TWI877798B - Memory device and operating method thereof - Google Patents
Memory device and operating method thereof Download PDFInfo
- Publication number
- TWI877798B TWI877798B TW112135761A TW112135761A TWI877798B TW I877798 B TWI877798 B TW I877798B TW 112135761 A TW112135761 A TW 112135761A TW 112135761 A TW112135761 A TW 112135761A TW I877798 B TWI877798 B TW I877798B
- Authority
- TW
- Taiwan
- Prior art keywords
- bit
- current signal
- current
- weight
- signal
- Prior art date
Links
Images
Landscapes
- Static Random-Access Memory (AREA)
- Read Only Memory (AREA)
Abstract
Description
本揭示內容是有關於一種記憶體技術,特別是關於一種記憶體裝置及記憶體裝置的操作方法。The present disclosure relates to a memory technology, and more particularly to a memory device and a method for operating the memory device.
記憶體裝置包含多個用以儲存資料位元的記憶體單元。記憶體單元可以藉由兩個電晶體的架構實施。資料位元儲存於兩個電晶體互相耦接的儲存節點。在讀取操作時,可以藉由流經記憶體單元的電流信號判斷資料位元的邏輯值。The memory device includes a plurality of memory cells for storing data bits. The memory cell can be implemented by a two-transistor architecture. The data bits are stored in a storage node where the two transistors are coupled to each other. During a read operation, the logical value of the data bit can be determined by a current signal flowing through the memory cell.
本發明實施例包含一種記憶體裝置。記憶體裝置包含第一記憶體單元。第一記憶體單元用以依據第一權重位元及第一輸入位元進行一邏輯操作。第一記憶體單元包含第一開關及第二開關。第一開關用以將第一權重位元寫入第一儲存節點。第二開關用以依據第一權重位元及第一輸入位元產生第一電流信號。第二開關的第一端用以接收攜載第一輸入位元的第一位元線信號。第二開關的第二端用以接收第一字元線信號。第二開關的控制端耦接第一儲存節點。當第一輸入位元具有第一邏輯值時,第一位元線信號及第一字元線信號的每一者具有第一電壓準位。當第一輸入位元具有第二邏輯值時,第一位元線信號具有小於第一電壓準位的第二電壓準位且第一字元線信號具有第一電壓準位。An embodiment of the present invention includes a memory device. The memory device includes a first memory unit. The first memory unit is used to perform a logic operation based on a first weight bit and a first input bit. The first memory unit includes a first switch and a second switch. The first switch is used to write the first weight bit into a first storage node. The second switch is used to generate a first current signal based on the first weight bit and the first input bit. The first end of the second switch is used to receive a first bit line signal carrying the first input bit. The second end of the second switch is used to receive a first word line signal. The control end of the second switch is coupled to the first storage node. When the first input bit has a first logic value, each of the first bit line signal and the first word line signal has a first voltage level. When the first input bit has a second logic value, the first bit line signal has a second voltage level less than the first voltage level and the first word line signal has the first voltage level.
在一些實施例中,當第一儲存節點具有第三電壓準位時,第一電流信號具有第一電流準位,當第一儲存節點第四電壓準位且第一輸入位元具有第一邏輯值時,第一電流信號具有第一電流準位,以及當第一儲存節點具有第四電壓準位且第一輸入位元具有第二邏輯值時,第一電流信號具有大於第一電流準位的第二電流準位。In some embodiments, when the first storage node has a third voltage level, the first current signal has a first current level, when the first storage node has a fourth voltage level and the first input bit has a first logic value, the first current signal has the first current level, and when the first storage node has a fourth voltage level and the first input bit has a second logic value, the first current signal has a second current level greater than the first current level.
在一些實施例中,邏輯操作是與邏輯操作、或邏輯操作、非與邏輯操作及非或邏輯操作的一者,當邏輯操作是與邏輯操作或或邏輯操作時,第四電壓準位及第二電流準位的每一者對應第二邏輯值,以及當邏輯操作是非與邏輯操作或非或邏輯操作時,第四電壓準位及第一電流準位的每一者對應第二邏輯值。In some embodiments, the logic operation is one of an AND logic operation, an OR logic operation, a NAND logic operation, and a NOR logic operation. When the logic operation is an AND logic operation or an OR logic operation, each of the fourth voltage level and the second current level corresponds to the second logic value, and when the logic operation is a NAND logic operation or a NOR logic operation, each of the fourth voltage level and the first current level corresponds to the second logic value.
在一些實施例中,記憶體裝置更包含第二記憶體單元。第二記憶體單元用以儲存第二權重位元並接收攜載第二輸入位元的第二位元線信號,以與第一記憶體單元共同進行邏輯操作。第二權重位元與第一權重位元互補,以及第二輸入位元與第一輸入位元互補。In some embodiments, the memory device further includes a second memory unit. The second memory unit is used to store a second weight bit and receive a second bit line signal carrying a second input bit to perform a logic operation with the first memory unit. The second weight bit complements the first weight bit, and the second input bit complements the first input bit.
在一些實施例中,第二記憶體單元包含第三開關及第四開關。第三開關用以將第二權重位元寫入第二儲存節點。第四開關用以依據第二權重位元及第二輸入位元產生第二電流信號。第四開關的第一端用以接收攜載第二位元線信號,第四開關的第二端用以接收第一字元線信號,以及第四開關的一控制端耦接第二儲存節點。In some embodiments, the second memory unit includes a third switch and a fourth switch. The third switch is used to write the second weight bit into the second storage node. The fourth switch is used to generate a second current signal according to the second weight bit and the second input bit. The first end of the fourth switch is used to receive the second bit line signal, the second end of the fourth switch is used to receive the first word line signal, and a control end of the fourth switch is coupled to the second storage node.
在一些實施例中,記憶體裝置更包含第二記憶體單元、第一放大器、第三記憶體單元、第四記憶體單元及第二放大器。第二記憶體單元用以儲存第二權重位元並接收攜載第二輸入位元的第二位元線信號,以產生第二電流信號。第一放大器耦接第一節點,並用以依據第三電流信號產生第一電壓信號。第三記憶體單元用以儲存第三權重位元並接收第一位元線信號,以產生第四電流信號。第四記憶體單元用以儲存第四權重位元並接收第二位元線信號,以產生第五電流信號。第二放大器耦接第二節點,並用以依據第六電流信號產生第二電壓信號。第二記憶體單元及第一記憶體單元用以將第二電流信號及第一電流信號於第一節點相加以產生第三電流信號,以及第三記憶體單元及第四記憶體單元用以將第四電流信號及第五電流信號於第二節點相加以產生第六電流信號。In some embodiments, the memory device further includes a second memory unit, a first amplifier, a third memory unit, a fourth memory unit, and a second amplifier. The second memory unit is used to store a second weight bit and receive a second bit line signal carrying a second input bit to generate a second current signal. The first amplifier is coupled to the first node and is used to generate a first voltage signal according to the third current signal. The third memory unit is used to store a third weight bit and receive a first bit line signal to generate a fourth current signal. The fourth memory unit is used to store a fourth weight bit and receive a second bit line signal to generate a fifth current signal. The second amplifier is coupled to the second node and is used to generate a second voltage signal according to the sixth current signal. The second memory unit and the first memory unit are used to add the second current signal and the first current signal at the first node to generate a third current signal, and the third memory unit and the fourth memory unit are used to add the fourth current signal and the fifth current signal at the second node to generate a sixth current signal.
在一些實施例中,記憶體裝置更包含第二記憶體單元、第三記憶體單元、第四記憶體單元及第一放大器。第二記憶體單元用以儲存第二權重位元並接收攜載第二輸入位元的第二位元線信號,以產生第二電流信號。第三記憶體單元用以儲存第三權重位元並接收攜載第三輸入位元的第三位元線信號,以產生第三電流信號。第四記憶體單元用以儲存第四權重位元並接收攜載第四輸入位元的第四位元線信號,以產生第四電流信號。第一放大器用以接收第五電流信號於第一節點。第一記憶體單元、第二記憶體單元、第三記憶體單元及第四記憶體單元用以將第一電流信號、第二電流信號、第三電流信號及第四電流信號在第一節點相加以產生第五電流信號,以及第一權重位元、第二權重位元、第一輸入位元及第二輸入位元分別互補於第三權重位元、第四權重位元、第三輸入位元及第四輸入位元。In some embodiments, the memory device further includes a second memory unit, a third memory unit, a fourth memory unit, and a first amplifier. The second memory unit is used to store a second weight bit and receive a second bit line signal carrying a second input bit to generate a second current signal. The third memory unit is used to store a third weight bit and receive a third bit line signal carrying a third input bit to generate a third current signal. The fourth memory unit is used to store a fourth weight bit and receive a fourth bit line signal carrying a fourth input bit to generate a fourth current signal. The first amplifier is used to receive a fifth current signal at the first node. The first memory unit, the second memory unit, the third memory unit and the fourth memory unit are used to add the first current signal, the second current signal, the third current signal and the fourth current signal at the first node to generate a fifth current signal, and the first weight bit, the second weight bit, the first input bit and the second input bit complement the third weight bit, the fourth weight bit, the third input bit and the fourth input bit respectively.
在一些實施例中,記憶體裝置更包含第二記憶體單元、第三記憶體單元、第四記憶體單元、第一放大器及第二放大器。第二記憶體單元用以儲存第二權重位元並接收攜載第二輸入位元的第二位元線信號,以產生第二電流信號。第三記憶體單元用以儲存第三權重位元並接收攜載第三輸入位元的第三位元線信號,以產生第三電流信號。第四記憶體單元用以儲存第四權重位元並接收攜載第四輸入位元的第四位元線信號,以產生第四電流信號。第一放大器用以接收第五電流信號於第一節點。第二放大器用以接收第六電流信號第二節點。第一記憶體單元及第二記憶體單元用以將第一電流信號及第二電流信號在第一節點相加以產生第五電流信號,第三記憶體單元及第四記憶體單元用以將第三電流信號及第四電流信號在第二節點相加以產生第六電流信號,以及第一權重位元、第二權重位元、第一輸入位元及第二輸入位元分別互補於第三權重位元、第四權重位元、第三輸入位元及第四輸入位元。In some embodiments, the memory device further includes a second memory unit, a third memory unit, a fourth memory unit, a first amplifier, and a second amplifier. The second memory unit is used to store a second weight bit and receive a second bit line signal carrying a second input bit to generate a second current signal. The third memory unit is used to store a third weight bit and receive a third bit line signal carrying a third input bit to generate a third current signal. The fourth memory unit is used to store a fourth weight bit and receive a fourth bit line signal carrying a fourth input bit to generate a fourth current signal. The first amplifier is used to receive a fifth current signal at the first node. The second amplifier is used to receive a sixth current signal at the second node. The first memory unit and the second memory unit are used to add the first current signal and the second current signal at the first node to generate a fifth current signal, the third memory unit and the fourth memory unit are used to add the third current signal and the fourth current signal at the second node to generate a sixth current signal, and the first weight bit, the second weight bit, the first input bit and the second input bit complement the third weight bit, the fourth weight bit, the third input bit and the fourth input bit respectively.
