US20150085586A1 - Memory device and method of operation of such a memory device - Google Patents
Memory device and method of operation of such a memory device Download PDFInfo
- Publication number
- US20150085586A1 US20150085586A1 US14/037,413 US201314037413A US2015085586A1 US 20150085586 A1 US20150085586 A1 US 20150085586A1 US 201314037413 A US201314037413 A US 201314037413A US 2015085586 A1 US2015085586 A1 US 2015085586A1
- Authority
- US
- United States
- Prior art keywords
- write
- circuitry
- voltage level
- column
- word line
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims description 16
- 230000002093 peripheral effect Effects 0.000 claims abstract description 72
- 238000004590 computer program Methods 0.000 claims description 10
- 238000007599 discharging Methods 0.000 claims description 4
- 230000008878 coupling Effects 0.000 claims description 3
- 238000010168 coupling process Methods 0.000 claims description 3
- 238000005859 coupling reaction Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 2
- 230000001419 dependent effect Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 238000013459 approach Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000007246 mechanism Effects 0.000 description 5
- 230000006872 improvement Effects 0.000 description 4
- 230000007704 transition Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000014759 maintenance of location Effects 0.000 description 3
- 230000000717 retained effect Effects 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001808 coupling effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/413—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C7/00—Arrangements for writing information into, or reading information out from, a digital store
- G11C7/12—Bit line control circuits, e.g. drivers, boosters, pull-up circuits, pull-down circuits, precharging circuits, equalising circuits, for bit lines
-
- G06F17/505—
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/413—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction
- G11C11/417—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction for memory cells of the field-effect type
- G11C11/418—Address circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C11/00—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor
- G11C11/21—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements
- G11C11/34—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices
- G11C11/40—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors
- G11C11/41—Digital stores characterised by the use of particular electric or magnetic storage elements; Storage elements therefor using electric elements using semiconductor devices using transistors forming static cells with positive feedback, i.e. cells not needing refreshing or charge regeneration, e.g. bistable multivibrator or Schmitt trigger
- G11C11/413—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction
- G11C11/417—Auxiliary circuits, e.g. for addressing, decoding, driving, writing, sensing, timing or power reduction for memory cells of the field-effect type
- G11C11/419—Read-write [R-W] circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C16/00—Erasable programmable read-only memories
- G11C16/02—Erasable programmable read-only memories electrically programmable
- G11C16/06—Auxiliary circuits, e.g. for writing into memory
- G11C16/08—Address circuits; Decoders; Word-line control circuits
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C7/00—Arrangements for writing information into, or reading information out from, a digital store
- G11C7/10—Input/output [I/O] data interface arrangements, e.g. I/O data control circuits, I/O data buffers
- G11C7/1078—Data input circuits, e.g. write amplifiers, data input buffers, data input registers, data input level conversion circuits
- G11C7/1096—Write circuits, e.g. I/O line write drivers
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11C—STATIC STORES
- G11C8/00—Arrangements for selecting an address in a digital store
- G11C8/10—Decoders
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03K—PULSE TECHNIQUE
- H03K17/00—Electronic switching or gating, i.e. not by contact-making and –breaking
- H03K17/002—Switching arrangements with several input- or output terminals
- H03K17/005—Switching arrangements with several input- or output terminals with several inputs only
-
- G06F2217/78—
Definitions
- the present invention relates to a memory device and to a method of operation of such a memory device, and in particular to a mechanism for operating such a memory device when the access circuitry used to access the array of memory cells is operated from a peripheral voltage domain with a supply voltage less than the supply voltage in a core voltage domain used by the array of memory cells.
- embedded SRAM bit cells may use a higher voltage supply (providing a core voltage level) to guarantee state retention, whilst the rest of the system, including the access logic circuitry employed to access those bit cells within the memory device, may use a lower voltage supply (providing a peripheral voltage level) to reduce power consumption.
- a level shifting mechanism is needed to pass signals between these different voltage domains.
- FIG. 1 is a block diagram schematically illustrating a known memory device 10 where the access logic circuitry 20 of the memory device is provided in a first voltage domain operating with a first voltage supply providing a peripheral voltage level VDDP and a ground voltage level (not shown), whilst the array of bit cells 30 is provided in a second voltage domain operating with a second voltage supply providing a core voltage level VDDC and the ground voltage level.
- the core voltage level VDDC is higher than the peripheral voltage level VDDP.
- the access logic circuitry 20 will receive a number of control signals over path 55 , these control signals identifying write transactions, read transactions, the addresses of those transactions, etc. It will also receive over path 60 write data for write transactions to be performed within the array of bit cells 30 , and will output over path 75 the read data resulting from read transactions performed within the array of bit cells 30 .
- the access logic circuitry 20 will include a number of components, such as address latches, various stages of word line decoding circuitry to decode the address in order to generate appropriate word line enable signals to activate addressed rows within the array of bit cells 30 , write data path logic circuits used to control the voltages on the bit lines within the array of bit cells during a write operation, and various read data path logic circuits for processing the data read out of the array of bit cells in response to a read operation. Accordingly, for write operations, a number of signals will be generated for issuing to the array of bit cells 30 (these signals being schematically illustrated by the arrow 65 in FIG. 1 ).
- these signals are subjected to a level up shifting function 40 in order to convert the voltage of those signals from the lower voltage domain to the higher voltage domain.
- any data read out from the bit cells 30 over path 70 is subjected to a level down shifting function 50 in order to convert the voltage levels from the higher (core) voltage domain to the lower (peripheral) voltage domain, before those signals are then subsequently processed by the access logic circuitry 20 .
- the level up shifting mechanism 40 is generally more problematic to implement than the level down shifting mechanism 50 (in fact in many instances no specific level down shifting circuitry may be required), since when performing level up shifting there is the potential for establishing various DC paths that can result in significant power consumption, and which may potentially create short circuit current paths.
- the voltage difference between the core voltage domain and the peripheral voltage domain is relatively small (for example up to 250 mV) it is not always necessary to level shift all of the control signals routed from the access logic circuitry to the bit cells, and read and write operations performed in respect of the bit cells will still operate correctly.
- the voltage difference between the core voltage domain and the peripheral voltage domain is getting larger, and may for example be in the range of 400 mV.
- level shifting of the control signals is required in order to ensure correct operation of the bit cells within the array when performing read and write operations.
- the write margin can be insufficient unless such level shifting has been performed.
- the present invention provides a memory device comprising: an array of memory cells connected to a core voltage level and arranged as a plurality of rows and columns, each row being connected to an associated word line, and each column being connected to an associated at least one bit line, the columns being arranged into a plurality of column groups, each column group containing a plurality of columns; access circuitry configured to perform a write operation in order to write data into a plurality of addressed memory cells, said plurality of addressed memory cells comprising an addressed memory cell within each of a plurality of the column groups, the access circuitry being configured to receive both the core voltage level and a peripheral voltage level less than said core voltage level; the access circuitry comprising: word line driver circuitry configured to assert a word line signal at the core voltage level on the word line associated with a row of the array containing the plurality of addressed memory cells; precharge circuitry configured to precharge, to said peripheral voltage level, the at least one bit line associated with at least each column containing an addressed memory cell, prior to said
- the access circuitry is configured to receive both the core voltage level and the peripheral voltage level.
- Word line driver circuitry within the access circuitry is arranged to assert a word line signal at the core voltage level in order to activate a row containing the addressed memory cells for a write operation, and in addition write multiplexing driver circuitry asserts the required mux control signal(s) to write multiplexing circuitry at the core voltage level in order to cause the relevant components within the write multiplexing circuitry to couple the required columns of the array to the write driver circuitry.
- the precharge circuitry precharges the bit lines to the peripheral voltage level prior to the write operation being performed.
- the word line driver circuitry and the write multiplexing driver circuitry are the only components within the access circuitry that are configured to operate from the core voltage level, and all of the remaining components operate from the peripheral voltage level. This optimises the power consumption benefits achievable, whilst ensuring sufficient write margin is maintained.
- the write multiplexing driver circuitry includes level shifting circuitry used to convert at least one signal from the peripheral voltage level to the core voltage level in order to cause the asserted mux control signal to be at the core voltage level.
- level shifting circuitry is provided within the write multiplexing driver circuitry, in addition to the components required to perform the write multiplexing driver function.
- the at least one signal whose voltage level is converted by the level shifting circuitry can vary dependent on embodiment.
- the level shifting circuitry operates on at least one input signal to the write multiplexing driver circuitry in order to generate a corresponding at least one level shifted input signal, and the write multiplexing driver circuitry is configured to then generate the asserted mux control signal in dependence on said at least one level shifted input signal.
- all of the input signals to the write multiplexing driver circuitry are subjected to this level shifting operation, and accordingly the functional components performing the write multiplexing driver operation within the write multiplexing driver circuitry operate entirely on input signals within the core voltage domain.
- level shifting circuits Whilst in one embodiment the above-described dedicated level shifting circuits are used, such level shifting circuits occupy a relatively large area, and consume additional power to perform their level shifting operations. In an alternative embodiment, the requirement for such separate level shifting circuits can be avoided by instead applying a P/N skew to the transistors forming at least one of the components within the write multiplexing driver circuitry.
- the write multiplexing driver circuitry comprises at least one circuit component formed of both NMOS transistors and PMOS transistors, and a drive strength of one of the NMOS transistors and the PMOS transistors is skewed so as to enable the write multiplexing driver circuitry to level shift at least one signal from the peripheral voltage level to the core voltage level in order to cause the asserted mux control signal to be at the core voltage level.
- the write multiplexing driver circuitry includes a NAND circuit component and a P/N skew is incorporated within the NAND circuit component to increase the drive strength of the NMOS transistors relative to the PMOS transistors.
- word line driver circuitry used to assert the word line signal at the core voltage level.
- word line driver circuitry may incorporate dedicated level shifting circuits, or can provide P/N skew within the transistors of one or more components of the word line driver circuitry in order to achieve the required shift in the voltage level from the peripheral voltage domain to the core voltage domain.
- the write multiplexing circuitry can be constructed in a variety of ways. However, in one embodiment the write multiplexing circuitry comprises, for each column group, switch circuitry configured to couple to the write driver circuitry the at least one bit line associated with one column in that column group. Hence, one column in each column group is coupled to the write driver circuitry, the column selected being the column that contains the addressed memory cell.
- the switch circuitry can take a variety of forms, but in one embodiment comprises a switch element for each column within the column group, and the mux control signal generated by the write multiplexing driver circuitry comprises a separate mux control signal for each switch element, with one of the separate mux control signals being asserted at the core voltage level to turn on the associated switch element.
- the individual switch elements can take a variety of forms. In one embodiment, they are formed by an NMOS transistor circuit whose gate is driven by the appropriate mux control signal. In an alternative embodiment, a transmission gate structure can be used including an NMOS transistor and a PMOS transistor placed back to back, with the input signal provided to the gate of the NMOS transistor being inverted prior to provision to the gate of the PMOS transistor.
