Miscellaneous active electrical nonlinear devices – circuits – and – Signal converting – shaping – or generating – Synchronizing
Reexamination Certificate
2000-03-08
2001-10-23
Tran, Toan (Department: 2816)
Miscellaneous active electrical nonlinear devices, circuits, and
Signal converting, shaping, or generating
Synchronizing
C327S212000
Reexamination Certificate
active
06307410
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a semiconductor integrated circuit device and, particularly, to a semiconductor integrated circuit device having a main memory portion and a sub-memory portion formed in a semiconductor substrate and a data transfer circuit provided between the main memory portion and the sub-memory portion.
This application is based on Japanese Patent Application No. 11-62393, the contents of which are incorporated herein by reference.
2. Description of the Related Art
In general, a relatively low speed, inexpensive semiconductor device having the large memory capacity, such as general purpose DRAM, is used as the main memory in a computer system.
In recent computer systems, the operating speed of a DRAM constituting the main memory is increased along with the increase of the operating speed of the system, particularly, of the MPU. However, the operating speed of the DRAM is still insufficient and, in order to solve this problem, it is usual to provide a sub-memory between the MPU and the main memory. Such a sub-memory is generally called as a cache memory and is constructed with a high speed SRAM or an ECLRAM.
The cache memory is generally provided to the MPU externally or internally. Recently, a semiconductor device in which the DRAM constituting the main memory and the cache memory are mounted on the same semiconductor substrate has gained attention. Japanese Unexamined Patent Application, First publication Nos. Sho 57-20983, Sho 607690, Sho 62-38590 and Hei 1-146187 disclose examples of such semiconductor memory. Such semiconductor memory is sometimes called cache DRAM or CDRAM since it includes the DRAM and the cache memory. Data are bi-directionally transferred between the SRAM, which functions as the cache memory, and the DRAM which is the main memory.
These conventional arts have problems such as delays in data transfer operation in a case of cache mis-hit, and techniques which solve such problem has been proposed. Examples of the proposed techniques are disclosed in Japanese Unexamined Patent Applications, First Publication Nos. Hei 4-252486, Hei 4-318389 and Hei 5-2872. In the techniques, a latch or register function is provided in a bi-directional data transfer circuit between a DRAM portion and an SRAM portion, so that data transfer from the SRAM portion to the DRAM portion and data transfer from the DRAM portion to the SRAM portion can be done simultaneously, and the speed of data transfer (copy back) at the cache his-hit can be increased.
The above mentioned technique, however, has a limitation in the number of the circuits and in the number of the transfer bus lines because the bi-directional transfer gate circuit occupies a significant part of the area. Therefore, the number of bits which can be transferred between the DRAM array and the SRAM array at a time is limited to 16 bits. In general, as the number of bits transferred at a time decreases, the cache hit rate decreases.
Further, there is the recent problem of degradation of the cache hit rate when there are access requests from a plurality of processing devices as shown in FIG.
70
. When the CDRAM or the EDRAM is used for the main memory as shown in FIG.
70
and there are access requests from a plurality of processing devices (memory masters), the cache hit rate is lowered and the speeding up of the whole system operation is restricted since the number of address requests for different sets (rows) may be increased. As the number of the systems having processing devices is increased, the memory portions must respond not only to one type of an access request but also various types of access requests.
In addition to the above-mentioned problem, this type of semiconductor memory receives an address signal synchronously with an external clock, and has the problem in that it takes time to generate an internal address signal.
FIG. 71
shows the construction of the circuit for generating the internal address signal within the conventional semiconductor memory. As shown in
FIG. 71
, a clock signal CLK, an external address signal Ai, various control signals CSB, RASB, CASB, and WEB are input through receiver circuits
6001
to
6006
. The receiver circuits
6001
to
6006
convert these external signals into signals suitable for handling within the device. The clock signal CLK, which is input through the receiver circuit
6001
, is input into an internal clock signal generating circuit
6010
, which then generates an internal clock signal ICLK having a predetermined duty.
The address signal Ai, which is input through the receiver circuit
6002
, is latched in an address latch circuit
6011
, which then generates an internal address signal IAi. The various control signals CSB, RASB, CASB, and WEB are input into a command latch circuit
6012
, which then generates a read/write command signal and an active command signal. A read/write signal generating circuit
6013
receives the read/write command signal, and generates a read/write signal and a column address latch signal. An active signal generating circuit
6014
receives the active command signal, and generates an active signal and a row address latch signal.
A column address latch circuit
6020
uses the column address latch signal from the read/write signal generating circuit as the trigger, latches the internal address signal LAi from the address latch circuit, and generates a column address signal Yi. This column address signal Yi is input into a counter circuit
6021
, which then generates a counter output address signal. The counter output address signal is input into a column address latch circuit
6020
and is used, for example, as a column address signal in a burst mode. A row address latch circuit
6022
uses the row address latch signal from the active signal generating circuit
6014
as the trigger, latches the internal address signal IAi from the address latch circuit, and generates a row address signal Xi.
FIG. 72
shows the construction of the address latch circuit
6011
.
As shown in
FIG. 72
, the address latch circuit
6011
comprises: a master latch circuit
6011
A for allowing the address signal CAi, which is externally input, to pass when the internal clock signal is at a L (low) level, and latching the address signal CAi when the internal clock signal is at a H (high) level; and a slave latch circuit
6011
B for allowing the signal from the master to pass when the internal clock signal is at the H level, and latching the signal when the internal clock signal is at the L level. That is, according to this construction, the address signal is latched by the master at the rising edge of the clock signal ICLK, and is latched by the slave at the falling edge of the clock signal ICLK, to thereby output the internal address signal IAi.
FIG. 73
shows the construction of the command latch circuit
6012
.
As shown in
FIG. 73
, the various control signals CCS (control signals corresponding to CSB), CRAS (control signal corresponding to RASB), CCAS (control signal corresponding to CASB), and CWE (control signal corresponding to WEB), which are output from the above-mentioned receiver circuits
6003
to
6006
, are latched in master latch circuits
6012
A at the rising edge of the internal clock signal ICLK, and their logical product is calculated by a gate circuit
6012
B. The signal, obtained from the logical product, is latched in a slave latch circuits
6012
C at the falling edge of the internal clock signal ICLK, and the read/write command signal and the active command signal are generated.
FIG. 74
shows the construction of the read/write signal generating circuit.
The internal clock signal ICLK is delayed by a predetermined time by an inverter chain
6013
A, and is input into one of the NAND circuits
6013
B and
6013
C. The above-described read/write command signal and the burst signal for activating the burst mode are input into the other NAND circuits
6013
B or
6013
C. Signals output from the NAND circuits
6013
B and
6013
C are input into the NAND circuit
6013
D. A s
Cox Cassandra
NEC Corporation
Sughrue Mion Zinn Macpeak & Seas, PLLC
Tran Toan
LandOfFree
Semiconductor integrated circuit device does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Semiconductor integrated circuit device, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Semiconductor integrated circuit device will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2590460