本發明實施例包含一種記憶體裝置的操作方法。操作方法包含:將第一權重位元寫入第一記憶體單元的第一儲存節點;藉由第一記憶體單元接收攜載第一輸入位元的第一位元線信號;依據第一輸入位元及第一權重位元進行一邏輯操作,以產生通過第一記憶體單元中的第一開關的第一電流信號;當第一儲存節點具有第一電壓準位時,關閉第一開關且第一電流信號具有第一電流準位;當第一儲存節點具有第二電壓準位且第一位元線信號具有第三電壓準位時,導通第一開關且第一電流信號具有第一電流準位;以及當第一儲存節點具有第二電壓準位且第一位元線信號具有第四電壓準位時,導通第一開關且第一電流信號具有第二電流準位。第二電流準位大於第一電流準位,以及第三電壓準位大於第四電壓準位。An embodiment of the present invention includes an operating method of a memory device. The operation method includes: writing a first weight bit into a first storage node of a first memory unit; receiving a first bit line signal carrying a first input bit by the first memory unit; performing a logic operation according to the first input bit and the first weight bit to generate a first current signal passing through a first switch in the first memory unit; when the first storage node has a first voltage level, closing the first switch and the first current signal has a first current level; when the first storage node has a second voltage level and the first bit line signal has a third voltage level, turning on the first switch and the first current signal has a first current level; and when the first storage node has a second voltage level and the first bit line signal has a fourth voltage level, turning on the first switch and the first current signal has a second current level. The second current level is greater than the first current level, and the third voltage level is greater than the fourth voltage level.
在一些實施例中,操作方法更包含:反相第一輸入位元以產生第二輸入位元;反相第一權重位元以產生第二權重位元;將第二權重位元寫入第二記憶體單元的第二儲存節點;藉由第二記憶體單元接收攜載第二輸入位元的第二位元線信號;依據第二輸入位元及第二權重位元產生通過第二記憶體單元中的第二開關的第二電流信號;以及將第一電流信號及第二電流信號相加以產生對應邏輯操作的輸出的第三電流信號。In some embodiments, the operating method further includes: inverting the first input bit to generate a second input bit; inverting the first weight bit to generate a second weight bit; writing the second weight bit into a second storage node of a second memory unit; receiving a second bit line signal carrying the second input bit by the second memory unit; generating a second current signal through a second switch in the second memory unit based on the second input bit and the second weight bit; and adding the first current signal and the second current signal to generate a third current signal corresponding to the output of the logic operation.
於本文中,當一元件被稱為「連接」或「耦接」時,可指「電性連接」或「電性耦接」。「連接」或「耦接」亦可用以表示二或多個元件間相互搭配操作或互動。此外,雖然本文中使用「第一」、「第二」、…等用語描述不同元件,該用語僅是用以區別以相同技術用語描述的元件或操作。除非上下文清楚指明,否則該用語並非特別指稱或暗示次序或順位,亦非用以限定本案。In this article, when an element is referred to as "connected" or "coupled", it may refer to "electrically connected" or "electrically coupled". "Connected" or "coupled" may also be used to indicate that two or more elements cooperate with each other or interact with each other. In addition, although the terms "first", "second", etc. are used in this article to describe different elements, the terms are only used to distinguish between elements or operations described with the same technical terms. Unless the context clearly indicates otherwise, the terms do not specifically refer to or imply an order or sequence, nor are they used to limit the present case.
除非另有定義,本文使用的所有術語(包括技術和科學術語)具有與本案所屬領域的普通技術人員通常理解的相同的含義。將進一步理解的是,諸如在通常使用的字典中定義的那些術語應當被解釋為具有與它們在相關技術和本案的上下文中的含義一致的含義,並且將不被解釋為理想化的或過度正式的意義,除非本文中明確地這樣定義。Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by ordinary technicians in the field to which this case belongs. It will be further understood that those terms as defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology and this case, and will not be interpreted as an idealized or overly formal meaning unless expressly defined as such in this document.
這裡使用的術語僅僅是為了描述特定實施例的目的,而不是限制性的。如本文所使用的,除非內容清楚地指示,否則單數形式「一」、「一個」和「該」旨在包括複數形式,包括「至少一個」。「或」表示「及/或」。如本文所使用的,術語「及/或」包括一個或多個相關所列項目的任何和所有組合。還應當理解,當在本說明書中使用時,術語「包括」及/或「包含」指定所述特徵、區域、整體、步驟、操作、元件的存在及/或部件,但不排除一個或多個其它特徵、區域整體、步驟、操作、元件、部件及/或其組合的存在或添加。The terms used herein are for the purpose of describing specific embodiments only and are not restrictive. As used herein, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms, including "at least one". "Or" means "and/or". As used herein, the term "and/or" includes any and all combinations of one or more of the relevant listed items. It should also be understood that when used in this specification, the terms "include" and/or "comprise" specify the presence and/or parts of the features, regions, wholes, steps, operations, elements, but do not exclude the presence or addition of one or more other features, regions, wholes, steps, operations, elements, parts and/or their combinations.
以下將以圖式揭露本案之複數個實施方式,為明確說明起見,許多實務上的細節將在以下敘述中一併說明。然而,應瞭解到,這些實務上的細節不應用以限制本案。也就是說,在本揭示內容部分實施方式中,這些實務上的細節是非必要的。此外,為簡化圖式起見,一些習知慣用的結構與元件在圖式中將以簡單示意的方式繪示之。The following will disclose multiple implementations of the present invention with drawings. For the purpose of clarity, many practical details will be described together in the following description. However, it should be understood that these practical details should not be used to limit the present invention. In other words, in some implementations of the disclosed content, these practical details are not necessary. In addition, in order to simplify the drawings, some commonly used structures and components will be depicted in the drawings in a simple schematic manner.
第1圖為根據本案之一實施例所繪示之記憶體裝置100的示意圖。如第1圖所示,記憶體裝置100包含記憶體單元110。在一些實施例中,記憶體裝置100更包含以多個行及多個列排列而成的多個記憶體單元(圖中未示出)。FIG. 1 is a schematic diagram of a
在一些實施例中,記憶體單元110用以接收字元線信號WWL1、RWL1及位元線信號WBL1、RBL1,並依據位元線信號WBL1攜載的權重位元W11及位元線信號RBL1攜載的輸入位元IPT1進行邏輯操作以產生電流信號IS1。權重位元W11及輸入位元IPT1對應邏輯操作的輸入,且電流信號IS1的電流準位對應邏輯操作的輸出。In some embodiments, the
在一些實施例中,位元線信號WBL1的不同電壓準位對應權重位元W11的不同邏輯值,且位元線信號RBL11的不同電壓準位對應輸入位元IPT1的不同邏輯值。In some embodiments, different voltage levels of the bit line signal WBL1 correspond to different logical values of the weight bit W11, and different voltage levels of the bit line signal RBL11 correspond to different logical values of the input bit IPT1.
舉例來說,當位元線信號WBL1具有電壓準位VWL時,權重位元W11具有邏輯值0及1的一者,且當位元線信號WBL1具有電壓準位VWH時,權重位元W11具有邏輯值0及1的另一者。類似地,當位元線信號RBL1具有電壓準位VL時,輸入位元IPT1具有邏輯值0及1的一者,且當位元線信號RBL1具有電壓準位VH時,輸入位元IPT1具有邏輯值0及1的另一者。在一些實施例中,電壓準位VWH大於電壓準位VWL,且電壓準位VH大於電壓準位VL。For example, when the bit line signal WBL1 has a voltage level VWL, the weight bit W11 has one of logical values 0 and 1, and when the bit line signal WBL1 has a voltage level VWH, the weight bit W11 has the other of logical values 0 and 1. Similarly, when the bit line signal RBL1 has a voltage level VL, the input bit IPT1 has one of logical values 0 and 1, and when the bit line signal RBL1 has a voltage level VH, the input bit IPT1 has the other of logical values 0 and 1. In some embodiments, the voltage level VWH is greater than the voltage level VWL, and the voltage level VH is greater than the voltage level VL.
如第1圖所示,記憶體單元110包含開關T11及T12。開關T11的一端用以接收位元線信號WBL1,開關T11的另一端耦接儲存節點N11,且開關T11的控制端用以接收字元線信號WWL1。開關T12的一端用以接收字元線信號RWL1,開關T12的另一端用以接收位元線信號RBL1,且開關T12的控制端耦接儲存節點N11。As shown in FIG. 1 , the
在一些實施例中,開關T11及T12可以藉由N型金屬氧化半導體(NMOS)電晶體、浮接閘極(FG)電晶體、矽-氧化物-氮化物-氧化物-矽(SONOS)電晶體或氧化銦鎵鋅(IGZO)電晶體實施。In some embodiments, the switches T11 and T12 may be implemented by N-type metal oxide semiconductor (NMOS) transistors, floating gate (FG) transistors, silicon-oxide-nitride-oxide-silicon (SONOS) transistors, or indium gallium zinc oxide (IGZO) transistors.
在一些實施例中,記憶體單元110可以更包含電容C11。電容C11的一端耦接儲存節點N11,且電容C11的另一端接地。在各種實施例中,電容C11可以藉由一個或多個電容實施,也可以藉由儲存節點N11的寄生電容實施。In some embodiments, the
操作上,記憶體單元110可以進行寫入操作以將權重位元W11寫入儲存節點N11,並可以在寫入操作之後進行讀取操作以依據權重位元W11及輸入位元IPT1產生流經開關T12的電流信號IS1。In operation, the
在寫入操作時,字元線信號WWL1具有致能電壓準位,使得開關T11導通。此時,位元線信號WBL1將權重位元W11通過開關T11寫入儲存節點N11。在寫入操作結束時,電容C11可以儲存權重位元W11對應的電壓準位。During the write operation, the word line signal WWL1 has an enable voltage level, which turns on the switch T11. At this time, the bit line signal WBL1 writes the weight bit W11 into the storage node N11 through the switch T11. When the write operation is completed, the capacitor C11 can store the voltage level corresponding to the weight bit W11.