- operation of the precharge circuitry and the write driver circuitry is controlled by associated control signals switchable between a ground voltage level and said peripheral voltage level.
- control signals switchable between a ground voltage level and said peripheral voltage level.
- each column may comprise a single bit line
- each column is connected to a pair of bit lines.
- the write driver circuitry may be configured to control the voltage on the pair of bit lines of each of the addressed memory cells during the write operation by maintaining one of the bit lines in said pair at the peripheral voltage level and discharging the voltage on the other of the bit lines in said pair. Since the bit lines are precharged to the peripheral voltage level, the power consumption involved in discharging one of the bit lines during the write operation is significantly reduced, when compared with an approach where the bit lines were precharged to the core voltage level. Further, the power consumption involved in subsequently precharging the discharged bit line is also reduced.
- the access circuitry may also perform read operations in order to read data from a plurality of addressed memory cells.
- the word line driver circuitry and the precharge circuitry operate in the same manner as for the write operation.
- the access circuitry further comprises: sense amplifier circuitry configured to determine the data stored in the addressed memory cells by monitoring the voltage on the at least one bit line of each of the addressed memory cells during the read operation; read multiplexing circuitry configured, during the read operation, to couple the at least one bit line of each addressed memory cell to the sense amplifier circuitry, in dependence on a read mux control signal identifying which column in each of said plurality of column groups contains the addressed memory cells; and read multiplexing driver circuitry configured to assert said read mux control signal to the read multiplexing circuitry at said peripheral voltage level.
- the read multiplexing driver circuitry asserts the read mux control signal at the peripheral voltage level. Further, since the bit lines are precharged to the peripheral voltage level, the data as sensed by the sense amplifier circuitry can be returned as read data to the peripheral voltage domain without any level down shifting being required.
- operation of the precharge circuitry and the sense amplifier circuitry is controlled by associated control signals switchable between a ground voltage level and said peripheral voltage level, thereby reducing power consumption in those components when compared with a situation where the control signals are in the core voltage domain.
- the present invention provides a method of operating a memory device to perform a write operation, the memory device comprising an array of memory cells connected to a core voltage level and arranged as a plurality of rows and columns, each row being connected to an associated word line, and each column being connected to an associated at least one bit line, the columns being arranged into a plurality of column groups, each column group containing a plurality of columns, the method comprising: employing access circuitry to perform the write operation in order to write data into a plurality of addressed memory cells, said plurality of addressed memory cells comprising an addressed memory cell within each of a plurality of the column groups, the access circuitry receiving both the core voltage level and a peripheral voltage level less than said core voltage level; precharging, to said peripheral voltage level, the at least one bit line associated with at least each column containing an addressed memory cell, prior to said write operation being performed; during the write operation causing the access circuitry to: assert a word line signal at the core voltage level on the word line associated with a
- the present invention provides a memory device comprising: an array of memory cell means connected to a core voltage level and arranged as a plurality of rows and columns, each row being connected to an associated word line means, and each column being connected to an associated at least one bit line means, the columns being arranged into a plurality of column groups, each column group containing a plurality of columns; access means for performing a write operation in order to write data into a plurality of addressed memory cell means, said plurality of addressed memory cell means comprising an addressed memory cell means within each of a plurality of the column groups, the access means for receiving both the core voltage level and a peripheral voltage level less than said core voltage level; the access means comprising: word line driver means for asserting a word line signal at the core voltage level on the word line means associated with a row of the array containing the plurality of addressed memory cell means; precharge means for precharging, to said peripheral voltage level, the at least one bit line means associated with at least each column containing an addressed memory cell means, prior
- the present invention provides a computer program storage medium (for example a non-transitory storage medium) storing a memory compiler computer program for controlling a computer to generate an instance of a memory device from a memory architecture associated with the memory compiler computer program, the memory architecture specifying a definition of circuit elements and data defining rules for combining those circuit elements, such that said instance generated specifies a memory device in accordance with the first aspect of the present invention.
- a computer program storage medium for example a non-transitory storage medium
- a memory compiler computer program for controlling a computer to generate an instance of a memory device from a memory architecture associated with the memory compiler computer program, the memory architecture specifying a definition of circuit elements and data defining rules for combining those circuit elements, such that said instance generated specifies a memory device in accordance with the first aspect of the present invention.
- FIG. 1 is a diagram schematically illustrating a known memory device where the array of bit cells operate from a core voltage domain and the access logic circuitry operates from a peripheral voltage domain;
- FIG. 2 schematically illustrates a memory device in accordance with one embodiment
- FIG. 3 schematically illustrates the operation of the write column select circuitry of FIG. 2 in accordance with one embodiment
- FIG. 4 illustrates certain components within the memory device used when performing a write operation in accordance with one embodiment
- FIG. 5 schematically illustrates how the voltage levels change at various points within the circuitry of FIG. 4 in accordance with one embodiment
- FIG. 6 schematically illustrates address decoding circuitry and associated driver circuitry in accordance with one embodiment
- FIG. 7A illustrates a form of driver that may be used within the circuitry of FIG. 6 in accordance with one embodiment
- FIG. 7B illustrates a form of driver that may be used within the circuitry of FIG. 6 in accordance with an alternative embodiment
- FIG. 8 is a diagram schematically illustrating the operation of a memory compiler to generate a memory instance including a level shifting function on the word line and write column select signals in accordance with the described embodiments.
- FIG. 9 is a diagram of a computer system on which a memory compiler operation may be performed to generate a memory instance conforming to the above described embodiments.
- FIG. 2 is a diagram schematically illustrating a logical arrangement of a memory device in accordance with one embodiment.
- the memory device 100 includes a memory array 105 comprising an array of memory cells arranged in a plurality of rows and columns.
- a plurality of word lines 145 are provided through the array in order to allow individual rows of memory cells to be addressed by the word line driver circuitry 140 during write and read operations.
- a plurality of bit lines are provided in association with the columns of memory cells, in this embodiment each column having an associated pair of bit lines, to enable data to be written into an addressed memory cell of that column during a write operation, and for data to be read from an addressed memory cell of that column during a read operation.
- Precharge circuitry 115 is used to precharge the voltage level on the bit lines under the control of control circuitry 110 . Following the precharge operation, a write operation or a read operation can be performed.
- the various columns of memory cells within the memory array are arranged to form a plurality of column groups, with each column group containing a plurality of columns.
- the control circuitry 110 When a particular word line is activated during an access operation in order to activate a row of memory cells, the control circuitry 110 also issues a column select signal to the read column select circuitry 120 in the event of a read operation or to the write column select circuitry in the event of a write operation.
- the column select signal causes the appropriate column select circuitry to select one column within each of the column groups, the selected column being the column containing the addressed memory cell.
- the read column select circuitry 120 causes the bit lines of each selected column to be connected to the sense amplifier circuitry 130
- the write column select circuitry 125 causes the bit lines of each selected column to be connected to the write driver circuitry 135 .
- the control circuitry 110 will issue appropriate control signals to the word line driver circuitry 140 in order to cause a particular row of memory cells to be activated via the associated word line, and will also issue a write column select signal (in practice this is a collection of separate write column select signals) to the write column select circuitry 125 to cause one column from each column group to be selected, as a result of which the bit lines coupled to each addressed memory cell will then be coupled to the write driver circuitry 135 .
- the control circuitry 110 will also issue a write clock signal to the write driver circuitry 135 to cause the write driver circuitry 135 to control the voltages on the relevant bit line pairs, in order to cause the required data values to be written into the addressed memory cells.
- the write driver circuitry will selectively discharge one of the bit lines for each bit line pair connected to an addressed memory cell, in order to write the required data value into that addressed memory cell.
- control circuitry 110 will issue a control signal to the word line driver circuitry 140 in order to cause a particular row of memory cells to be activated via the appropriate word line, and the control circuitry 110 will issue a read column select signal (in practice a collection of signals) to the read column select circuitry 120 to cause one column in each column group to have its bit lines coupled to the sense amplifier circuitry 130 .
- the control circuitry 110 will also issue one or more control signals to the sense amplifier circuitry 130 in order to control the sense amplifier circuitry to evaluate the voltages on the relevant bit lines, with the sensed data then being output as the read data.
- bit lines will have been precharged to the logic one level
- one of the bit lines in the bit line pair of each addressed memory cell will discharge towards the logic zero level, and at some point during that discharge process, the sense amplifier circuitry 130 will be activated to sense the difference in voltage between the bit lines, and hence detect the data values stored in the addressed memory cells.
- control circuitry will receive a read/write select signal and an address used to determine which memory cells within the array should be written to, or read from.
- the control circuitry 110 also receives a clock signal and a clock enable signal to control operation of the circuitry.
- the memory array operates in a core voltage domain, and accordingly receives a core voltage level VDDC.
- the majority of the access circuitry formed by the other components shown in FIG. 2 operates in a peripheral voltage domain from a peripheral voltage level which is less than the core voltage level. In one particular embodiment, the peripheral voltage level is approximately 400 mV less than the core voltage level.
- all of the various control signals generated by the control circuitry 110 and issued to the word line driver circuitry 140 , the precharge circuitry 115 , the read column select circuitry 120 , the sense amplifier circuitry 130 and the write driver circuitry 135 are generated in the peripheral voltage domain.
- the precharge circuitry operates from the peripheral voltage supply, and accordingly precharges the bit lines to the peripheral voltage level VDDP prior to a read or a write operation.
- control circuitry 110 also receives the core voltage level VDDC, with that voltage level being used by write multiplexing driver circuitry within the control circuitry 110 such that the write column select signals over path 150 are issued in the core voltage domain.
- VDDC core voltage level
- the word line driver circuitry also receives the core voltage supply VDDC, such that the driver circuits within the word line driver circuitry used to assert word line signals on the word lines 145 generate word line signals in the core voltage domain.
- one of the word lines will be asserted by setting the voltage on that word line to the core voltage level VDDC.
- FIG. 3 schematically illustrates the operation of the write column select circuitry 125 in accordance with one embodiment.
- the write column select circuitry 125 comprises separate multiplexing circuits 220 , 250 for each bit of a data value to be written into the memory array.
- the memory array is configured in a four-way multiplexing manner, such that each group of memory cell columns 200 , 245 contains four columns of memory cells.
- the various memory cells 205 are coupled between the associated bit line pair 210 , 215 .
- the write driver circuitry 135 then consists of separate write driver circuits for each bit value to be written into the memory cells, and hence in FIG. 3 there is shown write driver circuitry 240 for bit zero and write driver circuitry 255 for bit one. It will be appreciated that the general layout shown in FIG. 3 will then be replicated for each additional bit of the data value to be written.
