its embodiments, it should not be limited thereto
but can naturally be modified in various manners without departing
from the gist thereof.
[0082] For example, the foregoing embodiments have been described
in case the peripheral circuit is exemplified by the memory such as
the RAM. However, the peripheral circuit should not be limited
thereto but can be applied not only to the peripheral circuit shown
in FIG. 15 but also to a variety of other peripheral circuits.
Moreover, the application of such peripheral circuit can cover not
only a CPU or a direct memory access controller but also a variety
of data processors such as a microprocessor, a microcomputer, a
single-chip microcomputer or a digital-signal processor.
[0083] On the other hand, the buffer memory should not be limited
to the perfect dual port buffer as in the foregoing embodiment but
can be used as if it acts as a dual port for operating a uni-port
buffer memory in a time sharing manner. From the standpoint of the
chip area of the data processor, on the other hand, the depth (or
storage capacity) of the buffer memory is important, but the much
inferior function will have the less contribution to the
improvement in the bus rate. Thus, this belongs to the design item
which is determined by considering the trade-off from the cost
performance. Incidentally, the depth of the buffer memory is
thought to aid in simplification of the buffer control circuit if
it is restricted to the number of words to be handled by a data
transfer of one time (i.e., the maximum number of words of the
burst transfer).
[0084] The effects to be obtained by the representative of the
invention disclosed herein will be briefly described in the
following.
[0085] Specifically, the peripheral circuit is operated in
synchronism with the output signal from its own self-excited
oscillator and asynchronously of the operation clock signal of the
data processor which sends an access request to that peripheral
circuit. In this relation, the interface between their data is is
realized the mutually equivalent access requests and the response
requests in response to the access request. As a result, a series
of data transfer time periods, which have been limited in the prior
art to an integer times as high as the fundamental operation clock
of the data processor, can be determined according to the clock
cycle of the response request depending upon the self-excited
oscillation frequency which is produced according to the
characteristics such as the operating speed of the peripheral
circuit such as the memory.
[0086] From the above discussion, it is easy to realize the data
transfer for the limit time periods of the individual
characteristics of the peripheral circuit and the data processor.
In other words, it is possible to reduce the unnecessary time
period spent for synchronization with the operation clock of the
data processor, as has been troubled in the prior art.
[0087] From the above discussion, moreover, the wait state control
circuit for the interface between the data processor and each of
the peripheral circuits can be dispensed with to simplify the
circuit connecting means.
[0088] The data processor equipped on the common chip with the
buffer memory to be interfaced with the peripheral circuit can
internally absorb the difference in the data transfer rate between
the internal unit of the data processor and the outside to shorten
the time period for the sequential standby in the processing of the
read data and the write data by the access request.
[0089] Thus, the data processing system can be freely constructed
by interfacing the data processor with a plurality of peripheral
circuits of different kinds and by interfacing mutually identical
peripheral circuits having a multi-bit input/output functions of
1/2n bits for the bit number of the data bus with the data
processor.
* * * * *