在讀取操作時,字元線信號WWL1具有禁能電壓準位,使得開關T11關閉。開關T12依據儲存節點N11的電壓準位(亦即,權重位元W11的邏輯值)導通或關閉。當儲存節點N11具有電壓準位VWL時,開關T12關閉,使得電流信號IS1具有電流準位ILL。在一些實施例中,電流準位ILL大約等於零。當儲存節點N11具有電壓準位VWH時,開關T12導通。During a read operation, the word line signal WWL1 has a disable voltage level, so that the switch T11 is closed. The switch T12 is turned on or off according to the voltage level of the storage node N11 (i.e., the logical value of the weight bit W11). When the storage node N11 has a voltage level VWL, the switch T12 is closed, so that the current signal IS1 has a current level ILL. In some embodiments, the current level ILL is approximately equal to zero. When the storage node N11 has a voltage level VWH, the switch T12 is turned on.
另外,在讀取操作時,字元線信號RWL1具有電壓準位VH。對應地,在開關T12導通且位元線信號RBL1具有電壓準位VH時,開關T12兩端的電壓準位相同,使得電流信號IS1具有電流準位ILL。在開關T12導通且位元線信號RBL1具有電壓準位VL時,電流信號IS1具有大於電流準位ILL的電流準位ILH。在一些實施例中,具有電流準位ILL的電流信號IS1對應邏輯值0及1的一者,且具有電流準位ILH的電流信號IS1對應邏輯值0及1的另一者。In addition, during the read operation, the word line signal RWL1 has a voltage level VH. Correspondingly, when the switch T12 is turned on and the bit line signal RBL1 has a voltage level VH, the voltage levels at both ends of the switch T12 are the same, so that the current signal IS1 has a current level ILL. When the switch T12 is turned on and the bit line signal RBL1 has a voltage level VL, the current signal IS1 has a current level ILH greater than the current level ILL. In some embodiments, the current signal IS1 having the current level ILL corresponds to one of the logical values 0 and 1, and the current signal IS1 having the current level ILH corresponds to the other of the logical values 0 and 1.
在一些實施例中,記憶體單元110用以進行與(AND)邏輯操作。在上述實施例中,位元線信號WBL1的電壓準位VWL及VWH分別對應邏輯值0及1,位元線信號RBL1的電壓準位VL及VH分別對應邏輯值1及0,且電流信號IS1的電流準位ILL及ILH分別對應邏輯值0及1。In some embodiments, the
在一些實施例中,記憶體單元110用以進行或(OR)邏輯操作。在上述實施例中,位元線信號WBL1的電壓準位VWL及VWH分別對應邏輯值1及0,位元線信號RBL1的電壓準位VL及VH分別對應邏輯值0及1,且電流信號IS1的電流準位ILL及ILH分別對應邏輯值1及0。In some embodiments, the
在一些實施例中,記憶體單元110用以進行非與(NAND)邏輯操作。在上述實施例中,位元線信號WBL1的電壓準位VWL及VWH分別對應邏輯值0及1,位元線信號RBL1的電壓準位VL及VH分別對應邏輯值1及0,且電流信號IS1的電流準位ILL及ILH分別對應邏輯值1及0。In some embodiments, the
在一些實施例中,記憶體單元110用以進行非或(NOR)邏輯操作。在上述實施例中,位元線信號WBL1的電壓準位VWL及VWH分別對應邏輯值1及0,位元線信號RBL1電壓準位VL及VH分別對應邏輯值0及1,且電流信號IS1的電流準位ILL及ILH分別對應邏輯值0及1。In some embodiments, the
在一些作法中,記憶體單元只用於儲存資料位元,並沒有用於進行邏輯操作。In some implementations, memory cells are used only to store data bits and are not used to perform logical operations.
相較於上述作法,在本揭示內容的實施例中,記憶體單元110可以將輸入位元IPT1及權重位元W11作為邏輯操作的輸入,並產生電流信號IS1作為邏輯操作的輸出。如此一來,可以藉由記憶體單元110進行及、或、非與、非或等不同的邏輯操作。Compared to the above, in the embodiment of the present disclosure, the
第2圖為根據本案之一實施例所繪示之記憶體裝置200的示意圖。請參照第2圖及第1圖,記憶體裝置200係記憶體裝置100的一種變化例。記憶體裝置200的部份元件沿用記憶體裝置100的標號方式。為簡潔起見,討論將集中在記憶體裝置200不同於記憶體裝置100的部份而非相同之處。FIG. 2 is a schematic diagram of a
相較於記憶體裝置100,記憶體裝置200更包含記憶體單元210。記憶體單元210用以接收字元線信號WWL1、RWL1及位元線信號WBL2、RBL2,並依據位元線信號WBL2攜載的權重位元WB11及位元線信號RBL2攜載的輸入位元IPTB1進行邏輯操作以產生電流信號IS2。記憶體裝置200更用以電流信號IS2及IS1在節點N22相加以產生電流信號IM2。Compared to the
在一些實施例中,權重位元WB11與權重位元W11互補,且輸入位元IPTB1與輸入位元IPT1互補。舉例來說,當權重位元W11具有邏輯值0及1的一者時,權重位元WB11具有邏輯值0及1的另一者。當輸入位元IPT1具有邏輯值0及1的一者時,輸入位元IPTB1具有邏輯值0及1的另一者。In some embodiments, weight bit WB11 and weight bit W11 complement each other, and input bit IPTB1 and input bit IPT1 complement each other. For example, when weight bit W11 has one of logical values 0 and 1, weight bit WB11 has the other of logical values 0 and 1. When input bit IPT1 has one of logical values 0 and 1, input bit IPTB1 has the other of logical values 0 and 1.
權重位元WB11及輸入位元IPTB1的邏輯值與位元線信號WBL2及RBL2的電壓準位之間的關係類似於權重位元W11及輸入位元IPT1的邏輯值與位元線信號WBL1及RBL1的電壓準位之間的關係,因此,為簡潔起見,部分敘述不再重複說明。The relationship between the logic value of the weight bit WB11 and the input bit IPTB1 and the voltage level of the bit line signals WBL2 and RBL2 is similar to the relationship between the logic value of the weight bit W11 and the input bit IPT1 and the voltage level of the bit line signals WBL1 and RBL1, so for the sake of brevity, some descriptions will not be repeated.
在一些實施例中,記憶體裝置200更包含反相器NV21及NV22。反相器NV21用以接收位元線信號RBL1並輸出位元線信號RBL2。反相器NV22用以接收位元線信號WBL1並輸出位元線信號WBL2。換言之,反相器NV21用以反相輸入位元IPT1以產生輸入位元IPTB1,且反相器NV22用以反相權重位元W11以產生權重位元WB11。In some embodiments, the
如第2圖所示,記憶體單元210包含開關T21及T22。開關T21的一端用以接收位元線信號WBL2,開關T21的另一端耦接儲存節點N21,且開關T21的控制端用以接收字元線信號WWL1。開關T22的一端用以接收字元線信號RWL1,開關T22的另一端用以接收位元線信號RBL2,且開關T22的控制端耦接儲存節點N21。As shown in FIG. 2 , the
在一些實施例中,記憶體單元210可以更包含電容C21。電容C21的一端耦接儲存節點N21,且電容C21的另一端接地。在各種實施例中,電容C21可以藉由一個或多個電容實施,也可以藉由儲存節點N21的寄生電容實施。In some embodiments, the
操作上,記憶體單元210可以進行寫入操作以將權重位元WB11寫入節點N21,並可以在寫入操作之後進行讀取操作以依據權重位元WB11及輸入位元IPTB1產生流經開關T22的電流信號IS2。In operation, the
在寫入操作時,字元線信號WWL1具有致能電壓準位,使得開關T21導通。此時,位元線信號WBL2將權重位元WB11通過開關T21寫入儲存節點N21。在寫入操作結束時,電容C21可以儲存權重位元WB11對應的電壓準位。During the write operation, the word line signal WWL1 has an enabling voltage level, which turns on the switch T21. At this time, the bit line signal WBL2 writes the weight bit WB11 into the storage node N21 through the switch T21. When the write operation is completed, the capacitor C21 can store the voltage level corresponding to the weight bit WB11.
在讀取操作時,字元線信號WWL1具有禁能電壓準位,使得開關T21關閉。開關T22依據儲存節點N21的電壓準位導通或關閉。當儲存節點N21具有電壓準位VWL時,開關T22關閉,使得電流信號IS2具有電流準位ILL。當儲存節點N21具有電壓準位VWH時,開關T22導通。During the read operation, the word line signal WWL1 has a disable voltage level, so that the switch T21 is closed. The switch T22 is turned on or off according to the voltage level of the storage node N21. When the storage node N21 has a voltage level VWL, the switch T22 is closed, so that the current signal IS2 has a current level ILL. When the storage node N21 has a voltage level VWH, the switch T22 is turned on.
另外,在讀取操作時,字元線信號RWL1具有電壓準位VH。對應地,在開關T22導通且位元線信號RBL2具有電壓準位VH時,開關T22兩端的電壓準位相同,使得電流信號IS2具有電流準位ILL。在開關T22導通且位元線信號RBL2具有電壓準位VL時,電流信號IS2具有電流準位ILH。在一些實施例中,具有電流準位ILL的電流信號IS2對應邏輯值0及1的一者,且具有電流準位ILH的電流信號IS2對應邏輯值0及1的另一者。In addition, during the read operation, the word line signal RWL1 has a voltage level VH. Correspondingly, when the switch T22 is turned on and the bit line signal RBL2 has a voltage level VH, the voltage levels at both ends of the switch T22 are the same, so that the current signal IS2 has a current level ILL. When the switch T22 is turned on and the bit line signal RBL2 has a voltage level VL, the current signal IS2 has a current level ILH. In some embodiments, the current signal IS2 having the current level ILL corresponds to one of the logical values 0 and 1, and the current signal IS2 having the current level ILH corresponds to the other of the logical values 0 and 1.