- Each multiplexing circuit 220 , 250 comprises a plurality of switching elements, and in particular separate switching elements for each column.
- a separate write column select signal is then generated for each column, and hence in this example four write column select signals are generated. For any particular write operation, only one of the write column select signals will be asserted (in this example at a logic one level) whilst all of the other write column select signals will remain de-asserted.
- each switching element consists of a pair of NMOS transistors which receive the write column select signal at their gate, and accordingly, when the write column select signal is asserted, serve to connect the corresponding bit line pair to the write driver circuitry.
- the switching element consisting of transistors 222 , 224 is provided in association with a column zero
- the switch element consisting of transistors 226 , 228 is provided in association with column one
- the switch element 230 , 232 is provided in association with column two
- the switch element consisting of transistors 234 , 236 is provided in association with column three. This same structure of switching elements is then replicated in each of the other multiplexers 250 .
- each of the switching elements is shown as a pair of NMOS transistors, it will be appreciated that the switching elements can be constructed in a variety of ways.
- each of the switching elements may be formed by a pair of transmission gates, each transmission gate comprising an NMOS transistor placed back to back with a PMOS transistor, with the NMOS transistor receiving the write column select signal and the PMOS transistor receiving an inverted version of that write column select signal.
- FIG. 4 schematically illustrates how a write operation is performed in respect of a single addressed memory cell.
- the memory cell is a 6T SRAM memory cell consisting of two NMOS pass gates 260 , 262 that are activated by a word line signal, and a storage element consisting of the PMOS transistors 264 , 266 and NMOS transistors 268 , 270 .
- the storage element is powered using the core voltage level VDDC in order to guarantee state retention within the memory cell.
- the precharge circuitry formed from the two PMOS transistors 274 , 278 will be activated by an asserted precharge signal at the logic zero level to precharge the bit lines BL and BLB to the peripheral voltage level VDDP.
- the precharge signal will be deasserted and the word line signal will be asserted at the core voltage level VDDC in order to strongly turn on the pass gates 260 , 262 .
- the relevant write column select signal will be asserted at the core voltage level VDDC in order to strongly turn on the switch elements 280 , 282 within the write column select circuitry 125 , and hence couple the bit lines 272 , 276 to the write driver circuitry 290 .
- the write driver circuitry operates in the peripheral voltage domain and during the write operation will discharge the voltage on one of the bit lines 272 , 276 dependent on the write data received over path 292 (which will either be at a zero voltage or the peripheral voltage dependent on the value to be written).
- FIG. 5 schematically illustrates how the voltages vary at various points within the circuitry of FIG. 4 in accordance with one embodiment.
- the word line signal is asserted, causing the voltage on the word line to transition from the logic zero level to the core voltage level VDDC at point 300 .
- a similar transition occurs on the write column signal at point 302 , with again the voltage transitioning to the core voltage level VDDC.
- the write driver circuitry will selectively discharge the voltage on one of the bit lines 272 , 276 .
- both bit lines will be at the peripheral voltage level VDDP.
- the bit line that is pulled down to logic zero level will have its voltage discharged as shown by the reference numeral 306 .
- bit line 292 Due to coupling effects, there will often be an initial dip in the other bit line as shown by the reference numeral 304 , but that bit line will then return to the VDDP level.
- the voltage on the write data line 292 will either be at the VDDP or the voltage zero level as shown by reference numeral 308 , dependent on the data to be written.
- the write operation is terminated by de-asserting the word line signal, as shown by the transition 318 , and similarly de-asserting the write column select signal, as shown by the transition 320 .
- the precharge transistors 274 , 276 can be reactivated to begin precharging the discharged bit line back up to the VDDP voltage level, as shown by the reference numeral 322 .
- FIG. 6 illustrates address decoder circuitry 337 and associated driver circuitry 340 that may be used in accordance with one embodiment.
- the address decoder circuitry 337 can be constructed in a variety of ways, but in this embodiment includes an internal address generator block 330 receiving a predetermined number of the bits of the address input to the control circuitry 110 , and an internal clock generator block 335 which receives one or more bits of the address along with an internal clock signal having a particular pulse width, the internal clock signal being derived from the clock signal received by the control circuitry 110 .
- the internal address generator 330 performs a predetermined logical operation on the received address bits in order to generate a plurality of internal address signals, each separate internal address signal being provided to a separate driver 342 , 344 , 346 , 348 within the driver circuitry 340 . Further, the internal clock generator 335 performs a logical operation based on the received signals in order to generate a further internal clock signal routed to each of the drivers 342 , 344 , 246 , 348 .
- Each driver then performs a logical operation on the two received input signals in order to generate an output signal at either a logic zero or a logic one level.
- the driver circuitry 340 operates from the core voltage level, and accordingly the output signals generated will be in the core voltage domain.
- one of the drivers 342 , 344 , 346 , 348 will assert its output signal at the VDDC voltage level, and all other drivers will output signals at a zero voltage level.
- the driver circuitry operates in the core voltage domain VDDC
- the entirety of the address decoder circuitry 337 operates in the peripheral voltage domain, and hence the various input signals received by the driver circuitry 340 will be in the peripheral voltage domain, and will need shifting into the core voltage domain.
- the general arrangement shown in FIG. 6 can be used to implement the write multiplexing driver circuitry that is configured to assert the write column select signals over path 150 to the write column select circuitry 125 of FIG. 2 .
- the driver circuitry 340 will include four separate drivers as shown in FIG. 6 , and each of the output signals will represent one of the write column select signals. In this instance, all of the circuitry shown in FIG. 6 resides within the control circuitry 110 of FIG. 2 .
- the general arrangement of circuitry shown in FIG. 6 can also be used to implement the word line driver circuits used to assert the word line signals over path 145 .
- the address decoder circuitry 337 may reside within the control circuitry 110 , or at least a part of that address decoder circuitry may reside within the word line driver circuitry 140 .
- the driver circuitry 340 will then reside within the word line driver circuitry 140 of FIG. 2 .
- each driver 342 will assert its output at the logic one level (in the VDDC voltage domain) if both of its inputs are at a logic one level (in the peripheral voltage domain VDDP).
- FIG. 7A illustrates one form of driver that can be used within each of the four driver blocks 342 , 344 , 346 , 348 shown in FIG. 6 .
- the two inputs (denoted as inputs A and B) in FIG. 7A are provided to dedicated level shifting circuitry 355 , which converts those input signals into signals in the core voltage domain.
- the driver logic 360 can then perform the required logical operation in order to generate the output signal (denoted by the letter C in FIG. 7A ) in the core voltage domain (at either a zero voltage or the VDDC voltage level dependent on the inputs A and B).
- the output signal denoted by the letter C in FIG. 7A
- the driver logic circuitry 360 performs a NAND operation on the inputs A and B, followed by an inverter operation (and hence the driver logic circuit 360 operates as an AND gate so that the output C is only asserted at a logic one level when both of the inputs are at the logic one level).
- the level shift circuitry 355 does occupy a relatively large area, and will consume power when performing the level shifting operations.
- the requirement for such a dedicated level shift circuitry can be avoided by applying a P/N skew process to the components forming the driver logic 360 . This is illustrated schematically in FIG. 7B , where the transistors forming both the NAND gate 365 and the inverter 370 are illustrated. In this embodiment, the inputs A and B are received in the peripheral voltage domain.
- the NAND gate is provided in the core voltage domain and accordingly, prior to the write operation taking place, the internal node X will be at the VDDC level by virtue of the PMOS transistors both being turned on.
- the NMOS transistors with the NAND gate of that driver are sufficiently strongly turned on to discharge the internal node X towards the logic zero level during the write operation, so that the output of the inverter 370 flips from the logic zero level to the output level at the voltage level VDDC.
- Table 1 illustrates some simulation results achieved for both a high performance memory cell array (HP1080) and a high current memory array (HC907), and contrasts the write margin (WRM) achieved when both the write column select signals and the word line drive signals are in the core voltage domain, as compared with a situation where only the word line drive signals are in the core voltage domain, and the write column select signals are retained in the peripheral voltage domain.
- WRM write margin
- the data also shows that the use of the described techniques shows an improved and adequate yield compared with an approach where only the word line drive signals are asserted in the core voltage domain.
- Both write margin and write performance improve when using the above described techniques, due to the stronger write multiplexing function.
- the write pulse width improves from 217 ps to 142 ps for the HP1080 bit cell when VDDC is at 0.935 volts and VDDP is at 0.535 volts.
- the access disturb margin (ADM) associated with read operations also improves as the voltage difference improves between VDDC and VDDP, due to less bit line disturbance as VDDP is lower.
- ADM access disturb margin
- Table 2 eventually there is a significant drop in the ADM value but only when the voltage difference reaches approximately 500 mV.
- both good ADM and WRM can be achieved with improved write performance when both the word line drive signals and the write column select signals are asserted in the core voltage domain.
- V 16FF Bitcell ADM Improvement with WL and Bitcell on VDDC FSG/125C, 256rpbl, LL, v0.1_2p1 VDDC VDDP HP1080 HC907 (V)
- V ADM ADM 0.935 0.935 8.1 7.5 0.935 0.835 8.4 7.7 0.935 0.735 8.9 8.1 0.935 0.635 9.7 8.7 0.935 0.535 18 15.4 0.935 0.435 2.9 2.3 0.765 0.765 7.7 6.87 0.765 0.635 8.1 7.18 0.765 0.535 8.67 7.6 0.765 0.435 15.9 13.28
- VDDC VDDP ADM V (V) (sigma) 1 1 6.3 1 0.9 6.5 1 0.8 7.0 1 0.7 7.4 1 0.6 6.9 1 0.5 Failed 0.8 0.8 5.4 0.8 0.7 5.7 0.8 0.6 6.1 0.8 0.5 5.8
- the cell has a write margin of 5.8 sigma when the WCOLSEL signal is also coupled to VDDC, compared with a fail condition if the WCOLSEL signal is coupled to VDDP.
- FIG. 8 schematically illustrates how a memory instance including the above described level shifting functionality in respect of the word line signals and the write column select signals may be created from a memory compiler 400 with reference to a memory architecture 410 .
- the memory architecture 410 specifies a definition of circuit elements and data defining rules for combining those circuit elements in order to create a memory instance.
- Particular requirements for the memory instance are entered into the memory compiler 400 as input parameters via a graphical user interface (GUI).
- GUI graphical user interface
- such input parameters can specify various aspects of the desired memory instance, for example defining the size of the memory array, the multiplexing arrangements of the memory array, selection of various optional features such as power gating features, built-in-self-test (BIST) modes to be supported, etc.
- the memory compiler 400 then generates the required memory instance based on the input parameters and the memory architecture 410 .
- the memory compiler includes the above described level shifting functionality in respect of the word line signals and the write column select signals to ensure that the memory instance will have sufficient write margin, but keeps other control signals in the peripheral voltage domain to reduce power consumption.
