CN114882920A - Approximate search TCAM matching system based on current mode measurement - Google Patents

Approximate search TCAM matching system based on current mode measurement Download PDF

Info

Publication number
CN114882920A
CN114882920A CN202210427548.9A CN202210427548A CN114882920A CN 114882920 A CN114882920 A CN 114882920A CN 202210427548 A CN202210427548 A CN 202210427548A CN 114882920 A CN114882920 A CN 114882920A
Authority
CN
China
Prior art keywords
tcam
output
search
signal
array
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.)
Pending
Application number
CN202210427548.9A
Other languages
Chinese (zh)
Inventor
尹勋钊
陈雨过
王观涛
黄文韬
卓成
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang University ZJU
Original Assignee
Zhejiang University ZJU
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang University ZJU filed Critical Zhejiang University ZJU
Priority to CN202210427548.9A priority Critical patent/CN114882920A/en
Publication of CN114882920A publication Critical patent/CN114882920A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • G11C7/1051—Data output circuits, e.g. read-out amplifiers, data output buffers, data output registers, data output level conversion circuits
    • G11C7/1063—Control signal output circuits, e.g. status or busy flags, feedback command signals
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C7/00—Arrangements for writing information into, or reading information out from, a digital store
    • G11C7/06—Sense amplifiers; Associated circuits, e.g. timing or triggering circuits
    • G11C7/062—Differential amplifiers of non-latching type, e.g. comparators, long-tailed pairs
    • 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
    • G11C7/109—Control signal input circuits
    • G—PHYSICS
    • G11—INFORMATION STORAGE
    • G11C—STATIC STORES
    • G11C7/00—Arrangements for writing information into, or reading information out from, a digital store
    • G11C7/16—Storage of analogue signals in digital stores using an arrangement comprising analogue/digital [A/D] converters, digital memories and digital/analogue [D/A] converters 
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

The invention discloses an approximate search TCAM matching system based on current mode measurement, which comprises a TCAM array and a sensing amplifier, wherein the TCAM array is used as a front-stage circuit, and the sensing amplifier is used as a rear-stage circuit and is used for receiving analog current signals output by each row of TCAM units in the TCAM array and converting the analog current signals into digital pulses with the same mismatching degree. The invention fully utilizes the matching discharge characteristic of the TCAM, realizes the design of a brand-new TCAM approximate search array and a measurement circuit, realizes the measurement of the number of search unmatched bits, and obtains the expandability of a storage array.