由於權重位元WB11及W11彼此互補,且輸入位元IPTB1及IPT1彼此互補,位元線信號RBL2、WBL2的電壓準位與位元線信號RBL1、WBL1的電壓準位相關聯。Since the weight bits WB11 and W11 complement each other, and the input bits IPTB1 and IPT1 complement each other, the voltage levels of the bit line signals RBL2 and WBL2 are related to the voltage levels of the bit line signals RBL1 and WBL1.
舉例來說,當位元線信號RBL1及WBL1分別具有電壓準位VL及VWH時,位元線信號RBL2及WBL2分別具有電壓準位VH及VWL。對應地,電流信號IS1及IS2分別具有電流準位ILH及ILL,使得電流信號IM2具有電流準位ILH。For example, when the bit line signals RBL1 and WBL1 have voltage levels VL and VWH, respectively, the bit line signals RBL2 and WBL2 have voltage levels VH and VWL, respectively. Correspondingly, the current signals IS1 and IS2 have current levels ILH and ILL, respectively, so that the current signal IM2 has a current level ILH.
類似地,當位元線信號RBL1及WBL1分別具有電壓準位VH及VWL時,位元線信號RBL2及WBL2分別具有電壓準位VL及VWH。對應地,電流信號IS1及IS2分別具有電流準位ILL及ILH,使得電流信號IM2具有電流準位ILH。Similarly, when the bit line signals RBL1 and WBL1 have voltage levels VH and VWL, respectively, the bit line signals RBL2 and WBL2 have voltage levels VL and VWH, respectively. Correspondingly, the current signals IS1 and IS2 have current levels ILL and ILH, respectively, so that the current signal IM2 has a current level ILH.
類似地,當位元線信號RBL1及WBL1分別具有電壓準位VL及VWL時,位元線信號RBL2及WBL2分別具有電壓準位VH及VWH。對應地,電流信號IS1及IS2的每一者具有電流準位ILL,使得電流信號IM2具有電流準位ILL。Similarly, when the bit line signals RBL1 and WBL1 have voltage levels VL and VWL, respectively, the bit line signals RBL2 and WBL2 have voltage levels VH and VWH, respectively. Correspondingly, each of the current signals IS1 and IS2 has a current level ILL, so that the current signal IM2 has a current level ILL.
類似地,當位元線信號RBL1及WBL1分別具有電壓準位VH及VWH時,位元線信號RBL2及WBL2分別具有電壓準位VL及VWL。對應地,電流信號IS1及IS2的每一者具有電流準位ILL,使得電流信號IM2具有電流準位ILL。Similarly, when the bit line signals RBL1 and WBL1 have voltage levels VH and VWH, respectively, the bit line signals RBL2 and WBL2 have voltage levels VL and VWL, respectively. Correspondingly, each of the current signals IS1 and IS2 has a current level ILL, so that the current signal IM2 has a current level ILL.
在一些實施例中,記憶體單元110及210用以共同進行邏輯操作,並且將電流信號IM2作為邏輯操作的輸出。具有電流準位ILL的電流信號IM2對應邏輯值0及1的一者,且具有電流準位ILH的電流信號IM2對應邏輯值0及1的另一者。In some embodiments, the
在一些實施例中,記憶體裝置200用以進行互斥或(XOR)邏輯操作。在上述實施例中,位元線信號WBL1的電壓準位VWL及VWH分別對應邏輯值0及1,位元線信號RBL1電壓準位VL及VH分別對應邏輯值1及0,且電流信號IM2的電流準位ILL及ILH分別對應邏輯值0及1。In some embodiments, the
在一些實施例中,記憶體裝置200用以進行非互斥或(XNOR)邏輯操作。在上述實施例中,位元線信號WBL1的電壓準位VWL及VWH分別對應邏輯值0及1,位元線信號RBL1電壓準位VL及VH分別對應邏輯值1及0,且電流信號IM2的電流準位ILL及ILH分別對應邏輯值1及0。In some embodiments, the
在一些作法中,記憶體單元只用於儲存資料位元,並沒有用於進行邏輯操作。In some implementations, memory cells are used only to store data bits and are not used to perform logical operations.
相較於上述作法,在本揭示內容的實施例中,記憶體裝置200可以將輸入位元IPT1、權重位元W11及互補的輸入位元IPTB1、權重位元WB11作為邏輯操作的輸入,並產生電流信號IM2作為邏輯操作的輸出。如此一來,可以藉由記憶體裝置200進行互斥或、非互斥或等不同的邏輯操作。Compared to the above method, in the embodiment of the present disclosure, the
第3圖為根據本案之一實施例所繪示之記憶體裝置300的示意圖。請參照第3圖及第1圖,記憶體裝置300係記憶體裝置100的一種變化例。記憶體裝置300的部份元件沿用記憶體裝置100的標號方式。為簡潔起見,討論將集中在記憶體裝置300不同於記憶體裝置100的部份而非相同之處。FIG. 3 is a schematic diagram of a
相較於記憶體裝置100,記憶體裝置300更包含電阻R3。電阻R3的一端用以接收字元線信號RWL1,且電阻R3的另一端耦接開關T12於節點N31。換言之,開關T12通過電阻R3接收字元線信號RWL1。Compared to the
在一些實施例中,電阻R3用以降低電流信號IS1,以避免過大的電流信號IS1造成電路運作異常。在各種實施例中,電阻R3的電阻值大約在10 5到10 9歐姆的範圍內,且電阻R3可以藉由氧化層或多晶矽電阻實施。 In some embodiments, the resistor R3 is used to reduce the current signal IS1 to prevent the excessive current signal IS1 from causing abnormal circuit operation. In various embodiments, the resistance value of the resistor R3 is approximately in the range of 10 5 to 10 9 ohms, and the resistor R3 can be implemented by an oxide layer or a polysilicon resistor.
在一些實施例中,開關T12具有臨界電壓準位VTH,且電壓準位VWH減去電壓準位VL大於臨界電壓準位VTH。當儲存節點N11的電壓準位減去位元線信號RBL1的電壓準位大於臨界電壓準位VTH時,電流信號IS1具有電流準位ILH。反之,當儲存節點N11的電壓準位減去位元線信號RBL1的電壓準位小於臨界電壓準位VTH時,電流信號IS1具有電流準位ILL。In some embodiments, the switch T12 has a critical voltage level VTH, and the voltage level VWH minus the voltage level VL is greater than the critical voltage level VTH. When the voltage level of the storage node N11 minus the voltage level of the bit line signal RBL1 is greater than the critical voltage level VTH, the current signal IS1 has a current level ILH. Conversely, when the voltage level of the storage node N11 minus the voltage level of the bit line signal RBL1 is less than the critical voltage level VTH, the current signal IS1 has a current level ILL.
具體來說,當位元線信號RBL1具有電壓準位VL且儲存節點N11具有電壓準位VWH時,電流信號IS1具有電流準位ILH。當位元線信號RBL1具有電壓準位VH且儲存節點N11具有電壓準位VWH時,電流信號IS1具有電流準位ILL。當位元線信號RBL1具有電壓準位VH且儲存節點N11具有電壓準位VWL時,電流信號IS1具有電流準位ILL。當位元線信號RBL1具有電壓準位VL且儲存節點N11具有電壓準位VWL時,電流信號IS1具有電流準位ILL。Specifically, when the bit line signal RBL1 has a voltage level VL and the storage node N11 has a voltage level VWH, the current signal IS1 has a current level ILH. When the bit line signal RBL1 has a voltage level VH and the storage node N11 has a voltage level VWH, the current signal IS1 has a current level ILL. When the bit line signal RBL1 has a voltage level VH and the storage node N11 has a voltage level VWL, the current signal IS1 has a current level ILL. When the bit line signal RBL1 has a voltage level VL and the storage node N11 has a voltage level VWL, the current signal IS1 has a current level ILL.
在一些實施例中,位元線信號RBL1的電壓準位VL及VH分別對應輸入位元IPT1的邏輯值1及0,儲存節點N11的電壓準位VWL及VWH分別對應權重位元W11的邏輯值0及1,且電流信號IS1的電流準位ILL及ILH分別邏輯值0及1。In some embodiments, the voltage levels VL and VH of the bit line signal RBL1 correspond to the logic values 1 and 0 of the input bit IPT1, respectively, the voltage levels VWL and VWH of the storage node N11 correspond to the logic values 0 and 1 of the weight bit W11, respectively, and the current levels ILL and ILH of the current signal IS1 have logic values 0 and 1, respectively.