- FIG. 9 schematically illustrates a general purpose computer 500 of the type that may be used to implement the above described memory compilation operation in order to generate a memory instance.
- the general purpose computer 500 includes a central processing unit 502 , a random access memory 504 , a read only memory 506 , a network interface card 508 , a hard disk drive 510 , a display driver 512 and monitor 514 and a user input/output circuit 516 with a keyboard 518 and mouse 520 all connected via a common bus 522 .
- the central processing unit 502 will execute computer program instructions that may be stored in one or more of the random access memory 504 , the read only memory 506 and the hard disk drive 510 or dynamically downloaded via the network interface card 508 .
- the results of the processing performed may be displayed to a user via the display driver 512 and the monitor 514 .
- User inputs for controlling the operation of the general purpose computer 500 may be received via the user input output circuit 516 from the keyboard 518 or the mouse 520 (and hence for example the input parameters used to determine certain properties of the required memory instance can be entered via this mechanism).
- the computer program could be written in a variety of different computer languages.
- the computer program may be stored and distributed on a recording medium or dynamically downloaded to the general purpose computer 500 .
- the general purpose computer 500 can perform the above described memory compiler operation and can be considered to form an apparatus for performing the above described memory compiler operation.
- the architecture of the general purpose computer 500 could vary considerably and FIG. 9 is only one example.
Landscapes
- Engineering & Computer Science (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Computer Hardware Design (AREA)
- Static Random-Access Memory (AREA)
Abstract
Description
- 1. Field of the Invention
- The present invention relates to a memory device and to a method of operation of such a memory device, and in particular to a mechanism for operating such a memory device when the access circuitry used to access the array of memory cells is operated from a peripheral voltage domain with a supply voltage less than the supply voltage in a core voltage domain used by the array of memory cells.
- 2. Description of the Prior Art
- In modern data processing systems, it is becoming more and more common for certain parts of the data processing system to operate in a different voltage domain to one or more other parts. For example, a trend within integrated circuits is the increasingly common use of embedded memory, such as SRAM memory. With the reduction in size of process geometries, the individual memory cells within the memory are becoming less stable. To reduce the power consumption of the integrated circuit, it is desirable to reduce the operating voltage of the components within the integrated circuit. However, whilst this can be done for many of the components within the integrated circuit, including access logic circuitry associated with the memory device, it is often the case that a higher voltage is needed to drive the array of memory cells within the memory device in order to enhance the stability of those cells. Hence, embedded SRAM bit cells may use a higher voltage supply (providing a core voltage level) to guarantee state retention, whilst the rest of the system, including the access logic circuitry employed to access those bit cells within the memory device, may use a lower voltage supply (providing a peripheral voltage level) to reduce power consumption. In order to maintain performance, and reduce switching power, a level shifting mechanism is needed to pass signals between these different voltage domains.
-
FIG. 1 is a block diagram schematically illustrating aknown memory device 10 where theaccess logic circuitry 20 of the memory device is provided in a first voltage domain operating with a first voltage supply providing a peripheral voltage level VDDP and a ground voltage level (not shown), whilst the array ofbit cells 30 is provided in a second voltage domain operating with a second voltage supply providing a core voltage level VDDC and the ground voltage level. The core voltage level VDDC is higher than the peripheral voltage level VDDP. This enables theaccess logic circuitry 20 to operate with a reduced power consumption, whilst the array ofbit cells 30 is operated at a voltage level sufficient to guarantee state retention. - The
access logic circuitry 20 will receive a number of control signals overpath 55, these control signals identifying write transactions, read transactions, the addresses of those transactions, etc. It will also receive overpath 60 write data for write transactions to be performed within the array ofbit cells 30, and will output overpath 75 the read data resulting from read transactions performed within the array ofbit cells 30. As will be understood by those skilled in the art, theaccess logic circuitry 20 will include a number of components, such as address latches, various stages of word line decoding circuitry to decode the address in order to generate appropriate word line enable signals to activate addressed rows within the array ofbit cells 30, write data path logic circuits used to control the voltages on the bit lines within the array of bit cells during a write operation, and various read data path logic circuits for processing the data read out of the array of bit cells in response to a read operation. Accordingly, for write operations, a number of signals will be generated for issuing to the array of bit cells 30 (these signals being schematically illustrated by thearrow 65 inFIG. 1 ). In one known prior art approach, these signals are subjected to a level up shiftingfunction 40 in order to convert the voltage of those signals from the lower voltage domain to the higher voltage domain. Similarly, any data read out from thebit cells 30 overpath 70 is subjected to a level down shiftingfunction 50 in order to convert the voltage levels from the higher (core) voltage domain to the lower (peripheral) voltage domain, before those signals are then subsequently processed by theaccess logic circuitry 20. - The level up
shifting mechanism 40 is generally more problematic to implement than the level down shifting mechanism 50 (in fact in many instances no specific level down shifting circuitry may be required), since when performing level up shifting there is the potential for establishing various DC paths that can result in significant power consumption, and which may potentially create short circuit current paths. - When the voltage difference between the core voltage domain and the peripheral voltage domain is relatively small (for example up to 250 mV) it is not always necessary to level shift all of the control signals routed from the access logic circuitry to the bit cells, and read and write operations performed in respect of the bit cells will still operate correctly. However, in modern systems the voltage difference between the core voltage domain and the peripheral voltage domain is getting larger, and may for example be in the range of 400 mV. With such a large difference in the voltage domains, it has been found that level shifting of the control signals is required in order to ensure correct operation of the bit cells within the array when performing read and write operations. In particular, it has been found that the write margin can be insufficient unless such level shifting has been performed. Accordingly, when the difference between the core voltage domain and the peripheral voltage domain is of the order of 400 mV, it is known to perform level shifting on all of the control signals provided from the access logic circuitry to the bit cells. Whilst this ensures correct operation, it has a significant impact on the overall power consumption of the memory device.
- Accordingly, it would be desirable to provide a memory device which operates correctly when the voltage difference between the core voltage domain and the peripheral voltage domain is relatively large (e.g. of the order of 400 mV), but with a reduced power consumption compared with the known prior art approach.
- Viewed from a first aspect, the present invention provides a memory device comprising: an array of memory cells connected to a core voltage level and arranged as a plurality of rows and columns, each row being connected to an associated word line, and each column being connected to an associated at least one bit line, the columns being arranged into a plurality of column groups, each column group containing a plurality of columns; access circuitry configured to perform a write operation in order to write data into a plurality of addressed memory cells, said plurality of addressed memory cells comprising an addressed memory cell within each of a plurality of the column groups, the access circuitry being configured to receive both the core voltage level and a peripheral voltage level less than said core voltage level; the access circuitry comprising: word line driver circuitry configured to assert a word line signal at the core voltage level on the word line associated with a row of the array containing the plurality of addressed memory cells; precharge circuitry configured to precharge, to said peripheral voltage level, the at least one bit line associated with at least each column containing an addressed memory cell, prior to said write operation being performed; write driver circuitry configured to control the voltage on the at least one bit line of each of the addressed memory cells during the write operation in order to store write data into the plurality of addressed memory cells; write multiplexing circuitry configured, during the write operation, to couple the at least one bit line of each addressed memory cell to the write driver circuitry, in dependence on a mux control signal identifying which column in each of said plurality of column groups contains the addressed memory cells; and write multiplexing driver circuitry configured to assert said mux control signal to the write multiplexing circuitry at said core voltage level.
- In accordance with the present invention, the access circuitry is configured to receive both the core voltage level and the peripheral voltage level. Word line driver circuitry within the access circuitry is arranged to assert a word line signal at the core voltage level in order to activate a row containing the addressed memory cells for a write operation, and in addition write multiplexing driver circuitry asserts the required mux control signal(s) to write multiplexing circuitry at the core voltage level in order to cause the relevant components within the write multiplexing circuitry to couple the required columns of the array to the write driver circuitry. However, the precharge circuitry precharges the bit lines to the peripheral voltage level prior to the write operation being performed.
- By this approach, and in particular by asserting the word line signal and the mux control signal at the core voltage level, it has been found that sufficient write margin can be achieved to ensure write operations are performed correctly, even when the difference between the core voltage domain and the peripheral voltage domain is relatively large, for example 400 mV. However, since other control signals are retained at the peripheral voltage level, the overall power consumption of the memory device can be significantly reduced. For example, since the precharge circuitry precharges the bit lines only to the peripheral voltage level, rather than to the core voltage level, the power consumption associated with the discharging of certain bit lines during a write operation, and the subsequent precharging of those discharged bit lines back up to the peripheral voltage level prior to the next access operation, is significantly reduced.
- In one particular embodiment, the word line driver circuitry and the write multiplexing driver circuitry are the only components within the access circuitry that are configured to operate from the core voltage level, and all of the remaining components operate from the peripheral voltage level. This optimises the power consumption benefits achievable, whilst ensuring sufficient write margin is maintained.
- There are a number of ways in which the mux control signal can be asserted at the core voltage level by the write multiplexing driver circuitry. In one embodiment, the write multiplexing driver circuitry includes level shifting circuitry used to convert at least one signal from the peripheral voltage level to the core voltage level in order to cause the asserted mux control signal to be at the core voltage level. Hence, in this embodiment, dedicated level shifting circuitry is provided within the write multiplexing driver circuitry, in addition to the components required to perform the write multiplexing driver function.
- The at least one signal whose voltage level is converted by the level shifting circuitry can vary dependent on embodiment. However, in one embodiment the level shifting circuitry operates on at least one input signal to the write multiplexing driver circuitry in order to generate a corresponding at least one level shifted input signal, and the write multiplexing driver circuitry is configured to then generate the asserted mux control signal in dependence on said at least one level shifted input signal. In one particular embodiment, all of the input signals to the write multiplexing driver circuitry are subjected to this level shifting operation, and accordingly the functional components performing the write multiplexing driver operation within the write multiplexing driver circuitry operate entirely on input signals within the core voltage domain.
- Whilst in one embodiment the above-described dedicated level shifting circuits are used, such level shifting circuits occupy a relatively large area, and consume additional power to perform their level shifting operations. In an alternative embodiment, the requirement for such separate level shifting circuits can be avoided by instead applying a P/N skew to the transistors forming at least one of the components within the write multiplexing driver circuitry. In particular, in one embodiment, the write multiplexing driver circuitry comprises at least one circuit component formed of both NMOS transistors and PMOS transistors, and a drive strength of one of the NMOS transistors and the PMOS transistors is skewed so as to enable the write multiplexing driver circuitry to level shift at least one signal from the peripheral voltage level to the core voltage level in order to cause the asserted mux control signal to be at the core voltage level. Whilst the shifting range available from such a P/N skewing process is not as large as that available when using dedicated level shifting circuitry, it has been found that it can still provide a sufficient level shifting range to accommodate a 400 mV difference between the peripheral voltage domain and the core voltage domain, and accordingly can be used in order to reduce the area and power consumption associated with dedicated level shifting circuits.