Description

Approximate search TCAM matching system based on current mode measurement
Technical Field
The invention relates to the field of storage and calculation integration, in particular to an approximate search TCAM matching system based on current mode measurement, which considers the application of a current mode measurement method and is used for designing a low-power-consumption high-performance TCAM with an approximate search function.
Background
In the big data era, data interaction between memory and processors has become a short plate of computer architecture performance. As one of the most potential solutions, the computing body has made great progress in parallel search and the like. Ternary Content Addressable Memory (TCAM) has attracted wide attention because it realizes a function of matching only part of keywords while addressing in high parallel. The unique functionality of TCAMs makes them very versatile in applications including routers, data mining, artificial intelligence models, and the like.
In the aspect of devices, the TCAM based on the CMOSMOS tube has the problems of large exposed area, low energy efficiency and the like. For this reason, new nonvolatile devices are introduced into the field of storage, including Ferroelectric field effect MOS transistors (fefets), Magnetic Tunnel Junctions (Magnetic Tunnel Junctions), and the like. The basic TCAM storage unit structure based on the FeFET is successfully applied to scenes such as a few-sample learning hardware architecture.
Functionally, the exact search pattern of conventional TCAMs requires that the search data match the stored content exactly. When mass data generated by a user are mostly output unmatched signals, the energy efficiency of the system is improved to a limited extent by the highly parallel hardware searching module, and a large amount of energy consumption overhead and searching time waste are caused. The storage content of approximate search output and the search data which are closest to each other can greatly reduce the search time and power consumption by allowing the output deviation within a reasonable fault tolerance range, and the contradiction between the CAM hardware scale cost and massive data query search operation, which can only accurately search limited data, can be relieved. The technical problem of how to obtain a quantitative comparison result of the degree of mismatch between the search data and the stored content has not been completely overcome.
Disclosure of Invention
The invention aims to provide an approximate search TCAM matching system based on current mode measurement to obtain a lower energy delay product aiming at the problems of lower working efficiency and larger energy consumption and delay of the existing accurate search TCAM.
The purpose of the invention is realized by the following technical scheme:
the invention provides an approximate search TCAM matching system based on current mode measurement, which comprises a TCAM array and a sensing amplifier, wherein the TCAM array is used as a front-stage circuit, and the sensing amplifier is used as a rear-stage circuit and is used for receiving analog current signals output by each row of TCAM units in the TCAM array and converting the analog current signals into digital pulses with the same mismatching degree.
Further, the current measuring method of the TCAM cell includes: the pre-charging mechanism of original VDD to ML is cancelled, original ML is connected with VDD and is maintained at high voltage, TCAM is output to ground to be used as new ML, and output current is used as a measurement index.
Furthermore, the sensing amplifier comprises an input transmission gate, a capacitor, an output transmission gate and a Schmitt trigger which are sequentially and electrically connected, when the sensing amplifier works, an input current signal sequentially passes through the input transmission gate controlled by the enable control and the output signal control to charge the capacitor, the output voltage of the capacitor is output through the Schmitt trigger, and the output signal is fed back to the discharge transmission gate.
Further, there are two input transmission gates, which are the first transmission gate TG1 and the second transmission gate TG2, respectively, two discharge transmission gates, which are the third transmission gate TG3 and the fourth transmission gate TG4, respectively, and two diplexers, which are the first diplexer and the second diplexer, respectively;
by enabling signal V EN And outputs a feedback signal V SPIKE The first transmission gate TG1 and the second transmission gate TG2 which are respectively controlled are connected in series and are used as isolation between the front-stage TCAM array and the measurement capacitor;
the output current of the TCAM array passes through a first transmission gate TG1 and a second transmission gate TG2 and then is coupled with a capacitor C SA Charging so that the capacitor C SA Output voltage linearly rising, capacitor C SA The other end is grounded;
by enabling signal V EN And outputs a feedback signal V SPIKE A third transmission gate TG3 and a fourth transmission gate TG4 which are respectively controlled in an inverted way are connected in parallel with the capacitor C SA Output terminal and reset signal V RESET Form a fast discharge loop;
capacitor C SA The output signal sequentially passes through a first phase inverter for amplitude expansion, a periodically-turned Schmitt trigger and a second phase inverter to obtain an output pulse signal V SPIKE ；