第4圖為根據本案之一實施例所繪示之記憶體裝置400的示意圖。請參照第4圖及第3圖,記憶體裝置400係記憶體裝置300的一種變化例。記憶體裝置400的部份元件沿用記憶體裝置300的標號方式。為簡潔起見,討論將集中在記憶體裝置400不同於記憶體裝置300的部份而非相同之處。FIG. 4 is a schematic diagram of a
相較於記憶體裝置300,記憶體裝置400更包含記憶體單元411~417及放大器451、452。在一些實施例中,記憶體單元110及411~413用以進行第一乘積累加(multiply and accumulate,MAC)邏輯操作,以產生電流信號IM41作為第一乘積累加操作的輸出。記憶體單元414~417用以進行第二乘積累加運算,以產生電流信號IM42作為第二乘積累加操作的輸出。放大器451用以依據電流信號IM41產生電壓信號S41。放大器452用以依據電流信號IM42產生電壓信號S42。在一些實施例中,放大器451及452可以藉由感測放大器(sensing amplifier,SA)實施。Compared to the
如第4圖所示,記憶體單元411包含開關T41、T42及電阻R41。開關T41的一端用以接收位元線信號WBL2,開關T41的另一端耦接儲存節點N41,且開關T41的控制端用以接收字元線信號WWL1。開關T42的一端耦接電阻R41,開關T42的另一端用以接收位元線信號RBL2,且開關T42的控制端耦接儲存節點N41。電阻R41的一端耦接開關T42,且電阻R41的另一端用以接收字元線信號RWL1。As shown in FIG. 4 , the
類似地,記憶體單元412包含開關T43、T44及電阻R42。開關T43的一端用以接收位元線信號WBL3,開關T43的另一端耦接儲存節點N42,且開關T43的控制端用以接收字元線信號WWL1。開關T44的一端耦接電阻R42,開關T44的另一端用以接收位元線信號RBL3,且開關T44的控制端耦接儲存節點N42。電阻R42的一端耦接開關T44,且電阻R42的另一端用以接收字元線信號RWL1。Similarly, the
類似地,記憶體單元413包含開關T45、T46及電阻R43。開關T45的一端用以接收位元線信號WBL4,開關T45的另一端耦接儲存節點N43,且開關T45的控制端用以接收字元線信號WWL1。開關T46的一端耦接電阻R43,開關T46的另一端用以接收位元線信號RBL4,且開關T46的控制端耦接儲存節點N43。電阻R43的一端耦接開關T46,且電阻R43的另一端用以接收字元線信號RWL1。Similarly, the
類似地,記憶體單元414包含開關T47、T48及電阻R44。開關T47的一端用以接收位元線信號WBL1,開關T47的另一端耦接儲存節點N44,且開關T47的控制端用以接收字元線信號WWL2。開關T48的一端耦接電阻R44,開關T48的另一端用以接收位元線信號RBL1,且開關T48的控制端耦接儲存節點N44。電阻R44的一端耦接開關T48,且電阻R44的另一端用以接收字元線信號RWL2。Similarly, the
類似地,記憶體單元415包含開關T49、T410及電阻R45。開關T49的一端用以接收位元線信號WBL2,開關T49的另一端耦接儲存節點N45,且開關T49的控制端用以接收字元線信號WWL2。開關T410的一端耦接電阻R45,開關T410的另一端用以接收位元線信號RBL2,且開關T410的控制端耦接儲存節點N45。電阻R45的一端耦接開關T410,且電阻R45的另一端用以接收字元線信號RWL2。Similarly, the
類似地,記憶體單元416包含開關T411、T412及電阻R46。開關T411的一端用以接收位元線信號WBL3,開關T411的另一端耦接儲存節點N46,且開關T411的控制端用以接收字元線信號WWL2。開關T412的一端耦接電阻R46,開關T412的另一端用以接收位元線信號RBL3,且開關T412的控制端耦接儲存節點N46。電阻R46的一端耦接開關T412,且電阻R46的另一端用以接收字元線信號RWL2。Similarly, the
類似地,記憶體單元417包含開關T413、T414及電阻R47。開關T413的一端用以接收位元線信號WBL4,開關T413的另一端耦接儲存節點N47,且開關T413的控制端用以接收字元線信號WWL2。開關T414的一端耦接電阻R47,開關T414的另一端用以接收位元線信號RBL4,且開關T414的控制端耦接儲存節點N47。電阻R47的一端耦接開關T414,且電阻R47的另一端用以接收字元線信號RWL2。Similarly, the
在一些實施例中,記憶體單元411~417更分別包含電容C41~C47。電容C41~C47分別耦接節點N41~N47。在一些實施例中,電容C41~C47的配置類似於電容C11的配置,且電阻R41~R47的配置類似於電阻R3的配置。因此,為簡潔起見,部分敘述不再重複說明。In some embodiments, the memory cells 411-417 further include capacitors C41-C47, respectively. The capacitors C41-C47 are coupled to nodes N41-N47, respectively. In some embodiments, the configuration of the capacitors C41-C47 is similar to the configuration of the capacitor C11, and the configuration of the resistors R41-R47 is similar to the configuration of the resistor R3. Therefore, for the sake of brevity, some descriptions are not repeated.
如第4圖所示,位元線信號RBL1~RBL4分別用以攜載輸入位元IPT1~IPT4。位元線信號WBL1用以攜載權重位元W11及W21。位元線信號WBL2用以攜載權重位元W12及W22。位元線信號WBL3用以攜載權重位元W13及W23。位元線信號WBL4用以攜載權重位元W14及W24。As shown in FIG. 4 , bit line signals RBL1 to RBL4 are used to carry input bits IPT1 to IPT4, respectively. Bit line signal WBL1 is used to carry weight bits W11 and W21. Bit line signal WBL2 is used to carry weight bits W12 and W22. Bit line signal WBL3 is used to carry weight bits W13 and W23. Bit line signal WBL4 is used to carry weight bits W14 and W24.
在一些實施例中,輸入位元IPT1~IPT4、權重位元W11~W14、W21~W24的邏輯值及位元線信號RBL1~RBL4、WBL1~WBL4的電壓準位之間的關係類似於輸入位元IPT1、權重位元W11的邏輯值及位元線信號RBL1、WBL1的電壓準位之間的關係。因此,為簡潔起見,部分敘述不再重複說明。In some embodiments, the relationship between the logic values of the input bits IPT1-IPT4, the weight bits W11-W14, W21-W24, and the voltage levels of the bit line signals RBL1-RBL4, WBL1-WBL4 is similar to the relationship between the logic value of the input bit IPT1, the weight bit W11, and the voltage levels of the bit line signals RBL1, WBL1. Therefore, for the sake of brevity, some descriptions are not repeated.
操作上,記憶體單元110及411~413可以進行第一寫入操作以將權重位元W11~W14分別寫入儲存節點N11及N41~N43。在第一寫入操作之後,記憶體單元414~417可以進行第二寫入操作以將權重位元W21~W24分別寫入儲存節點N44~N47。In operation, the
在第一寫入操作時,字元線信號WWL1及WWL2分別具有致能電壓準位及禁能電壓準位,使得開關T11、T41、T43及T45導通,且開關T47、T49、T411及T413關閉。此時,位元線信號WBL1~WBL4分別具有權重位元W11~W14,使得儲存節點N11及N41~N43可以寫入權重位元W11~W14。在第一寫入操作結束時,電容C11及C41~C43可以儲存權重位元W11~W14對應的電壓準位。During the first write operation, word line signals WWL1 and WWL2 have an enable voltage level and a disable voltage level, respectively, so that switches T11, T41, T43 and T45 are turned on, and switches T47, T49, T411 and T413 are turned off. At this time, bit line signals WBL1-WBL4 have weight bits W11-W14, respectively, so that storage nodes N11 and N41-N43 can write weight bits W11-W14. When the first write operation is completed, capacitors C11 and C41-C43 can store voltage levels corresponding to weight bits W11-W14.
在第二寫入操作時,字元線信號WWL1及WWL2分別具有禁能電壓準位及致能電壓準位,使得開關T11、T41、T43及T45關閉,且開關T47、T49、T411及T413導通。此時,位元線信號WBL1~WBL4分別具有權重位元W21~W24,使得儲存節點N44~N47可以寫入權重位元W21~W24。在第二寫入操作結束時,電容C11及C44~C47可以儲存權重位元W21~W24對應的電壓準位。During the second write operation, the word line signals WWL1 and WWL2 have a disable voltage level and an enable voltage level, respectively, so that the switches T11, T41, T43 and T45 are closed, and the switches T47, T49, T411 and T413 are turned on. At this time, the bit line signals WBL1-WBL4 have weight bits W21-W24, respectively, so that the storage nodes N44-N47 can write the weight bits W21-W24. When the second write operation is completed, the capacitors C11 and C44-C47 can store the voltage levels corresponding to the weight bits W21-W24.
在一些實施例中,在第二寫入操作之後,記憶體裝置400還可以進行更多的寫入操作以將其他的權重位元寫入其他列的記憶體單元(圖中未示出)。在上述寫入操作之後,記憶體裝置400可以進行讀取操作以依據權重位元W11~W14、W21~W24及輸入位元IPT1~IPT4產生電流信號IS1及IS41~IS47。其中電流信號IS41~IS47分別流經開關T42、T44、T46、T48、T410、T412及T414。In some embodiments, after the second write operation, the
在讀取操作時,字元線信號WWL1及WWL2的每一者具有禁能電壓準位,使得開關T11、T41、T43、T45、T47、T49、T411及T413關閉。開關T12、T42、T44、T46、T48、T410、T412及T414分別依據儲存節點N11及N41~N47的電壓準位(亦即,權重位元W11~W14及W21~W24的邏輯值)導通或關閉。During the read operation, each of the word line signals WWL1 and WWL2 has a disable voltage level, so that the switches T11, T41, T43, T45, T47, T49, T411 and T413 are turned off. The switches T12, T42, T44, T46, T48, T410, T412 and T414 are turned on or off according to the voltage levels of the storage nodes N11 and N41-N47 (i.e., the logic values of the weight bits W11-W14 and W21-W24), respectively.
當儲存節點N11及N41~N47的一或多者具有電壓準位VWL時,開關T12、T42、T44、T46、T48、T410、T412及T414的對應一或多者關閉,使得電流信號IS1及IS41~IS47的對應一或多者具有電流準位ILL。When one or more of the storage nodes N11 and N41-N47 have a voltage level VWL, a corresponding one or more of the switches T12, T42, T44, T46, T48, T410, T412 and T414 are turned off, so that the current signal IS1 and a corresponding one or more of IS41-IS47 have a current level ILL.
當儲存節點N11及N41~N47的一或多者具有電壓準位VWH時,開關T12、T42、T44、T46、T48、T410、T412及T414的對應一或多者導通,使得電流信號IS1及IS41~IS47的對應一或多者具有對應輸入位元IPT1~IPT4的對應一或多者的電流準位。When one or more of the storage nodes N11 and N41-N47 have a voltage level VWH, a corresponding one or more of the switches T12, T42, T44, T46, T48, T410, T412 and T414 are turned on, so that the current signals IS1 and a corresponding one or more of IS41-IS47 have a current level corresponding to a corresponding one or more of the input bits IPT1-IPT4.
舉例來說,輸入位元IPT1~IPT4分別具有邏輯值1、1、0、0,其中邏輯值1及0分別對應電壓準位VL及VH。權重位元W11~W14及W21~W24分別具有邏輯值1、0、1、0、0、1、0、1,其中邏輯值1及0分別對應電壓準位VWH及VWL。For example, input bits IPT1-IPT4 have logic values 1, 1, 0, 0, respectively, wherein logic values 1 and 0 correspond to voltage levels VL and VH, respectively. Weight bits W11-W14 and W21-W24 have logic values 1, 0, 1, 0, 0, 1, 0, 1, respectively, wherein logic values 1 and 0 correspond to voltage levels VWH and VWL, respectively.