- In one particular embodiment, the write multiplexing driver circuitry includes a NAND circuit component and a P/N skew is incorporated within the NAND circuit component to increase the drive strength of the NMOS transistors relative to the PMOS transistors.
- The above described options for constructing the write multiplexing driver circuitry also apply when constructing the word line driver circuitry used to assert the word line signal at the core voltage level. In particular, such word line driver circuitry may incorporate dedicated level shifting circuits, or can provide P/N skew within the transistors of one or more components of the word line driver circuitry in order to achieve the required shift in the voltage level from the peripheral voltage domain to the core voltage domain.
- The write multiplexing circuitry can be constructed in a variety of ways. However, in one embodiment the write multiplexing circuitry comprises, for each column group, switch circuitry configured to couple to the write driver circuitry the at least one bit line associated with one column in that column group. Hence, one column in each column group is coupled to the write driver circuitry, the column selected being the column that contains the addressed memory cell.
- The switch circuitry can take a variety of forms, but in one embodiment comprises a switch element for each column within the column group, and the mux control signal generated by the write multiplexing driver circuitry comprises a separate mux control signal for each switch element, with one of the separate mux control signals being asserted at the core voltage level to turn on the associated switch element.
- The individual switch elements can take a variety of forms. In one embodiment, they are formed by an NMOS transistor circuit whose gate is driven by the appropriate mux control signal. In an alternative embodiment, a transmission gate structure can be used including an NMOS transistor and a PMOS transistor placed back to back, with the input signal provided to the gate of the NMOS transistor being inverted prior to provision to the gate of the PMOS transistor.
- In one embodiment, operation of the precharge circuitry and the write driver circuitry is controlled by associated control signals switchable between a ground voltage level and said peripheral voltage level. Hence, the power consumption associated with these circuits is significantly reduced, when compared with driving those circuits using control signals that vary between the ground voltage level and the core voltage level.
- Whilst each column may comprise a single bit line, in one embodiment each column is connected to a pair of bit lines. In that embodiment, the write driver circuitry may be configured to control the voltage on the pair of bit lines of each of the addressed memory cells during the write operation by maintaining one of the bit lines in said pair at the peripheral voltage level and discharging the voltage on the other of the bit lines in said pair. Since the bit lines are precharged to the peripheral voltage level, the power consumption involved in discharging one of the bit lines during the write operation is significantly reduced, when compared with an approach where the bit lines were precharged to the core voltage level. Further, the power consumption involved in subsequently precharging the discharged bit line is also reduced.
- In addition to performing write operations, the access circuitry may also perform read operations in order to read data from a plurality of addressed memory cells. For the read operation the word line driver circuitry and the precharge circuitry operate in the same manner as for the write operation. However, to support read operations, the access circuitry further comprises: sense amplifier circuitry configured to determine the data stored in the addressed memory cells by monitoring the voltage on the at least one bit line of each of the addressed memory cells during the read operation; read multiplexing circuitry configured, during the read operation, to couple the at least one bit line of each addressed memory cell to the sense amplifier circuitry, in dependence on a read mux control signal identifying which column in each of said plurality of column groups contains the addressed memory cells; and read multiplexing driver circuitry configured to assert said read mux control signal to the read multiplexing circuitry at said peripheral voltage level.
- In contrast to the write multiplexing driver circuitry, it should be noted that the read multiplexing driver circuitry asserts the read mux control signal at the peripheral voltage level. Further, since the bit lines are precharged to the peripheral voltage level, the data as sensed by the sense amplifier circuitry can be returned as read data to the peripheral voltage domain without any level down shifting being required.
- In one embodiment, operation of the precharge circuitry and the sense amplifier circuitry is controlled by associated control signals switchable between a ground voltage level and said peripheral voltage level, thereby reducing power consumption in those components when compared with a situation where the control signals are in the core voltage domain.
- Viewed from a second aspect, the present invention provides a method of operating a memory device to perform a write operation, the memory device comprising an array of memory cells connected to a core voltage level and arranged as a plurality of rows and columns, each row being connected to an associated word line, and each column being connected to an associated at least one bit line, the columns being arranged into a plurality of column groups, each column group containing a plurality of columns, the method comprising: employing access circuitry to perform the write operation in order to write data into a plurality of addressed memory cells, said plurality of addressed memory cells comprising an addressed memory cell within each of a plurality of the column groups, the access circuitry receiving both the core voltage level and a peripheral voltage level less than said core voltage level; precharging, to said peripheral voltage level, the at least one bit line associated with at least each column containing an addressed memory cell, prior to said write operation being performed; during the write operation causing the access circuitry to: assert a word line signal at the core voltage level on the word line associated with a row of the array containing the plurality of addressed memory cells; employ write driver circuitry to control the voltage on the at least one bit line of each of the addressed memory cells in order to store write data into the plurality of addressed memory cells; couple the at least one bit line of each addressed memory cell to the write driver circuitry, in dependence on a mux control signal identifying which column in each of said plurality of column groups contains the addressed memory cells; and assert said mux control signal to the write multiplexing circuitry at said core voltage level.
- Viewed from a third aspect, the present invention provides a memory device comprising: an array of memory cell means connected to a core voltage level and arranged as a plurality of rows and columns, each row being connected to an associated word line means, and each column being connected to an associated at least one bit line means, the columns being arranged into a plurality of column groups, each column group containing a plurality of columns; access means for performing a write operation in order to write data into a plurality of addressed memory cell means, said plurality of addressed memory cell means comprising an addressed memory cell means within each of a plurality of the column groups, the access means for receiving both the core voltage level and a peripheral voltage level less than said core voltage level; the access means comprising: word line driver means for asserting a word line signal at the core voltage level on the word line means associated with a row of the array containing the plurality of addressed memory cell means; precharge means for precharging, to said peripheral voltage level, the at least one bit line means associated with at least each column containing an addressed memory cell means, prior to said write operation being performed; write driver means for controlling the voltage on the at least one bit line means of each of the addressed memory cell means during the write operation in order to store write data into the plurality of addressed memory cell means; write multiplexing means for coupling, during the write operation, the at least one bit line means of each addressed memory cell means to the write driver means, in dependence on a mux control signal identifying which column in each of said plurality of column groups contains the addressed memory cell means; and write multiplexing driver means for asserting said mux control signal to the write multiplexing means at said core voltage level.
- Viewed from a fourth aspect, the present invention provides a computer program storage medium (for example a non-transitory storage medium) storing a memory compiler computer program for controlling a computer to generate an instance of a memory device from a memory architecture associated with the memory compiler computer program, the memory architecture specifying a definition of circuit elements and data defining rules for combining those circuit elements, such that said instance generated specifies a memory device in accordance with the first aspect of the present invention.
- The present invention will be described further, by way of example only, with reference to embodiments thereof as illustrated in the accompanying drawings, in which:
-
FIG. 1 is a diagram schematically illustrating a known memory device where the array of bit cells operate from a core voltage domain and the access logic circuitry operates from a peripheral voltage domain; -
FIG. 2 schematically illustrates a memory device in accordance with one embodiment; -
FIG. 3 schematically illustrates the operation of the write column select circuitry ofFIG. 2 in accordance with one embodiment; -
FIG. 4 illustrates certain components within the memory device used when performing a write operation in accordance with one embodiment; -
FIG. 5 schematically illustrates how the voltage levels change at various points within the circuitry ofFIG. 4 in accordance with one embodiment; -
FIG. 6 schematically illustrates address decoding circuitry and associated driver circuitry in accordance with one embodiment; -
FIG. 7A illustrates a form of driver that may be used within the circuitry ofFIG. 6 in accordance with one embodiment; -
FIG. 7B illustrates a form of driver that may be used within the circuitry ofFIG. 6 in accordance with an alternative embodiment; -
FIG. 8 is a diagram schematically illustrating the operation of a memory compiler to generate a memory instance including a level shifting function on the word line and write column select signals in accordance with the described embodiments; and -
FIG. 9 is a diagram of a computer system on which a memory compiler operation may be performed to generate a memory instance conforming to the above described embodiments. -
FIG. 2 is a diagram schematically illustrating a logical arrangement of a memory device in accordance with one embodiment. In particular, thememory device 100 includes amemory array 105 comprising an array of memory cells arranged in a plurality of rows and columns. A plurality ofword lines 145 are provided through the array in order to allow individual rows of memory cells to be addressed by the wordline driver circuitry 140 during write and read operations. In addition, a plurality of bit lines are provided in association with the columns of memory cells, in this embodiment each column having an associated pair of bit lines, to enable data to be written into an addressed memory cell of that column during a write operation, and for data to be read from an addressed memory cell of that column during a read operation. -
Precharge circuitry 115 is used to precharge the voltage level on the bit lines under the control ofcontrol circuitry 110. Following the precharge operation, a write operation or a read operation can be performed. - The various columns of memory cells within the memory array are arranged to form a plurality of column groups, with each column group containing a plurality of columns. When a particular word line is activated during an access operation in order to activate a row of memory cells, the
control circuitry 110 also issues a column select signal to the read columnselect circuitry 120 in the event of a read operation or to the write column select circuitry in the event of a write operation. The column select signal causes the appropriate column select circuitry to select one column within each of the column groups, the selected column being the column containing the addressed memory cell. For a read operation, the read columnselect circuitry 120 causes the bit lines of each selected column to be connected to thesense amplifier circuitry 130, whilst for a write operation, the write columnselect circuitry 125 causes the bit lines of each selected column to be connected to thewrite driver circuitry 135. - For a write operation, the