Capacitor C SA The voltage rise triggers the output signal of the Schmitt trigger to turn over, resulting in the capacitor C SA The discharge circuit is conducted and the capacitor C is connected SA Voltage is reduced to form charge-discharge circulation, the circulation period is inversely proportional to the input current, and the charge-discharge period number, namely the output pulse peak number, is directly proportional to the input current in the same measurement time;
the operation method of the sense amplifier comprises the following steps:
before measurement, the enable signal is turned off, the sensing amplifier is disconnected with the front-stage storage array, and the capacitor discharge loop is conducted to be completely discharged;
when the measurement is started, the enable signal controls the transmission gate to be turned on or off, the system automatically works according to the input circuit, and the digital signal in the spike pulse form is output.
Furthermore, the TCAM cell is a FeFET-TCAM cell, and is composed of two FeFETs, wherein the drain terminals of the FeFETs are connected with VVD, and the grid electrodes of the FeFETs are respectively connected with the search lines SL and SL
Figure BDA0003610258140000021
The sources are connected with a matching line ML in common, and the matching line of each row of TCAM cells in the TCAM array is connected.
Further, the source-drain voltage of the FeFET is constant, and the output current is constant during searching; each column of TCAM cells of the TCAM array shares the same longitudinal SL and
Figure BDA0003610258140000022
each row ML is further signal processed by a subsequent stage sense amplifier.
Further, the method of operating the TCAM array includes:
before the TCAM array starts to work, binary coded data and its reverse signal are respectively passed through SL and SL by high voltage
Figure BDA0003610258140000031
Writing the array;
the search signal and its opposite signal are passed through at low voltage without pre-charging before the search is startedSL and
Figure BDA0003610258140000032
and if the currents are not matched, the ML outputs a constant current, and the output currents of all TCAM units in each row in the TCAM array are superposed to be used as the input of the rear-stage sense amplifier.
Furthermore, the TCAM cell is a MOSFET-TCAM cell, and is formed by two symmetric storage bits, each storage bit stores data through a ring inverter, the gate is connected to the word line WL to select the MOS transistor to control the connection of the bit line BL and the ring inverter, and the output of the ring inverter and the search line respectively control the discharge loop of VDD to ML through the gates of two MOS transistors connected in series.
Further, the voltage of the ML discharge loop and the circuit structure are kept stable by VDD in the searching process, and the loop outputs constant current; each row of TCAM cells in the TCAM array shares a pair of lateral bit lines BL and BL
Figure BDA0003610258140000033
Writing opposite signals, each column of TCAM cells sharing a pair of longitudinal write lines WL and
Figure BDA0003610258140000038
control the write unit to share a pair of vertical search lines SL and SL
Figure BDA0003610258140000034
The search function is realized.
Further, the method of operating the TCAM array includes:
before the TCAM array starts to work, pass through the write line WL and
Figure BDA0003610258140000035
control, the pairs of opposite signals from the bit lines BL and
Figure BDA0003610258140000036
writing into the ring inverter and storing;
the search signal and its inverse signal are passed through SL and SL respectively without need of precharging before search is started
Figure BDA0003610258140000037
If the input is not matched, ML outputs a constant current, and the output currents of all units in each row are superposed to be used as the input of the rear-stage sense amplifier.
The invention has the following beneficial effects:
1) the invention fully utilizes the matching discharge characteristic of the TCAM, realizes the design of a brand-new TCAM approximate search array and a measurement circuit, realizes the measurement of the number of search unmatched bits, and obtains the expandability of a storage array;
2) the current mode sensing amplifier can realize the conversion from an analog current signal to a digital spike pulse, thereby measuring the current;
3) for CMOS-TCAM design, ML voltage is rapidly reduced in inverse proportion to the matching degree after the existing TCAM is precharged, the difference between different mismatching degrees cannot be obtained quantitatively, the output current of the CMOS-TCAM based on a current mode is increased in equal proportion to the mismatching degree, the current corresponding to the unit mismatching degree is constant, and therefore quantitative measurement is achieved, and meanwhile the method is not limited by array scale.
4) For FeFET-TCAM design, on the basis of the beneficial effects of CMOS-TCAM, the non-volatile memory characteristic is provided, the energy consumption is lower, the area is smaller, and the memory array scale of a TCAM system can be improved.
Drawings
FIG. 1 is a schematic diagram of a conventional TCAM unit (a) and a modified TCAM unit (b) according to the present invention;
FIG. 2 is a diagram of a FeFET-TCAM cell (a) and its constituent FeFET-TCAM array (b);
FIG. 3 is a block diagram of a current-mode sense amplifier;
FIG. 4 is a block diagram of a CMOS-TCAM cell (a) and its constituent CMOS-TCAM array (b);
FIG. 5 is a simulation waveform (a) and statistical result (b) of a FeFET-TCAM design;