在上述範例中,回應於權重位元W11、W13、W22及W24的邏輯值1,儲存節點N11、N42、N45及N47具有電壓準位VWH,使得開關T12、T44、T410及T414導通。此時,回應於輸入位元IPT1及IPT2的邏輯值1,電流信號IS1及IS45具有電流準位ILH。另一方面,電流信號IS41~IS44及IS46~IS47具有電流準位ILL。此時,記憶體裝置400將電流信號IS1及IS41~IS43在節點N48相加以產生具有電流準位ILH的電流信號IM41,並將電流信號IS44~IS47在節點N49相加以產生具有電流準位ILH的電流信號IM42,使得放大器451及452產生對應電流準位ILH的電壓信號S41及S42。In the above example, in response to the logic value 1 of the weight bits W11, W13, W22 and W24, the storage nodes N11, N42, N45 and N47 have a voltage level VWH, so that the switches T12, T44, T410 and T414 are turned on. At this time, in response to the logic value 1 of the input bits IPT1 and IPT2, the current signals IS1 and IS45 have a current level ILH. On the other hand, the current signals IS41~IS44 and IS46~IS47 have a current level ILL. At this time, the
舉另一例來說,輸入位元IPT1~IPT4分別具有邏輯值1、1、1、0。權重位元W11~W14及W21~W24分別具有邏輯值1、1、0、1、0、0、0、1。For another example, input bits IPT1-IPT4 have logical values 1, 1, 1, 0, respectively. Weight bits W11-W14 and W21-W24 have logical values 1, 1, 0, 1, 0, 0, 0, 1, respectively.
在上述範例中,回應於權重位元W11、W12、W14及W24的邏輯值1,儲存節點N11、N41、N43及N47具有電壓準位VWH,使得開關T12、T42、T46及T414導通。此時,回應於輸入位元IPT1~IPT3的邏輯值1,電流信號IS1及IS41具有電流準位ILH。另一方面,電流信號IS42~IS47具有電流準位ILL。此時,記憶體裝置400將電流信號IS1及IS41~IS43在節點N48相加以產生具有電流準位2×ILH的電流信號IM41,並將電流信號IS44~IS47在節點N49相加以產生具有電流準位ILL的電流信號IM42,使得放大器451產生對應電流準位2×ILH的電壓信號S41,且放大器452產生對應電流準位ILL的電壓信號S42。In the above example, in response to the logic value 1 of the weight bits W11, W12, W14 and W24, the storage nodes N11, N41, N43 and N47 have a voltage level VWH, so that the switches T12, T42, T46 and T414 are turned on. At this time, in response to the logic value 1 of the input bits IPT1~IPT3, the current signals IS1 and IS41 have a current level ILH. On the other hand, the current signals IS42~IS47 have a current level ILL. At this time, the
在一些實施例中,電壓信號S41的電壓準位對應輸入位元IPT1~IPT4分別乘上權重位元W11~W14後,將乘積相加的結果。舉例來說,當IPT1×W11+IPT2×W12+IPT3×W13+IPT4×W14=1,電流信號IM41具有電流準位ILH,使得電壓信號S41具有對應邏輯值1的電壓準位VL。類似地,電壓信號S42的電壓準位對應輸入位元IPT1~IPT4分別乘上權重位元W21~W24後,將乘積相加的結果。舉例來說,當IPT1×W21+IPT2×W22+IPT3×W23+IPT4×W24=0,電流信號IM42具有電流準位ILL,使得電壓信號S42具有對應邏輯值0的電壓準位VH。In some embodiments, the voltage level of the voltage signal S41 corresponds to the result of adding the products of the input bits IPT1-IPT4 multiplied by the weight bits W11-W14. For example, when IPT1×W11+IPT2×W12+IPT3×W13+IPT4×W14=1, the current signal IM41 has a current level ILH, so that the voltage signal S41 has a voltage level VL corresponding to the logical value 1. Similarly, the voltage level of the voltage signal S42 corresponds to the result of adding the products of the input bits IPT1-IPT4 multiplied by the weight bits W21-W24. For example, when IPT1×W21+IPT2×W22+IPT3×W23+IPT4×W24=0, the current signal IM42 has a current level ILL, so that the voltage signal S42 has a voltage level VH corresponding to a logical value of 0.
綜上所述,記憶體裝置400將輸入位元IPT1~IPT4分別乘上權重位元W11~W14,並將電流IS1及IS41~IS43作為乘積相加以產生電流信號IM41。記憶體裝置400也將輸入位元IPT1~IPT4分別乘上權重位元W21~W24,並將電流IS44~IS47作為乘積相加以產生電流信號IM42。如此一來,記憶體裝置400可以達成輸入位元IPT1~IPT4及權重位元W11~W14、W21~W24的乘積累加邏輯操作。In summary, the
第5圖為根據本案之一實施例所繪示之記憶體裝置400的進一步細節的示意圖。請參照第5圖及第4圖,在一些變化例中,記憶體裝置400可以更包含記憶體單元511~514及放大器551、552。記憶體單元511及512用以對電壓信號S41及S42進行第三乘積累加運算,以產生電流信號IM51。記憶體單元513及514用以對電壓信號S41及S42進行第四乘積累加運算,以產生電流信號IM52。放大器551用以依據電流信號IM51產生電壓信號S51。放大器552用以依據電流信號IM52產生電壓信號S52。FIG. 5 is a schematic diagram showing further details of a
如第5圖所示,記憶體單元511包含開關T51、T52及電阻R51。開關T51的一端用以接收位元線信號WBL51,開關T51的另一端耦接儲存節點N51,且開關T51的控制端用以接收字元線信號WWL51。開關T52的一端耦接電阻R51,開關T52的另一端用以接收電壓信號S41,且開關T52的控制端耦接儲存節點N51。電阻R51的一端耦接開關T52,且電阻R51的另一端用以接收字元線信號RWL51。As shown in FIG. 5 , the
類似地,記憶體單元512包含開關T53、T54及電阻R52。開關T53的一端用以接收位元線信號WBL52,開關T53的另一端耦接儲存節點N52,且開關T53的控制端用以接收字元線信號WWL51。開關T54的一端耦接電阻R52,開關T54的另一端用以接收電壓信號S42,且開關T54的控制端耦接儲存節點N52。電阻R52的一端耦接開關T54,且電阻R52的另一端用以接收字元線信號RWL51。Similarly, the
類似地,記憶體單元513包含開關T55、T56及電阻R53。開關T55的一端用以接收位元線信號WBL51,開關T55的另一端耦接儲存節點N53,且開關T55的控制端用以接收字元線信號WWL52。開關T56的一端耦接電阻R53,開關T56的另一端用以接收電壓信號S41,且開關T56的控制端耦接儲存節點N53。電阻R53的一端耦接開關T56,且電阻R53的另一端用以接收字元線信號RWL52。Similarly, the
類似地,記憶體單元514包含開關T57、T58及電阻R54。開關T57的一端用以接收位元線信號WBL52,開關T57的另一端耦接儲存節點N54,且開關T57的控制端用以接收字元線信號WWL52。開關T58的一端耦接電阻R54,開關T58的另一端用以接收電壓信號S42,且開關T58的控制端耦接儲存節點N54。電阻R54的一端耦接開關T58,且電阻R54的另一端用以接收字元線信號RWL52。Similarly, the
在一些實施例中,記憶體單元511~514更分別包含電容C51~C54。電容C51~C54分別耦接節點N51~N54。在一些實施例中,電容C51~C54的配置類似於電容C11的配置,且電阻R51~R54的配置類似於電阻R3的配置。因此,為簡潔起見,部分敘述不再重複說明。In some embodiments, the memory cells 511-514 further include capacitors C51-C54, respectively. The capacitors C51-C54 are coupled to nodes N51-N54, respectively. In some embodiments, the configuration of the capacitors C51-C54 is similar to the configuration of the capacitor C11, and the configuration of the resistors R51-R54 is similar to the configuration of the resistor R3. Therefore, for the sake of brevity, some descriptions are not repeated.
在一些實施例中,位元線信號WBL51用以攜載權重位元W51及W53。位元線信號WBL52用以攜載權重位元W52及W54。記憶體單元511~514可以對權重位元W51~W54及電壓信號S41、S42(亦即,第一及第二乘積累加邏輯操作的輸出)進行第三及第四乘積累加邏輯操作,以產生電流信號IM51及IM52。In some embodiments, the bit line signal WBL51 is used to carry weight bits W51 and W53. The bit line signal WBL52 is used to carry weight bits W52 and W54. The memory units 511-514 can perform third and fourth multiplication and accumulation logic operations on the weight bits W51-W54 and the voltage signals S41 and S42 (i.e., the outputs of the first and second multiplication and accumulation logic operations) to generate current signals IM51 and IM52.
操作上,記憶體單元511及512可以進行第一寫入操作以將權重位元W51及W52分別寫入儲存節點N51及N52。在第一寫入操作之後,記憶體單元513及514可以進行第二寫入操作以將權重位元W53及W54分別寫入節點N53及N54。In operation,
在第一寫入操作時,字元線信號WWL51及WWL52分別具有致能電壓準位及禁能電壓準位,使得開關T51及T53導通,且開關T52及T54關閉。此時,位元線信號WBL51及WBL52分別具有權重位元W51及W52,使得儲存節點N51及N52可以寫入權重位元W51及W52。在第一寫入操作結束時,電容C51及C52可以儲存權重位元W51及W52對應的電壓準位。During the first write operation, word line signals WWL51 and WWL52 have an enable voltage level and a disable voltage level, respectively, so that switches T51 and T53 are turned on, and switches T52 and T54 are turned off. At this time, bit line signals WBL51 and WBL52 have weight bits W51 and W52, respectively, so that storage nodes N51 and N52 can write weight bits W51 and W52. When the first write operation is completed, capacitors C51 and C52 can store voltage levels corresponding to weight bits W51 and W52.
在第二寫入操作時,字元線信號WWL51及WWL52分別具有禁能電壓準位及致能電壓準位,使得開關T51及T53關閉,且開關T52及T54導通。此時,位元線信號WBL51及WBL52分別具有權重位元W53及W54,使得儲存節點N53及N54可以寫入權重位元W53及W54。在第二寫入操作結束時,電容C53及C54可以儲存權重位元W53及W54對應的電壓準位。During the second write operation, word line signals WWL51 and WWL52 have a disable voltage level and an enable voltage level, respectively, so that switches T51 and T53 are closed, and switches T52 and T54 are turned on. At this time, bit line signals WBL51 and WBL52 have weight bits W53 and W54, respectively, so that storage nodes N53 and N54 can write weight bits W53 and W54. When the second write operation is completed, capacitors C53 and C54 can store voltage levels corresponding to weight bits W53 and W54.