control circuitry 110 will issue appropriate control signals to the wordline driver circuitry 140 in order to cause a particular row of memory cells to be activated via the associated word line, and will also issue a write column select signal (in practice this is a collection of separate write column select signals) to the write columnselect circuitry 125 to cause one column from each column group to be selected, as a result of which the bit lines coupled to each addressed memory cell will then be coupled to thewrite driver circuitry 135. Thecontrol circuitry 110 will also issue a write clock signal to thewrite driver circuitry 135 to cause thewrite driver circuitry 135 to control the voltages on the relevant bit line pairs, in order to cause the required data values to be written into the addressed memory cells. In particular, whilst the precharge circuitry will have precharged each bit line pair to a logic one level, during the write operation the write driver circuitry will selectively discharge one of the bit lines for each bit line pair connected to an addressed memory cell, in order to write the required data value into that addressed memory cell. - For a read operation, the
control circuitry 110 will issue a control signal to the wordline driver circuitry 140 in order to cause a particular row of memory cells to be activated via the appropriate word line, and thecontrol circuitry 110 will issue a read column select signal (in practice a collection of signals) to the read columnselect circuitry 120 to cause one column in each column group to have its bit lines coupled to thesense amplifier circuitry 130. Thecontrol circuitry 110 will also issue one or more control signals to thesense amplifier circuitry 130 in order to control the sense amplifier circuitry to evaluate the voltages on the relevant bit lines, with the sensed data then being output as the read data. In particular, whilst the bit lines will have been precharged to the logic one level, during the read operation one of the bit lines in the bit line pair of each addressed memory cell will discharge towards the logic zero level, and at some point during that discharge process, thesense amplifier circuitry 130 will be activated to sense the difference in voltage between the bit lines, and hence detect the data values stored in the addressed memory cells. - As shown in
FIG. 2 , for each write operation or read operation to be performed, the control circuitry will receive a read/write select signal and an address used to determine which memory cells within the array should be written to, or read from. Thecontrol circuitry 110 also receives a clock signal and a clock enable signal to control operation of the circuitry. - In accordance with one embodiment, the memory array operates in a core voltage domain, and accordingly receives a core voltage level VDDC. However, the majority of the access circuitry formed by the other components shown in
FIG. 2 operates in a peripheral voltage domain from a peripheral voltage level which is less than the core voltage level. In one particular embodiment, the peripheral voltage level is approximately 400 mV less than the core voltage level. As shown inFIG. 2 , all of the various control signals generated by thecontrol circuitry 110 and issued to the wordline driver circuitry 140, theprecharge circuitry 115, the read columnselect circuitry 120, thesense amplifier circuitry 130 and thewrite driver circuitry 135 are generated in the peripheral voltage domain. The precharge circuitry operates from the peripheral voltage supply, and accordingly precharges the bit lines to the peripheral voltage level VDDP prior to a read or a write operation. - However, the
control circuitry 110 also receives the core voltage level VDDC, with that voltage level being used by write multiplexing driver circuitry within thecontrol circuitry 110 such that the write column select signals overpath 150 are issued in the core voltage domain. In one embodiment, when a particular write column select signal is asserted, it will be asserted at the core voltage level VDDC. Similarly, the word line driver circuitry also receives the core voltage supply VDDC, such that the driver circuits within the word line driver circuitry used to assert word line signals on the word lines 145 generate word line signals in the core voltage domain. In particular, for a read or a write operation, one of the word lines will be asserted by setting the voltage on that word line to the core voltage level VDDC. - It has been found that by asserting the required write column
select signal 150 at the core voltage level and the requiredword line 145 at the core voltage level, but leaving all other control signals in the peripheral voltage domain, this ensures sufficient write margin during the performance of write operations, whilst avoiding unnecessary power consumption that would arise as a result of issuing any of the other control signals in the core voltage domain. A significant source of power consumption arises from the precharging of the bit lines prior to each write or read operation. Since theprecharge circuitry 115 operates in the peripheral voltage domain, and precharges the bit lines to the peripheral voltage level, this consumes significantly less power than is required if those bit lines were to have to be precharged to the core voltage level VDDC. Further, less power consumption is subsequently consumed when those bit lines are selectively discharged during the write and read operations. -
FIG. 3 schematically illustrates the operation of the write columnselect circuitry 125 in accordance with one embodiment. In particular, the write columnselect circuitry 125 comprises 220, 250 for each bit of a data value to be written into the memory array. In the example illustrated, it is assumed that the memory array is configured in a four-way multiplexing manner, such that each group ofseparate multiplexing circuits 200, 245 contains four columns of memory cells. Within any particular column, thememory cell columns various memory cells 205 are coupled between the associated 210, 215. Thebit line pair write driver circuitry 135 then consists of separate write driver circuits for each bit value to be written into the memory cells, and hence inFIG. 3 there is shownwrite driver circuitry 240 for bit zero and writedriver circuitry 255 for bit one. It will be appreciated that the general layout shown inFIG. 3 will then be replicated for each additional bit of the data value to be written. - Each
220, 250 comprises a plurality of switching elements, and in particular separate switching elements for each column. A separate write column select signal is then generated for each column, and hence in this example four write column select signals are generated. For any particular write operation, only one of the write column select signals will be asserted (in this example at a logic one level) whilst all of the other write column select signals will remain de-asserted.multiplexing circuit - In this embodiment, each switching element consists of a pair of NMOS transistors which receive the write column select signal at their gate, and accordingly, when the write column select signal is asserted, serve to connect the corresponding bit line pair to the write driver circuitry. Hence, as shown in
FIG. 3 for the fourway multiplexer 220, the switching element consisting of 222, 224 is provided in association with a column zero, the switch element consisting oftransistors 226, 228 is provided in association with column one, thetransistors 230, 232 is provided in association with column two and the switch element consisting ofswitch element 234, 236 is provided in association with column three. This same structure of switching elements is then replicated in each of thetransistors other multiplexers 250. - Whilst in
FIG. 3 , each of the switching elements is shown as a pair of NMOS transistors, it will be appreciated that the switching elements can be constructed in a variety of ways. For example, in an alternative embodiment, each of the switching elements may be formed by a pair of transmission gates, each transmission gate comprising an NMOS transistor placed back to back with a PMOS transistor, with the NMOS transistor receiving the write column select signal and the PMOS transistor receiving an inverted version of that write column select signal. -
FIG. 4 schematically illustrates how a write operation is performed in respect of a single addressed memory cell. In this example, the memory cell is a 6T SRAM memory cell consisting of two 260, 262 that are activated by a word line signal, and a storage element consisting of theNMOS pass gates 264, 266 andPMOS transistors 268, 270. There are two internal nodes within the storage element, labelled as node zero and node one, and it will be appreciated that the data bit stored within the storage element is dependent on which of node zero or node one is at a logic one value (whichever node is at a logic one value, the other node will be at a logic zero value). The storage element is powered using the core voltage level VDDC in order to guarantee state retention within the memory cell.NMOS transistors - Prior to a write operation being performed, the precharge circuitry formed from the two
274, 278 will be activated by an asserted precharge signal at the logic zero level to precharge the bit lines BL and BLB to the peripheral voltage level VDDP.PMOS transistors - During a write operation, the precharge signal will be deasserted and the word line signal will be asserted at the core voltage level VDDC in order to strongly turn on the
260, 262. In addition, the relevant write column select signal will be asserted at the core voltage level VDDC in order to strongly turn on thepass gates 280, 282 within the write columnswitch elements select circuitry 125, and hence couple the 272, 276 to thebit lines write driver circuitry 290. The write driver circuitry operates in the peripheral voltage domain and during the write operation will discharge the voltage on one of the 272, 276 dependent on the write data received over path 292 (which will either be at a zero voltage or the peripheral voltage dependent on the value to be written).bit lines -
FIG. 5 schematically illustrates how the voltages vary at various points within the circuitry ofFIG. 4 in accordance with one embodiment. At the beginning of the write operation, the word line signal is asserted, causing the voltage on the word line to transition from the logic zero level to the core voltage level VDDC atpoint 300. A similar transition occurs on the write column signal atpoint 302, with again the voltage transitioning to the core voltage level VDDC. Thereafter, the write driver circuitry will selectively discharge the voltage on one of the 272, 276. At the start of this process, both bit lines will be at the peripheral voltage level VDDP. The bit line that is pulled down to logic zero level will have its voltage discharged as shown by thebit lines reference numeral 306. Due to coupling effects, there will often be an initial dip in the other bit line as shown by thereference numeral 304, but that bit line will then return to the VDDP level. The voltage on thewrite data line 292 will either be at the VDDP or the voltage zero level as shown byreference numeral 308, dependent on the data to be written. - In this example, it is assumed that the write data is going to flip the state of the storage element. At the start of the write operation, it assumed that node zero is at the core voltage level VDDC as shown by
reference numeral 310, and that node one is at the zero voltage level, as shown byreference numeral 312. As the voltage on one of the bit lines is discharged towards the logic zero level, this causes the state of the internal nodes to flip as shown generally by the voltage changes illustrated inFIG. 5 , such that at the end of the write operation the voltage on node one will at the VDDC voltage level as shown by thereference numeral 314, and the voltage on node zero will have changed to a zero voltage level, as shown by thereference numeral 316. At this point, the write operation is terminated by de-asserting the word line signal, as shown by thetransition 318, and similarly de-asserting the write column select signal, as shown by thetransition 320. At this point, the 274, 276 can be reactivated to begin precharging the discharged bit line back up to the VDDP voltage level, as shown by theprecharge transistors reference numeral 322. -
FIG. 6 illustratesaddress decoder circuitry 337 and associateddriver circuitry 340 that may be used in accordance with one embodiment. Theaddress decoder circuitry 337 can be constructed in a variety of ways, but in this embodiment includes an internaladdress generator block 330 receiving a predetermined number of the bits of the address input to thecontrol circuitry 110, and an internalclock generator block 335 which receives one or more bits of the address along with an internal clock signal having a particular pulse width, the internal clock signal being derived from the clock signal received by thecontrol circuitry 110. Theinternal address generator 330 performs a predetermined logical operation on the received address bits in order to generate a plurality of internal address signals, each separate internal address signal being provided to a 342, 344, 346, 348 within theseparate driver driver circuitry 340. Further, theinternal clock generator 335 performs a logical operation based on the received signals in order to generate a further internal clock signal routed to each of the 342, 344, 246, 348.drivers - Each driver then performs a logical operation on the two received input signals in order to generate an output signal at either a logic zero or a logic one level. As shown in