FIG. 6 is a simulation waveform (a) and statistical result (b) of a CMOS-TCAM design;
FIG. 7 is an expanded sawtooth simulation waveform of the current-mode sense amplifier design;
FIG. 8 is an actual pulse spike output simulation waveform for a current-mode sense amplifier design.
Detailed Description
The invention is described in further detail below with reference to the figures and specific examples.
Example 1
Referring to fig. 1-2, an approximate search TCAM matching system based on current mode measurement includes a TCAM array as a front stage circuit and a sense amplifier as a back stage circuit for receiving an analog current signal output by each row of TCAM cells in the TCAM array and converting the analog current signal into a digital pulse equal to a mismatch. The current measuring method of the TCAM unit comprises the following steps: the pre-charging mechanism of original VDD to ML is cancelled, original ML is connected with VDD and is maintained at high voltage, TCAM is output to ground to be used as new ML, and output current is used as a measurement index.
The TCAM cell of the present embodiment is preferably a FeFET-TCAM cell, and the overall structure and operation flow of the FeFET-TCAM cell and its array are as follows:
as shown in FIG. 2, each FeFET-TCAM cell is composed of 2 symmetrically connected FeFETs having drains connected to VDD and gates connected to search lines SL and SL, respectively
Figure BDA0003610258140000041
The sources are coupled to match lines ML, which are coupled to TCAM cell match lines in each row of the TCAM array. The source-drain voltage of the FeFET is constant, and the output current is constant during searching. In a TCAM array, a plurality of FeFET-TCAM units are connected in parallel to ML to form a row, and each column of FeFET-TCAM units share the same longitudinal SL and SL
Figure BDA0003610258140000042
The ML receives as an output the sense amplifier.
The operation of the entire FeFET-TCAM array is as follows:
(1) before the FeFET-TCAM array starts to work, binary coded data is written into the corresponding FeFET-TCAM, the operation scheme is shown in the following table, during writing, a reverse high-voltage signal (-4V) is added to reset the storage unit (state 0), a forward high-voltage signal (4V) is added to write the storage content into a state 1, and during searching, a reading voltage (1V) is adopted to read current, so that the storage content is prevented from being influenced.
TABLE 1 FeFET-TCAM operating protocol (voltage unit: V)
Figure BDA0003610258140000043
(2) A search phase for directly transmitting search signals from SL and
Figure BDA0003610258140000051
and (4) inputting. If the current is matched with the stored content, 2 FeFETs are all turned off, and ML has no current passing; if a mismatch occurs, then one and only one of the 2 fefets is on and ML outputs a constant current. Each column ML is connected in parallel with a plurality of memory cells, and the total output is the sum of the output currents of the cells.
The overall structure and operation flow of the current-mode sense amplifier are as follows:
as shown in FIG. 3, the sense amplifier includes a first transmission gate TG1, a second transmission gate TG2, a capacitor C SA A third transmission gate TG3, a fourth transmission gate TG4, a first diplexer 4, a schmitt trigger 1, and a second diplexer 2. By enabling signal V EN And output feedback signal V SPIKE The first transmission gate TG1 and the second transmission gate TG2 which are respectively controlled are connected in series and are used as isolation between the front-stage TCAM array and the measurement capacitor. The output current of the TCAM array passes through a first transmission gate TG1 and a second transmission gate TG2 and then is coupled with a capacitor C SA Charging so that the capacitor C SA Output voltage linearly rising, capacitor C SA The other end is grounded. By enabling signal V EN And outputs a feedback signal V SPIKE A third transmission gate TG3 and a fourth transmission gate TG4 which are respectively controlled in an inverted way are connected in parallel with the capacitor C SA Output terminal and reset signal V RESET And a quick discharge loop is formed. Capacitor C SA The output signal sequentially passes through a first phase inverter for amplitude expansion, a periodically-turned Schmitt trigger and a second phase inverter to obtain an output pulse signal V SPIKE . Capacitor C SA Schmitt initiation by voltage increaseFlip-flop output signal flips, resulting in a capacitor C SA The discharge circuit is conducted and the capacitor C is connected SA The voltage is reduced to form a charge-discharge cycle, the cycle period is inversely proportional to the input current, and the charge-discharge cycle number, namely the peak number of the output pulse is directly proportional to the input current in the same measurement time.
The operation of the whole current-mode sense amplifier is as follows:
(1) preparation phase, enabling signal V EN High, the first transfer gate TG1 closes and disconnects SA from the TCAM array, while TG4 opens to ensure that the capacitor voltage remains at V RESET 。
(2) Search phase, enabling signal V EN Set low, TG1 turns on and TG4 turns off, at which time the TCAM array is connected to the sense amplifiers. With the input of current, the capacitor C SA The voltage rises, the Schmitt trigger 1 overturns, and a pulse signal V is output SPIKE Put high, the second transmission gate TG2 is turned off, blocking the current input, and the third transmission gate TG3 is opened to make the capacitor C SA Discharge rapidly to V RESET And then the process is repeated again.