在第二寫入操作之後,記憶體裝置400可以進行讀取操作以依據權重位元W51~W54及電壓信號S41、S42產生電流信號IS51~IS54。其中電流信號IS51~IS54分別流經開關T52、T54、T56及T58。After the second write operation, the
在讀取操作時,字元線信號WWL51及WWL52的每一者具有禁能電壓準位,使得開關T51、T53、T55及T57關閉。開關T52、T54、T56及T58分別依據儲存節點N51~N54的電壓準位(亦即,權重位元W51~W54的邏輯值)導通或關閉。During the read operation, each of the word line signals WWL51 and WWL52 has a disable voltage level, so that switches T51, T53, T55 and T57 are closed. Switches T52, T54, T56 and T58 are turned on or off according to the voltage level of the storage nodes N51-N54 (i.e., the logical value of the weight bits W51-W54).
舉例來說,當權重位元W51~W54分別具有邏輯值1、0、0、1時,開關T52及T58回應於權重位元W51及W54的邏輯值1導通,且開關T54及T56回應於權重位元W52及W53的邏輯值0關閉。For example, when weight bits W51~W54 have logic values 1, 0, 0, 1 respectively, switches T52 and T58 are turned on in response to the logic value 1 of weight bits W51 and W54, and switches T54 and T56 are turned off in response to the logic value 0 of weight bits W52 and W53.
此時,若電壓信號S41及S42分別具有對應邏輯值1的電壓準位VL及對應邏輯值0的電壓準位VH,則電流信號IS51具有電流準位ILH,且電流信號IS52~IS54具有電流準位ILL。對應地,記憶體單元511及512將電流信號IS51及IS52在節點N55相加以產生具有電流準位ILH的電流信號IM51,且記憶體單元513及514將電流信號IS53及IS54在節點N56相加以產生具有電流準位ILL的電流信號IM52。At this time, if the voltage signals S41 and S42 respectively have a voltage level VL corresponding to a logic value of 1 and a voltage level VH corresponding to a logic value of 0, the current signal IS51 has a current level ILH, and the current signals IS52-IS54 have a current level ILL. Correspondingly, the
綜上所述,記憶體裝置400將電壓信號S41及S42的邏輯值分別乘上權重位元W51及W52,並將電流IS51及IS52作為乘積相加以產生電流信號IM51。記憶體裝置400也將電壓信號S41及S42的邏輯值分別乘上權重位元W53及W54,並將電流IS53及IS54作為乘積相加以產生電流信號IM52。如此一來,記憶體裝置400可以達成電壓信號S41及S42及權重位元W51~W54的乘積累加邏輯操作。In summary, the
請參照第4圖及第5圖,在記憶體單元110及411~417進行第一及第二乘積累加邏輯操作以產生電壓信號S41及S42之後,記憶體單元511~514對電壓信號S41及S42進行第三及第四乘積累加邏輯操作以產生電壓信號S51及S52。4 and 5 , after the
應用上,記憶體單元110及411~417可以包含於神經網路卷積層的第一層,且電壓信號S41及S42對應第一層的輸出。記憶體單元511~514可以包含於神經網路卷積層的第二層,電壓信號S41及S42對應第二層的輸入,且電壓信號S51及S52對應第二層的輸出。In application,
第6圖為根據本案之一實施例所繪示之記憶體裝置600的示意圖。請參照第6圖及第4圖,記憶體裝置600係記憶體裝置400的一種變化例。記憶體裝置600的部份元件沿用記憶體裝置400的標號方式。為簡潔起見,討論將集中在記憶體裝置600不同於記憶體裝置400的部份而非相同之處。FIG. 6 is a schematic diagram of a
相較於第4圖所示之實施例,在第6圖所示之實施例中,位元線信號RBL3及RBL4分別用以攜載輸入位元IPTB1及IPTB2,而非輸入位元IPT3及IPT4。位元線信號WBL3用以攜載權重位元WB11及WB21而非權重位元W13及W23。位元線信號WBL4用以攜載權重位元WB12及WB22而非權重位元W14及W24。其中輸入位元IPTB1、IPTB2及權重位元WB11、WB21、WB12、WB22分別與輸入位元IPT1、IPT2及權重位元W11、W21、W12、W22互補。Compared to the embodiment shown in FIG. 4, in the embodiment shown in FIG. 6, bit line signals RBL3 and RBL4 are used to carry input bits IPTB1 and IPTB2, respectively, instead of input bits IPT3 and IPT4. Bit line signal WBL3 is used to carry weight bits WB11 and WB21, instead of weight bits W13 and W23. Bit line signal WBL4 is used to carry weight bits WB12 and WB22, instead of weight bits W14 and W24. The input bits IPTB1, IPTB2 and weight bits WB11, WB21, WB12, WB22 complement the input bits IPT1, IPT2 and weight bits W11, W21, W12, W22, respectively.
在一些實施例中,記憶體裝置600更包含反相器NV61及NV62。反相器NV61用以接收位元線信號RBL1並輸出位元線信號RBL3。反相器NV62用以接收位元線信號RBL2並輸出位元線信號RBL4。換言之,反相器NV61用以反相輸入位元IPT1以產生輸入位元IPTB1,且反相器NV62用以反相輸入位元IPT2以產生輸入位元IPTB2。In some embodiments, the
回應於位元線信號WBL3攜載權重位元WB11及WB21且位元線信號WBL4攜載權重位元WB12及WB22,儲存節點N42、N43、N46及N47分別用以儲存對應權重位元WB11、WB12、WB21及WB22的電壓準位。In response to the bit line signal WBL3 carrying the weight bits WB11 and WB21 and the bit line signal WBL4 carrying the weight bits WB12 and WB22, the storage nodes N42, N43, N46 and N47 are used to store the voltage levels corresponding to the weight bits WB11, WB12, WB21 and WB22 respectively.
操作上,記憶體裝置600可以執行搜尋邏輯操作。具體而言,記憶體裝置600用以比較輸入位元及權重位元。當輸入位元及權重位元之間的相似度增加時,記憶體裝置600產生的電流信號的電流準位增加。In operation, the
舉例來說,當輸入位元IPT1及IPT2分別具有邏輯值1及0時,位元線信號RBL1~RBL4分別具有電壓準位VL、VH、VH、VL。若權重位元W11及W12也分別具有邏輯值1及0,則輸入位元IPT1、IPT2及權重位元W11、W12具有100%的相似度。此時,儲存節點N11及N41~N43分別具有電壓準位VWH、VWL、VWL、VWH。對應地,電流信號IS11及IS41~IS43分別具有電流準位ILH、ILL、ILL及ILH,使得電流信號IM41具有電流準位2×ILH。For example, when input bits IPT1 and IPT2 have logic values 1 and 0, respectively, bit line signals RBL1-RBL4 have voltage levels VL, VH, VH, VL, respectively. If weight bits W11 and W12 also have logic values 1 and 0, respectively, input bits IPT1, IPT2 and weight bits W11, W12 have 100% similarity. At this time, storage nodes N11 and N41-N43 have voltage levels VWH, VWL, VWL, VWH, respectively. Correspondingly, current signals IS11 and IS41-IS43 have current levels ILH, ILL, ILL, and ILH, respectively, so that current signal IM41 has a current level 2×ILH.
舉另一例來說,當輸入位元IPT1、IPT2分別具有邏輯值1、0且權重位元W21、W22分別具有邏輯值0、1時,輸入位元IPT1、IPT2及權重位元W21、W22具有0%的相似度。此時,位元線信號RBL1~RBL4分別具有電壓準位VL、VH、VH、VL,且儲存節點 N44~N47分別具有電壓準位VWL、VWH、VWH、VWL。對應地,電流信號IS44~IS47的每一者具有電流準位ILL,使得電流信號IM42具有電流準位ILL。For another example, when the input bits IPT1 and IPT2 have logic values 1 and 0 respectively and the weight bits W21 and W22 have logic values 0 and 1 respectively, the input bits IPT1 and IPT2 and the weight bits W21 and W22 have 0% similarity. At this time, the bit line signals RBL1-RBL4 have voltage levels VL, VH, VH, VL respectively, and the storage nodes N44-N47 have voltage levels VWL, VWH, VWH, VWL respectively. Correspondingly, each of the current signals IS44-IS47 has a current level ILL, so that the current signal IM42 has a current level ILL.
又舉另一例來說,當輸入位元IPT1、IPT2分別具有邏輯值1、0且權重位元W21、W22分別具有邏輯值1、1時,輸入位元IPT1、IPT2及權重位元W21、W22具有50%的相似度。此時,位元線信號RBL1~RBL4分別具有電壓準位VL、VH、VH、VL,且儲存節點 N44~N47分別具有電壓準位VWH、VWH、VWL、VWL。對應地,電流信號IS44~IS47分別具有電流準位ILH、ILL、ILL、ILL,使得電流信號IM42具有電流準位ILH。As another example, when the input bits IPT1 and IPT2 have logic values 1 and 0 respectively and the weight bits W21 and W22 have logic values 1 and 1 respectively, the input bits IPT1 and IPT2 and the weight bits W21 and W22 have a similarity of 50%. At this time, the bit line signals RBL1-RBL4 have voltage levels VL, VH, VH, VL respectively, and the storage nodes N44-N47 have voltage levels VWH, VWH, VWL, VWL respectively. Correspondingly, the current signals IS44-IS47 have current levels ILH, ILL, ILL, ILL respectively, so that the current signal IM42 has a current level ILH.
綜上所述,回應於輸入位元及權重位元之間具有0%、50%及100%的相似度,記憶體裝置600產生的電流信號分別具有電流準位ILL、ILH及2×ILH。如此一來,可以藉由電流準位判斷輸入位元及權重位元的相似度,以實行搜尋邏輯操作。In summary, in response to the similarity between the input bit and the weight bit being 0%, 50% and 100%, the current signal generated by the
第7圖為根據本案之一實施例所繪示之記憶體裝置700的示意圖。請參照第7圖及第6圖,記憶體裝置700係記憶體裝置600的一種變化例。記憶體裝置700的部份元件沿用記憶體裝置600的標號方式。為簡潔起見,討論將集中在記憶體裝置700不同於記憶體裝置600的部份而非相同之處。FIG. 7 is a schematic diagram of a
相較於記憶體裝置600,在記憶體裝置700中,記憶體單元110、411、414、415與記憶體單元412、413、416、417係布置於不同的記憶體陣列中。此外,相較於記憶體裝置600,記憶體裝置700更包含放大器751~754而非放大器451及452。Compared to the
如第7圖所示,放大器751耦接電阻R3及R41於節點N71。放大器752耦接電阻R44及R45於節點N72。放大器753耦接電阻R42及R43於節點N73。放大器754耦接電阻R46及R47於節點N74。As shown in FIG. 7 ,
操作上,記憶體單元110及411用以將電流信號IS11及IS41在節點N71相加以產生電流信號IM71。放大器751用以依據電流信號IM71產生電壓信號S71。記憶體單元414及415用以將電流信號IS44及IS45在節點N72相加以產生電流信號IM72。放大器752用以依據電流信號IM72產生電壓信號S72。In operation,
類似地,記憶體單元412及413用以將電流信號IS42及IS43在節點N73相加以產生電流信號IM73。放大器753用以依據電流信號IM73產生電壓信號S73。記憶體單元416及417用以將電流信號IS46及IS47在節點N74相加以產生電流信號IM74。放大器754用以依據電流信號IM74產生電壓信號S74。Similarly,
在一些實施例中,當電流信號IM71~IM74的一者具有對應邏輯值1的電流準位ILH時,電壓信號S71~S74的對應一者具有對應邏輯值1的電壓準位VL。當電流信號IM71~IM74的一者具有對應邏輯值0的電流準位ILL時,電壓信號S71~S74的對應一者具有對應邏輯值0的電壓準位VH。In some embodiments, when one of the current signals IM71-IM74 has a current level ILH corresponding to a logic value of 1, the corresponding one of the voltage signals S71-S74 has a voltage level VL corresponding to a logic value of 1. When one of the current signals IM71-IM74 has a current level ILL corresponding to a logic value of 0, the corresponding one of the voltage signals S71-S74 has a voltage level VH corresponding to a logic value of 0.
在一些實施例中,處理器(圖中未示出)用以依據電壓信號S71及S72判斷輸入位元IPT1、IPT2與權重位元W11、W12的相似度,並且依據電壓信號S73及S74判斷輸入位元IPT1、IPT2與權重位元W21、W22的相似度。In some embodiments, a processor (not shown) is used to determine the similarity between input bits IPT1, IPT2 and weight bits W11, W12 based on voltage signals S71 and S72, and to determine the similarity between input bits IPT1, IPT2 and weight bits W21, W22 based on voltage signals S73 and S74.
在第7圖及第6圖所示,記憶體裝置700或600的配置可以藉由相同或不同的記憶體陣列達成搜尋邏輯操作。如此一來,可以依據不同的規格需求進行設計,使得自由度提升。As shown in FIG7 and FIG6, the configuration of the
雖然本揭示內容已以實施例揭露如上,然其並非用以限定本揭示內容,任何所屬技術領域中具有通常知識者,在不脫離本揭示內容的精神和範圍內,當可作些許的更動與潤飾,故本揭示內容的保護範圍當視後附的申請專利範圍所界定者為準。Although the contents of this disclosure have been disclosed as above by way of embodiments, they are not intended to limit the contents of this disclosure. Any person with ordinary knowledge in the relevant technical field may make some changes and modifications without departing from the spirit and scope of the contents of this disclosure. Therefore, the protection scope of the contents of this disclosure shall be subject to the scope defined by the attached patent application.
100、200、300、400、600、700:記憶體裝置
110、210、411~417、511~514:記憶體單元
WWL1、RWL1、WWL2、RWL2、WWL51、RWL51、WWL52、RWL52:字元線信號
WBL1~WBL4、RBL1~RBL4、WBL51、WBL52:位元線信號
W11~W14、W21~W24、WB11、WB12、WB21、WB22、W51~W54:權重位元
IPT1~IPT4、IPTB1、IPTB2:輸入位元
IS1、IS2、IM41、IM42、IS41~IS47、IS51~IS54 IM51、IM52、IM71~IM74:電流信號
T11、T12、T21、T22、T41~T414、T51~T58:開關
N11、N21、N31、N41~N47、N51~N54:儲存節點
M48、N49、N55、N56、N71~N74:節點
C11、C21、C41~C47、C51~C54:電容
NV21、NV22、NV61、NV62:反相器
R3、R41~R47、R51~R54:電阻
451、452、551、552、751~754:放大器
S41、S42、S51、S52、S71~S74:電壓信號100, 200, 300, 400, 600, 700:
第1圖為根據本案之一實施例所繪示之記憶體裝置的示意圖。 第2圖為根據本案之一實施例所繪示之記憶體裝置的示意圖。 第3圖為根據本案之一實施例所繪示之記憶體裝置的示意圖。 第4圖為根據本案之一實施例所繪示之記憶體裝置的示意圖。 第5圖為根據本案之一實施例所繪示之記憶體裝置的進一步細節的示意圖。 第6圖為根據本案之一實施例所繪示之記憶體裝置的示意圖。 第7圖為根據本案之一實施例所繪示之記憶體裝置的示意圖。 FIG. 1 is a schematic diagram of a memory device according to one embodiment of the present invention. FIG. 2 is a schematic diagram of a memory device according to one embodiment of the present invention. FIG. 3 is a schematic diagram of a memory device according to one embodiment of the present invention. FIG. 4 is a schematic diagram of a memory device according to one embodiment of the present invention. FIG. 5 is a schematic diagram of further details of a memory device according to one embodiment of the present invention. FIG. 6 is a schematic diagram of a memory device according to one embodiment of the present invention. FIG. 7 is a schematic diagram of a memory device according to one embodiment of the present invention.
國內寄存資訊(請依寄存機構、日期、號碼順序註記) 無 國外寄存資訊(請依寄存國家、機構、日期、號碼順序註記) 無 Domestic storage information (please note in the order of storage institution, date, and number) None Foreign storage information (please note in the order of storage country, institution, date, and number) None
100:記憶體裝置 100: Memory device
110:記憶體單元 110: Memory unit
WWL1、RWL1:字元線信號 WWL1, RWL1: word line signal
WBL1、RBL1:位元線信號 WBL1, RBL1: bit line signal
W11:權重位元 W11: Weight bit
IPT1:輸入位元 IPT1: Input bits
IS1:電流信號 IS1: Current signal
T11、T12:開關 T11, T12: switch
N11:儲存節點 N11: Storage Node
C11:電容 C11: Capacitor
Claims (10)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW112135761A TWI877798B (en) | 2023-09-19 | 2023-09-19 | Memory device and operating method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW112135761A TWI877798B (en) | 2023-09-19 | 2023-09-19 | Memory device and operating method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
TWI877798B true TWI877798B (en) | 2025-03-21 |
TW202514609A TW202514609A (en) | 2025-04-01 |
Family
ID=95830329
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW112135761A TWI877798B (en) | 2023-09-19 | 2023-09-19 | Memory device and operating method thereof |
Country Status (1)
Country | Link |
---|---|
TW (1) | TWI877798B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5646903A (en) * | 1996-03-06 | 1997-07-08 | Xilinx, Inc. | Memory cell having a shared read/write line |
US6314017B1 (en) * | 1999-07-22 | 2001-11-06 | Sony Corporation | Semiconductor memory device |
US20040252542A1 (en) * | 2003-06-16 | 2004-12-16 | Katsuhiko Hoya | Ferroelectric memory device |
CN101004952A (en) * | 2006-01-19 | 2007-07-25 | 奇晶光电股份有限公司 | Shift register and flat display device using same |
WO2020150810A1 (en) * | 2019-01-21 | 2020-07-30 | Mitchell Miller | A system and method for bidirectionally based electrical information storage, processing and communication. |
-
2023
- 2023-09-19 TW TW112135761A patent/TWI877798B/en active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5646903A (en) * | 1996-03-06 | 1997-07-08 | Xilinx, Inc. | Memory cell having a shared read/write line |
US6314017B1 (en) * | 1999-07-22 | 2001-11-06 | Sony Corporation | Semiconductor memory device |
US20040252542A1 (en) * | 2003-06-16 | 2004-12-16 | Katsuhiko Hoya | Ferroelectric memory device |
CN101004952A (en) * | 2006-01-19 | 2007-07-25 | 奇晶光电股份有限公司 | Shift register and flat display device using same |
WO2020150810A1 (en) * | 2019-01-21 | 2020-07-30 | Mitchell Miller | A system and method for bidirectionally based electrical information storage, processing and communication. |
Also Published As
Publication number | Publication date |
---|---|
TW202514609A (en) | 2025-04-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US8582338B1 (en) | Ternary content addressable memory cell having single transistor pull-down stack | |
EP2939240B1 (en) | Static nand cell for ternary content addressable memory (tcam) | |
US8625320B1 (en) | Quaternary content addressable memory cell having one transistor pull-down stack | |
US8320148B1 (en) | PMC-based non-volatile CAM | |
EP3136398B1 (en) | Ternary content addressable memory (tcam) with programmable resistive elements | |
JP2010267373A (en) | Semiconductor device | |
US20040076065A1 (en) | Methods of reading and/or writing data to memory devices including virtual ground lines and/or multiple write circuits and related devices | |
JP5404559B2 (en) | Associative memory array | |
US8553441B1 (en) | Ternary content addressable memory cell having two transistor pull-down stack | |
JP4066638B2 (en) | Semiconductor device | |
JP2011081874A5 (en) | ||
JP2020135913A (en) | Semiconductor storage device | |
JP2002334585A (en) | Semiconductor memory | |
TWI877798B (en) | Memory device and operating method thereof | |
US8400802B2 (en) | Binary content addressable memory | |
US9312006B2 (en) | Non-volatile ternary content-addressable memory with resistive memory device | |
US20220415396A1 (en) | Ternary content addressable memory device based on ternary memory cell | |
US20200402552A1 (en) | Content addressable memory device with charge sharing based selective match line precharging scheme | |
JP2018092695A (en) | Semiconductor memory device | |
Zhang et al. | Fully integrated 3-D stackable CNTFET/RRAM 1T1R array as BEOL buffer macro for monolithic 3-D integration with analog RRAM-based computing-in-memory | |
CN119673235A (en) | Memory device and operation method thereof | |
TW201023182A (en) | Reading and comparison circuits with low voltage content address-enabling memory | |
US12283953B2 (en) | Inverter including transistors having different threshold voltages and memory cell including the same | |
CN111489779A (en) | Double-separation-gate flash memory circuit, storage device and reading method | |
JP2003123484A (en) | CAM cell |