FIG. 6 , thedriver circuitry 340 operates from the core voltage level, and accordingly the output signals generated will be in the core voltage domain. In one particular embodiment, during any particular read or write operation, one of the 342, 344, 346, 348 will assert its output signal at the VDDC voltage level, and all other drivers will output signals at a zero voltage level. Whilst the driver circuitry operates in the core voltage domain VDDC, the entirety of thedrivers address decoder circuitry 337 operates in the peripheral voltage domain, and hence the various input signals received by thedriver circuitry 340 will be in the peripheral voltage domain, and will need shifting into the core voltage domain. - The general arrangement shown in
FIG. 6 can be used to implement the write multiplexing driver circuitry that is configured to assert the write column select signals overpath 150 to the write columnselect circuitry 125 ofFIG. 2 . For the four-way multiplexing option discussed earlier with reference toFIG. 3 , thedriver circuitry 340 will include four separate drivers as shown inFIG. 6 , and each of the output signals will represent one of the write column select signals. In this instance, all of the circuitry shown inFIG. 6 resides within thecontrol circuitry 110 ofFIG. 2 . - The general arrangement of circuitry shown in
FIG. 6 can also be used to implement the word line driver circuits used to assert the word line signals overpath 145. In this embodiment, theaddress decoder circuitry 337 may reside within thecontrol circuitry 110, or at least a part of that address decoder circuitry may reside within the wordline driver circuitry 140. Thedriver circuitry 340 will then reside within the wordline driver circuitry 140 ofFIG. 2 . - As mentioned earlier, the
driver circuitry 340 needs to incorporate some level shifting functionality in order to shift the input signals from the peripheral voltage domain, so that the output signals can be generated in the core voltage domain. In one embodiment, eachdriver 342 will assert its output at the logic one level (in the VDDC voltage domain) if both of its inputs are at a logic one level (in the peripheral voltage domain VDDP). -
FIG. 7A illustrates one form of driver that can be used within each of the four 342, 344, 346, 348 shown indriver blocks FIG. 6 . The two inputs (denoted as inputs A and B) inFIG. 7A are provided to dedicatedlevel shifting circuitry 355, which converts those input signals into signals in the core voltage domain. Thedriver logic 360 can then perform the required logical operation in order to generate the output signal (denoted by the letter C inFIG. 7A ) in the core voltage domain (at either a zero voltage or the VDDC voltage level dependent on the inputs A and B). In one embodiment, as shown schematically inFIG. 7A , thedriver logic circuitry 360 performs a NAND operation on the inputs A and B, followed by an inverter operation (and hence thedriver logic circuit 360 operates as an AND gate so that the output C is only asserted at a logic one level when both of the inputs are at the logic one level). - Whilst an arrangement such as that shown in
FIG. 7A can support large level shifting functions with no leakage current, thelevel shift circuitry 355 does occupy a relatively large area, and will consume power when performing the level shifting operations. In an alternative embodiment, the requirement for such a dedicated level shift circuitry can be avoided by applying a P/N skew process to the components forming thedriver logic 360. This is illustrated schematically inFIG. 7B , where the transistors forming both theNAND gate 365 and theinverter 370 are illustrated. In this embodiment, the inputs A and B are received in the peripheral voltage domain. Accordingly, for the one driver that receives both inputs A and B at the logic one level, those inputs will only relatively weakly turn on the NMOS transistor pair coupling the output node to ground within theNAND gate 365. Further, the PMOS transistors of theNAND gate 365 will not be turned fully off, and hence will contribute some leakage current. However, this leakage current will only be present in the one driver circuit that is actually to assert a drive signal (whether that be a write column select drive signal or a word line drive signal). - The NAND gate is provided in the core voltage domain and accordingly, prior to the write operation taking place, the internal node X will be at the VDDC level by virtue of the PMOS transistors both being turned on. To ensure correct operation when asserting an output signal C within one of the drivers, it is important that the NMOS transistors with the NAND gate of that driver are sufficiently strongly turned on to discharge the internal node X towards the logic zero level during the write operation, so that the output of the
inverter 370 flips from the logic zero level to the output level at the voltage level VDDC. This can be achieved by applying a P/N skew process to the transistor components within theNAND gate 365, and in particular by making the two NMOS transistors relatively strong compared to the PMOS transistors, for example by increasing the width of the gate within both of those NMOS transistors. It has been found that such a P/N skew process can be sufficient to support a level up voltage shift of approximately 400 mV, and accordingly can avoid the need for separate level shifting circuits. - Table 1 below illustrates some simulation results achieved for both a high performance memory cell array (HP1080) and a high current memory array (HC907), and contrasts the write margin (WRM) achieved when both the write column select signals and the word line drive signals are in the core voltage domain, as compared with a situation where only the word line drive signals are in the core voltage domain, and the write column select signals are retained in the peripheral voltage domain. As shown in the table, the above described scheme where both the word line drive signals and the write column select signals are asserted in the core voltage domain provides significantly more write margin. This is true even up to a voltage difference of 500 mV between the core voltage domain and the peripheral voltage domain. The data also shows that the use of the described techniques shows an improved and adequate yield compared with an approach where only the word line drive signals are asserted in the core voltage domain. Both write margin and write performance (write pulse-width) improve when using the above described techniques, due to the stronger write multiplexing function. For example, the write pulse width improves from 217 ps to 142 ps for the HP1080 bit cell when VDDC is at 0.935 volts and VDDP is at 0.535 volts.
- As also shown by Table 2 below, the access disturb margin (ADM) associated with read operations also improves as the voltage difference improves between VDDC and VDDP, due to less bit line disturbance as VDDP is lower. As shown in Table 2, eventually there is a significant drop in the ADM value but only when the voltage difference reaches approximately 500 mV. At the anticipated useful voltage difference to be employed in many modern systems of approximately 400 mV, both good ADM and WRM can be achieved with improved write performance when both the word line drive signals and the write column select signals are asserted in the core voltage domain. Tables 1 and 2 are set out below:
-
TABLE 1 16FF Bitcell Write Margin Improvement with WL and WCOLSEL and Bitcell on VDDC HP1080 HC907 SFG/−40C, WCOLSEL = VDDP WCOLSEL = VDDC WCOLSEL = VDDC WCOLSEL = VDDC 128rpbl, LL, WL = VDDC WL = VDDC WL = VDDC WL = VDDC v0.1_2p1 5.2 sigma WL 5.2 sigma WL 5.2 sigma WL 5.2 sigma WL VDDC VDDP DC width DC width DC width DC width (V) (V) WRM (ps) WRM (ps) WRM (ps) WRM (ps) 0.935 0.935 13.2 174 13.2 174 8.8 185 8.8 185 0.935 0.635 12.5 184 12.9 149 8.3 197 8.5 164 0.935 0.535 11 217 12.6 142 7.6 227 8.3 158 0.935 0.435 7.1 523 11.4 144 4.2 Fail 7.6 163 0.765 0.765 8.8 215 8.8 215 5.8 279 5.8 279 0.765 0.635 8.6 226 8.7 203 5.6 309 5.7 279 0.765 0.535 8.3 251 8.6 195 5.4 380 5.6 285 0.765 0.435 7 346 8.4 194 4.8 Fail 5.4 323 -
TABLE 2 16FF Bitcell ADM Improvement with WL and Bitcell on VDDC FSG/125C, 256rpbl, LL, v0.1_2p1 VDDC VDDP HP1080 HC907 (V) (V) ADM ADM 0.935 0.935 8.1 7.5 0.935 0.835 8.4 7.7 0.935 0.735 8.9 8.1 0.935 0.635 9.7 8.7 0.935 0.535 18 15.4 0.935 0.435 2.9 2.3 0.765 0.765 7.7 6.87 0.765 0.635 8.1 7.18 0.765 0.535 8.67 7.6 0.765 0.435 15.9 13.28 - Tables 3 and 4 below show equivalent data obtained for another type of bitcell:
-
TABLE 3 14LPe Bitcell Write Margin Improvement with WL and WCOLSEL and Bitcell on VDDC HD Bitcell, HD Bitcell, WCOLSEL = VDDP WCOLSEL = VDDC WL = VDDC WL = VDDC 4.5 sigma DC 4.5 sigma VDDC VDDP DC WRM wl_pulsewidth WRM wl_pulsewidth (V) (V) (sigma) (ps) (sigma) (ps) 1 1 6.3 190.5 6.3 190.5 1 0.9 6.1 209.8 6.2 188.3 1 0.8 5.8 245.2 6.1 188.7 1 0.7 4.6 392.3 6.0 196.0 1 0.6 Fail Fail 5.8 227.3 1 0.5 Fail Fail 4.1 3.885 ns @ 4.1 sigma 3.0 sigma DC 3.0 sigma VDDC VDDP DC WRM wl_pulsewidth WRM wl_pulsewidth (V) (V) (sigma) (ps) (sigma) (ps) 0.8 0.8 4.1 265.3 4.1 265.3 0.8 0.7 3.9 338.4 4.1 265.0 0.8 0.6 3.1 664.2 3.9 280.9 0.8 0.6 fail fail 3.5 444.4 -
TABLE 4 14LPe Bitcell ADM Improvement with WL and Bitcell on VDDC HD Bitcell VDDC VDDP ADM (V) (V) (sigma) 1 1 6.3 1 0.9 6.5 1 0.8 7.0 1 0.7 7.4 1 0.6 6.9 1 0.5 Failed 0.8 0.8 5.4 0.8 0.7 5.7 0.8 0.6 6.1 0.8 0.5 5.8 - As can be seen from Table 3, in the 14 LPe bitcell, when the voltage difference is 400 mV (VDDC=1V, VDDP=0.6V), the cell has a write margin of 5.8 sigma when the WCOLSEL signal is also coupled to VDDC, compared with a fail condition if the WCOLSEL signal is coupled to VDDP.
-
FIG. 8 schematically illustrates how a memory instance including the above described level shifting functionality in respect of the word line signals and the write column select signals may be created from amemory compiler 400 with reference to amemory architecture 410. Thememory architecture 410 specifies a definition of circuit elements and data defining rules for combining those circuit elements in order to create a memory instance. Particular requirements for the memory instance are entered into thememory compiler 400 as input parameters via a graphical user interface (GUI). As will be appreciated by those skilled in the art, such input parameters can specify various aspects of the desired memory instance, for example defining the size of the memory array, the multiplexing arrangements of the memory array, selection of various optional features such as power gating features, built-in-self-test (BIST) modes to be supported, etc. - The
memory compiler 400 then generates the required memory instance based on the input parameters and thememory architecture 410. In accordance with one embodiment, the memory compiler includes the above described level shifting functionality in respect of the word line signals and the write column select signals to ensure that the memory instance will have sufficient write margin, but keeps other control signals in the peripheral voltage domain to reduce power consumption. -
FIG. 9 schematically illustrates ageneral purpose computer 500 of the type that may be used to implement the above described memory compilation operation in order to generate a memory instance. Thegeneral purpose computer 500 includes acentral processing unit 502, arandom access memory 504, a read onlymemory 506, anetwork interface card 508, ahard disk drive 510, adisplay driver 512 and monitor 514 and a user input/output circuit 516 with akeyboard 518 andmouse 520 all connected via acommon bus 522. In operation thecentral processing unit 502 will execute computer program instructions that may be stored in one or more of therandom access memory 504, the read onlymemory 506 and thehard disk drive 510 or dynamically downloaded via thenetwork interface card 508. The results of the processing performed may be displayed to a user via thedisplay driver 512 and themonitor 514. User inputs for controlling the operation of thegeneral purpose computer 500 may be received via the userinput output circuit 516 from thekeyboard 518 or the mouse 520 (and hence for example the input parameters used to determine certain properties of the required memory instance can be entered via this mechanism). It will be appreciated that the computer program could be written in a variety of different computer languages. The computer program may be stored and distributed on a recording medium or dynamically downloaded to thegeneral purpose computer 500. When operating under control of an appropriate computer program, thegeneral purpose computer 500 can perform the above described memory compiler operation and can be considered to form an apparatus for performing the above described memory compiler operation. The architecture of thegeneral purpose computer 500 could vary considerably andFIG. 9 is only one example. - From the above described embodiments, it will be appreciated that such embodiments enable write margin and write performance to be maintained in memory devices where the difference between the core voltage domain and the peripheral voltage domain is relatively large (for example of the order of 400 mV), whilst significantly reducing the power consumption in the memory device when compared with a memory device where all of the control signals from the access circuitry are level shifted to the core voltage domain. The above described techniques can hence result in memory devices occupying less area, having higher performance, and consuming less dynamic power than an equivalent memory device where a larger number of the control signals are level shifted, whilst ensuring sufficient write margin is achieved, and hence that write operations can be performed reliably.