Example 2
The difference between this embodiment and embodiment 1 is that the TCAM cell of this embodiment is preferably a CMOS-TCAM cell, and the overall structure and operation flow of the CMOS-TCAM cell and its constituent array are as follows:
as shown in fig. 4, each CMOS-TCAM cell contains 2 symmetric memory bits, and 16 MOS transistors. Taking the left storage bit as an example, for the input part, MOS transistor M1 and MOS transistor M3 have drains connected to BL and BL, respectively
Figure BDA0003610258140000052
The grid electrode is connected with WL, the source electrode is respectively connected with two ends of a ring-shaped inverter, and the ring-shaped inverter is composed of two inverters; for the storage part, two inverters are annularly connected to realize a storage function; for the output part, the drain of a MOS tube M5 is connected with VDD, the grid is connected with the output of the ring-shaped inverter, the source is connected with the drain of a MOS tube M7, the grid of the MOS tube M7 is connected with SL, the source is connected with ML, and the MOS tube M5 and the MOS tube M7 form a discharge branch.
The whole CMOS-TCAM array operates as follows:
(1) and writing binary coded data into the corresponding TCAM unit before the CMOS-TCAM array starts to work. During writing, the search lines are all held at low voltage, as shown in the following table, and the first stage
Figure BDA0003610258140000061
Set high voltage, write content from BL input,
Figure BDA0003610258140000062
and
Figure BDA0003610258140000063
holding low voltage, second stage WL holding high voltage, reverse writing content from
Figure BDA0003610258140000064
The inputs, BL and WL, are held low.
TABLE 2 CMOS-TCAM write scheme (voltage unit: V)
Figure BDA0003610258140000065
(2) During the search phase, the signals are as shown in the following table, WL,
Figure BDA0003610258140000066
BL and
Figure BDA0003610258140000067
all remaining low, the search signal is from SL and
Figure BDA0003610258140000068
and (4) inputting. If the current is matched with the stored content, the 2 discharging branches are all switched off, and the ML has no current passing; if mismatch occurs, only one of the 2 discharge branches is turned on, and the ML outputs a constant current. Each column ML is connected in parallel with a plurality of memory cells, and the total output is the sum of the output currents of the cells.
TABLE 3 CMOS-TCAM search scheme
Figure BDA0003610258140000069
The function and effect of the above examples 1 and 2 are further illustrated by the following simulation experiments:
1. simulation conditions
In the experiment, for CMOS-TCAM, a 45nm model is adopted to simulate on Hspice software, for FeFET-TCAM, simulation is carried out on spectra software, and 32bit single-line TCAM storage arrays are built for the CMOS-TCAM and the FeFET-TCAM.
To verify scalability, the applicant performs multiple searches, with the search mismatch increasing bit by bit starting from 0 bit. For example, for a 4-bit memory array with a memory content of 4 'b 0000, the input search data is 4' b0000, 4 'b 0001, 4' b 0011, 4 'b 0111, 4' b 1111 in sequence. The single search time and the search interval are both 5 ns.
For the current mode sensing amplifier, the applicant also simulates on Hspice software, and directly inputs a CMOS-TCAM output current signal into a sensing amplifier circuit for experiment.
2. Simulation result
1) Verification of extensibility
1.1) FIG. 5(a) shows the result of FeFET-TCAM simulation experiment, and the output current can keep good stability within one search time. Reading currents under different mismatching degrees and performing linear regression analysis, as shown in fig. 5(b), obtaining that the currents have a good linear relation, and the current corresponding to unit bit mismatching degree is 766.8nA, so that the expandability of the array is confirmed.
1.2) FIG. 6(a) shows the results of a CMOS-TCAM simulation experiment, which once again proves the effective improvement of the expandability of the present invention. The output current per bit mismatch corresponds to 43.5uA, which is a 57-fold reduction compared to the output current of CMOS-TCAM.
2) Assessment of energy consumption and latency
The following table gives the energy consumption and delay measurements for two TCAM designs.
TABLE 4 TCAM (64 × 64) array Performance comparison
Figure BDA0003610258140000071
As can be seen from the table, the introduction of fefets results in a significant reduction in power consumption, with FeFET-TCAM search energy being 57 times less compared to current-mode CMOS-TCAM. This fully demonstrates the advantage of fefets in the memory domain and also enables the high power dissipation problem of current-mode schemes to be solved.
3) Sense amplifier functional verification
Fig. 7 and 8 show simulation experiment results of the current mode sense amplifier, and it can be observed that the number of pulse spikes corresponds to the number of mismatched bits one to one, and we confirm that the proposed scheme realizes the mismatch measurement of the current of at most 4 levels. When the mismatch is low we can see that an additional small saw-tooth charging and discharging wave pattern appears at the end of the search time, which aims to keep a certain time margin to cope with the negative effect of the non-linearity of the discharge time.
The above-described embodiments are intended to illustrate rather than to limit the invention, and any modifications and variations of the present invention are within the spirit of the invention and the scope of the appended claims.