- Although particular embodiments have been described herein, it will be appreciated that the invention is not limited thereto and that many modifications and additions thereto may be made within the scope of the invention. For example, various combinations of the features of the following dependent claims could be made with the features of the independent claims without departing from the scope of the present invention.
Claims (19)
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US14/037,413 US8971133B1 (en) | 2013-09-26 | 2013-09-26 | Memory device and method of operation of such a memory device |
| GB1414373.9A GB2519410B (en) | 2013-09-26 | 2014-08-13 | A memory device and method of operation of such a memory device |
| TW103131183A TWI622993B (en) | 2013-09-26 | 2014-09-10 | A memory device and method of operation of such a memory device |
| KR1020140120820A KR102220632B1 (en) | 2013-09-26 | 2014-09-12 | A memory device and method of operation of such a memory device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US14/037,413 US8971133B1 (en) | 2013-09-26 | 2013-09-26 | Memory device and method of operation of such a memory device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| US8971133B1 US8971133B1 (en) | 2015-03-03 |
| US20150085586A1 true US20150085586A1 (en) | 2015-03-26 |
Family
ID=51629733
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US14/037,413 Active US8971133B1 (en) | 2013-09-26 | 2013-09-26 | Memory device and method of operation of such a memory device |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US8971133B1 (en) |
| KR (1) | KR102220632B1 (en) |
| GB (1) | GB2519410B (en) |
| TW (1) | TWI622993B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150169042A1 (en) * | 2013-12-16 | 2015-06-18 | Sandisk Technologies Inc. | Low power interface for a data storage device |
| GB2567420A (en) * | 2017-10-02 | 2019-04-17 | Advanced Risc Mach Ltd | Adaptive voltage scaling methods and systems therefor |
| US10832745B1 (en) | 2019-07-26 | 2020-11-10 | Micron Technology, Inc. | Apparatuses and methods for performing operations using sense amplifiers and intermediary circuitry |
| TWI711045B (en) * | 2018-06-28 | 2020-11-21 | 台灣積體電路製造股份有限公司 | Memory storage device, write driver and method for operating thereof |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9940987B2 (en) * | 2015-03-16 | 2018-04-10 | Qualcomm Incorporated | High-speed word line decoder and level-shifter |
| CN104978988B (en) * | 2015-05-22 | 2017-08-25 | 江苏时代全芯存储科技有限公司 | Memory device |
| US9966131B2 (en) * | 2015-08-21 | 2018-05-08 | Synopsys, Inc. | Using sense amplifier as a write booster in memory operating with a large dual rail voltage supply differential |
| US10163470B2 (en) * | 2015-09-18 | 2018-12-25 | Taiwan Semiconductor Manufacturing Company Ltd. | Dual rail memory, memory macro and associated hybrid power supply method |
| US9754923B1 (en) * | 2016-05-09 | 2017-09-05 | Qualcomm Incorporated | Power gate placement techniques in three-dimensional (3D) integrated circuits (ICs) (3DICs) |
| US10755770B2 (en) * | 2016-09-30 | 2020-08-25 | Taiwan Semiconductor Manufacturing Company, Ltd. | Circuit and method for writing to a bit cell |
| DE102017117791B4 (en) * | 2016-09-30 | 2025-06-12 | Taiwan Semiconductor Manufacturing Co. Ltd. | Circuit and method for writing to a bit cell |
| US9875774B1 (en) * | 2016-11-29 | 2018-01-23 | Taiwan Semiconductor Manufacturing Company, Ltd. | Memory device and method of operating same |
| CN117524279A (en) | 2017-11-15 | 2024-02-06 | 三星电子株式会社 | SRAM with virtualized architecture and systems and methods including the same |
| CN111289884A (en) * | 2018-12-27 | 2020-06-16 | 展讯通信(上海)有限公司 | Testing device, chip and method for testing voltage of memory |
| US11990179B2 (en) * | 2020-10-14 | 2024-05-21 | Samsung Electronics Co., Ltd. | Memory device using a plurality of supply voltages and operating method thereof |
| KR102734241B1 (en) * | 2021-10-28 | 2024-11-26 | 숙명여자대학교 산학협력단 | Embedded memory for artificial neural network accelerator and driving method thereof |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7830735B2 (en) * | 2001-02-28 | 2010-11-09 | Rambus Inc. | Asynchronous, high-bandwidth memory component using calibrated timing elements |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5633832A (en) * | 1995-09-26 | 1997-05-27 | Alliance Semiconductor Corporation | Reduced area word line driving circuit for random access memory |
| JP2001312893A (en) * | 2000-04-28 | 2001-11-09 | Toshiba Corp | Semiconductor device |
| JP2004234810A (en) * | 2003-02-03 | 2004-08-19 | Renesas Technology Corp | Semiconductor memory |
| US7463545B2 (en) * | 2006-03-17 | 2008-12-09 | Texas Instruments Incorporated | System and method for reducing latency in a memory array decoder circuit |
| US7368946B1 (en) * | 2006-06-16 | 2008-05-06 | Xilinx, Inc. | Level-shifting pass gate multiplexer |
| US7995410B2 (en) * | 2009-06-26 | 2011-08-09 | Apple Inc. | Leakage and NBTI reduction technique for memory |
| US8279687B2 (en) * | 2010-05-13 | 2012-10-02 | International Business Machines Corporation | Single supply sub VDD bit-line precharge SRAM and method for level shifting |
| US9299395B2 (en) * | 2012-03-26 | 2016-03-29 | Intel Corporation | Methods and systems to selectively boost an operating voltage of, and controls to an 8T bit-cell array and/or other logic blocks |
-
2013
- 2013-09-26 US US14/037,413 patent/US8971133B1/en active Active
-
2014
- 2014-08-13 GB GB1414373.9A patent/GB2519410B/en active Active
- 2014-09-10 TW TW103131183A patent/TWI622993B/en active
- 2014-09-12 KR KR1020140120820A patent/KR102220632B1/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7830735B2 (en) * | 2001-02-28 | 2010-11-09 | Rambus Inc. | Asynchronous, high-bandwidth memory component using calibrated timing elements |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150169042A1 (en) * | 2013-12-16 | 2015-06-18 | Sandisk Technologies Inc. | Low power interface for a data storage device |
| GB2567420A (en) * | 2017-10-02 | 2019-04-17 | Advanced Risc Mach Ltd | Adaptive voltage scaling methods and systems therefor |
| GB2567420B (en) * | 2017-10-02 | 2020-07-08 | Advanced Risc Mach Ltd | Adaptive voltage scaling methods and systems therefor |
| US11307244B2 (en) * | 2017-10-02 | 2022-04-19 | Arm Limited | Adaptive voltage scaling methods and systems therefor |
| TWI711045B (en) * | 2018-06-28 | 2020-11-21 | 台灣積體電路製造股份有限公司 | Memory storage device, write driver and method for operating thereof |
| US10832745B1 (en) | 2019-07-26 | 2020-11-10 | Micron Technology, Inc. | Apparatuses and methods for performing operations using sense amplifiers and intermediary circuitry |
| WO2021021288A1 (en) * | 2019-07-26 | 2021-02-04 | Micron Technology, Inc. | Apparatuses and methods for performing operations using sense amplifiers and intermediary circuitry |
| US11222669B2 (en) | 2019-07-26 | 2022-01-11 | Micron Technology, Inc. | Apparatuses and methods for performing operations using sense amplifiers and intermediary circuitry |
| US11783872B2 (en) | 2019-07-26 | 2023-10-10 | Micron Technology, Inc. | Apparatuses and methods for performing operations using sense amplifiers and intermediary circuitry |
Also Published As
| Publication number | Publication date |
|---|---|
| TWI622993B (en) | 2018-05-01 |
| US8971133B1 (en) | 2015-03-03 |
| KR102220632B1 (en) | 2021-02-26 |
| KR20150034613A (en) | 2015-04-03 |
| GB201414373D0 (en) | 2014-09-24 |
| GB2519410B (en) | 2017-02-08 |
| GB2519410A (en) | 2015-04-22 |
| TW201515006A (en) | 2015-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8971133B1 (en) | Memory device and method of operation of such a memory device | |
| US9171634B2 (en) | Memory device and method of controlling leakage current within such a memory device | |
| US10431269B2 (en) | Methods and apparatus for reducing power consumption in memory circuitry by controlling precharge duration | |
| US9147451B2 (en) | Memory device and method of controlling leakage current within such a memory device | |
| US9064559B2 (en) | Memory device and method of performing access operations within such a memory device | |
| US8963609B2 (en) | Combinatorial circuit and method of operation of such a combinatorial circuit | |
| US8947968B2 (en) | Memory having power saving mode | |
| KR102093636B1 (en) | A memory device and method of performing a read operation within such a memory device | |
| CN103229242A (en) | Static random access memory (SRAM) write assist circuit with leakage suppression and level control | |
| US20160049191A1 (en) | Integrated Circuit for Storing Data | |
| US9881669B1 (en) | Wordline driver with integrated voltage level shift function | |
| CN102842338A (en) | Storage device and method for controlling storage device | |
| TW201019343A (en) | A memory device and method of operating such a memory device | |
| US20050047220A1 (en) | Semiconductor memory device | |
| US20130077416A1 (en) | Memory device and method of performing a read operation within a memory device | |
| US8854902B2 (en) | Write self timing circuitry for self-timed memory | |
| US9679636B2 (en) | Memory architecture | |
| JP2010287266A (en) | SRAM (Static Random Access Memory) and method of accessing SRAM |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| AS | Assignment |
Owner name: ARM LIMITED, UNITED KINGDOM Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:ZHENG, BO;KWON, JUNGTAE;YEUNG, GUS;AND OTHERS;SIGNING DATES FROM 20130927 TO 20131018;REEL/FRAME:031776/0803 |
|
| FEPP | Fee payment procedure |
Free format text: PAYOR NUMBER ASSIGNED (ORIGINAL EVENT CODE: ASPN); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
| STCF | Information on status: patent grant |
Free format text: PATENTED CASE |
|
| MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 4TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1551); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 4 |
|
| MAFP | Maintenance fee payment |
Free format text: PAYMENT OF MAINTENANCE FEE, 8TH YEAR, LARGE ENTITY (ORIGINAL EVENT CODE: M1552); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY Year of fee payment: 8 |