Claims (10)

1. The approximate search TCAM matching system based on current mode measurement is characterized by comprising a TCAM array and a sense amplifier, wherein the TCAM array is used as a front-stage circuit, the sense amplifier is used as a rear-stage circuit, and the sense amplifier is used for receiving an analog current signal output by each row of TCAM units in the TCAM array and converting the analog current signal into a digital pulse with the same mismatching degree.
2. The approximate search TCAM matching system according to claim 1, wherein the current measurement method of the TCAM unit includes: the pre-charging mechanism of original VDD to ML is cancelled, original ML is connected with VDD and is maintained at high voltage, TCAM is output to ground to be used as new ML, and output current is used as a measurement index.
3. The approximate search TCAM matching system of claim 1, wherein the sense amplifier comprises an input transmission gate, a capacitor, an output transmission gate, a diplexer and a schmitt trigger electrically connected, the input current signal sequentially passes through the input transmission gate controlled by the enable control and the output signal control to charge the capacitor, the output voltage of the capacitor is output through the schmitt trigger, and the output signal is fed back to the discharging transmission gate.
4. The approximate search TCAM matching system according to claim 3, wherein there are two input transmission gates, a first transmission gate TG1 and a second transmission gate TG2, respectively, two discharge transmission gates, a third transmission gate TG3 and a fourth transmission gate TG4, respectively, and two homographs, a first homograph and a second homograph, respectively;
by enabling signal V EN And outputs a feedback signal V SPIKE The first transmission gate TG1 and the second transmission gate TG2 which are respectively controlled are connected in series and are used as a front-stage TCAM array and a measurement capacitor C SA Isolation between them;
the output current of the TCAM array passes through a first transmission gate TG1 and a second transmission gate TG2 and then is coupled with a capacitor C SA Charging so that the capacitor C SA Output voltage linearly rising, capacitor C SA The other end is grounded;
by enabling signal V EN And outputs a feedback signal V SPIKE A third transmission gate TG3 and a fourth transmission gate TG4 which are respectively controlled in an inverted way are connected in parallel with the capacitor C SA Output terminal and reset signal V RESET Form a fast discharge loop;
capacitor C SA The output signal sequentially passes through a first phase inverter for amplitude expansion, a periodically-turned Schmitt trigger and a second phase inverter to obtain an output pulse signal V SPIKE ；
Capacitor C SA The voltage rise triggers the output signal of the Schmitt trigger to turn over, resulting in the capacitor C SA The discharge circuit is conducted and the capacitor C is connected SA Voltage is reduced to form charge-discharge circulation, the circulation period is inversely proportional to the input current, and the charge-discharge period number, namely the output pulse peak number, is directly proportional to the input current in the same measurement time;
the operation method of the sense amplifier comprises the following steps:
before measurement, the enable signal is turned off, the sensing amplifier is disconnected with the front-stage storage array, and the capacitor discharge loop is conducted to be completely discharged;
when the measurement is started, the enable signal controls the transmission gate to be turned on or off, the system automatically works according to the input circuit, and the digital signal in the spike pulse form is output.
5. The approximate search TCAM matching system based on current-mode measurement as claimed in any one of claims 1-4, wherein the TCAM cell is a FeFET-TCAM cell, and is composed of two FeFETs having drain terminals commonly connected with VVD and gate electrodes connected with search lines SL and SL, respectively
Figure FDA0003610258130000021
The sources are connected with a matching line ML in common, and the matching line of each row of TCAM cells in the TCAM array is connected.
6. The approximate search TCAM matching system based on current mode measurement as claimed in claim 5, wherein FeFET source drain voltage is constant, output current is constant during search; each column of TCAM cells of the TCAM array shares the same longitudinal SL and
Figure FDA0003610258130000022
each row ML is further signal processed by a subsequent stage sense amplifier.
7. The approximate search TCAM matching system of claim 6, in which the method of operation of the TCAM array comprises:
before the TCAM array starts to work, binary coded data and its reverse signal are respectively passed through SL and SL by high voltage
Figure FDA0003610258130000023
Writing the array;
the search signal and its opposite signal are passed through SL and SL respectively at low voltage without need of pre-charging before search is started
Figure FDA0003610258130000024
And if the currents are not matched, the ML outputs a constant current, and the output currents of all TCAM units in each row in the TCAM array are superposed to be used as the input of the rear-stage sense amplifier.
8. The approximate search TCAM matching system of any one of claims 1 to 4, wherein the TCAM cell is a MOSFET-TCAM cell and is composed of two symmetric memory bits, each memory bit stores data through a ring inverter, the gate of each memory bit is connected to the word line WL to select the connection between the MOS transistor control bit line BL and the ring inverter, and the output and the search line of the ring inverter respectively control the discharging of VDD to ML through the gates of two MOS transistors in series.
9. The approximate search TCAM matching system of claim 8, in which VDD is stable to the voltage and circuit structure of ML discharge loop during search, and the loop outputs constant current; each row of TCAM cells in the TCAM array shares a pair of lateral bit lines BL and BL
Figure FDA0003610258130000029
Writing opposite signals, each column of TCAM cells sharing a pair of longitudinal write lines WL and
Figure FDA0003610258130000025
control the write unit to share a pair of vertical search lines SL and SL
Figure FDA0003610258130000026
The search function is realized.
10. The approximate search TCAM matching system of claim 9, wherein the method of operation of the TCAM array comprises:
before the TCAM array starts to work, the data is read through the write line WL and
Figure FDA0003610258130000027
control, the pairs of opposite signals from the bit lines BL and
Figure FDA00036102581300000210
writing into the ring inverter and storing;
the search signal and its inverse signal are passed through SL and SL respectively without need of precharging before search is started
Figure FDA0003610258130000028
If the input is not matched, ML outputs a constant current, and the output currents of all units in each row are superposed to be used as the input of the rear-stage sense amplifier.
CN202210427548.9A 2022-04-22 2022-04-22 Approximate search TCAM matching system based on current mode measurement Pending CN114882920A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210427548.9A CN114882920A (en) 2022-04-22 2022-04-22 Approximate search TCAM matching system based on current mode measurement

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210427548.9A CN114882920A (en) 2022-04-22 2022-04-22 Approximate search TCAM matching system based on current mode measurement

Publications (1)

Publication Number Publication Date
CN114882920A true CN114882920A (en) 2022-08-09

Family

ID=82671833

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210427548.9A Pending CN114882920A (en) 2022-04-22 2022-04-22 Approximate search TCAM matching system based on current mode measurement

Country Status (1)

Country Link
CN (1) CN114882920A (en)

Similar Documents

Publication Publication Date Title
CN110414677B (en) Memory computing circuit suitable for full-connection binarization neural network
US20220262424A1 (en) Compute in memory system
CN109979503B (en) Static random access memory circuit structure for realizing Hamming distance calculation in memory
Kazemi et al. In-memory nearest neighbor search with fefet multi-bit content-addressable memories
CN110058839A (en) A kind of circuit structure based on subtraction in Static RAM memory
CN113467751B (en) Analog domain memory internal computing array structure based on magnetic random access memory
CN115810374A (en) Memory circuit and memory computing circuit with BCAM addressing and logic operation functions
CN114496010A (en) Analog domain near memory computing array structure based on magnetic random access memory
US20040190320A1 (en) Semiconductor memory
CN113053434B (en) High-energy-efficiency TCAM based on FeFET structure and operation method thereof
US20230297235A1 (en) Sram-based cell for in-memory computing and hybrid computations/storage memory architecture
US6847566B1 (en) Method and circuit configuration for multiple charge recycling during refresh operations in a DRAM device
CN114882920A (en) Approximate search TCAM matching system based on current mode measurement
US20230377650A1 (en) Ultra-compact cam array based on single mtj and operating method thereof
CN116204490A (en) 7T memory circuit and multiply-accumulate operation circuit based on low-voltage technology
CN112597549B (en) Dynamic physical unclonable function circuit based on static random access memory
CN117316237B (en) Time domain 8T1C-SRAM memory cell and memory circuit for timing tracking quantization
US20240069780A1 (en) In-memory computing architecture for nearest neighbor search of cosine distance and operating method thereof
US20230274781A1 (en) Highly energy-efficient cam based on single fefet and operating method thereof
Kim et al. Binary content-addressable memory system using nanoelectromechanical memory switch
RU2680870C1 (en) Method for comparing data in the cell of the associative memory device, and module for comparing data in the cell of the associative memory device
US20230410862A1 (en) In-memory computation circuit using static random access memory (sram) array segmentation
US11049550B1 (en) Multi-bit current sense amplifier with pipeline current sampling of resistive memory array structure and sensing method thereof
US11875850B2 (en) Content addressable memory device, content addressable memory cell and method for data searching with a range or single-bit data
CN117275556A (en) Multi-bit content addressable memory cell, memory array and method of operation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination