TWI393189B - Design method and architecture for power gate switch placement - Google Patents

Design method and architecture for power gate switch placement Download PDF

Info

Publication number
TWI393189B
TWI393189B TW94129877A TW94129877A TWI393189B TW I393189 B TWI393189 B TW I393189B TW 94129877 A TW94129877 A TW 94129877A TW 94129877 A TW94129877 A TW 94129877A TW I393189 B TWI393189 B TW I393189B
Authority
TW
Taiwan
Prior art keywords
logic
unit
units
voltage reference
power gate
Prior art date
Application number
TW94129877A
Other languages
Chinese (zh)
Other versions
TW200620486A (en
Inventor
Gerald L Frenkil
Srinivasan Venkatraman
Original Assignee
Sequence Design Inc
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
Priority claimed from US10/998,204 external-priority patent/US7590962B2/en
Application filed by Sequence Design Inc filed Critical Sequence Design Inc
Publication of TW200620486A publication Critical patent/TW200620486A/en
Application granted granted Critical
Publication of TWI393189B publication Critical patent/TWI393189B/en

Links

Landscapes

  • Design And Manufacture Of Integrated Circuits (AREA)
  • Semiconductor Integrated Circuits (AREA)

Description

用於電源閘開關配置之設計方法與架構Design method and architecture for power gate switch configuration 發明領域Field of invention

本發明係有關於設計多臨界電壓互補金屬氧化物半導體(MTCMOS)電路之方法與由使用此方法結果所得之設計的實體架構。特別是,本發明係有關於在MTCMOS設計中佈置電源開關之方法及由使用此方法結果所得之設計的實體架構。The present invention relates to a method of designing a multi-threshold voltage complementary metal oxide semiconductor (MTCMOS) circuit and a physical architecture resulting from the results obtained using the method. In particular, the present invention relates to a method of arranging a power switch in an MTCMOS design and a physical architecture resulting from the results obtained using the method.

發明背景Background of the invention

在積體設計設計中之一重大關切為減少洩漏電流。洩漏電流因在邏輯設計中之電晶體的切換特徵為不理想的(即電晶體無法完全地被關閉)而在邏輯電路中由電源節點流入接地節點。One of the major concerns in integrated design is the reduction of leakage current. The leakage current flows into the ground node from the power supply node in the logic circuit due to the undesirable switching characteristics of the transistor in the logic design (ie, the transistor cannot be completely turned off).

在MTCMOS電路中,減少洩漏電流之一技術為在一邏輯閘之最低電位接頭(「虛擬接地」基準)與接地基準間佈置一「電源閘」(亦被習知為「電源開關」或簡單地為「開關單元」)。此技術在第1圖中被顯示,其顯示控制邏輯單元102至接地之洩漏電流路徑的電源閘或邏輯單元101。如第1圖顯示者,開關單元102使用低臨界電壓電晶體被形成以提供短路開關時間。此電源閘典型上為具有之臨界電壓比被用以施作該等邏輯單元之電晶體的臨界電壓為較高之電晶體。電源閘岔斷至接地之洩漏電流路徑。當電源閘101為導電的(即高電壓在控制節點106被提供)時,洩漏電流由電源節點104透過邏輯單元102流至虛擬接地節點103及透過電源閘101至真實接地節點105。然而,在待命之際(即當小於電源閘101之臨界電壓很多之電壓被施加至控制節點106時),電源閘101切斷由虛擬接地節點103至真實接地節點105之路徑。In an MTCMOS circuit, one technique for reducing the leakage current is to place a "power gate" between the lowest potential junction of a logic gate ("virtual ground" reference) and the ground reference (also known as a "power switch" or simply It is a "switch unit"). This technique is shown in Figure 1 which shows the power gate or logic unit 101 that controls the logic unit 102 to the grounded leakage current path. As shown in Figure 1, the switching unit 102 is formed using a low threshold voltage transistor to provide a short switching time. The power gate is typically a transistor having a threshold voltage that is higher than a threshold voltage of a transistor used to apply the logic cells. The leakage current path from the power gate to ground. When the power gate 101 is electrically conductive (i.e., a high voltage is provided at the control node 106), the leakage current is passed by the power supply node 104 through the logic unit 102 to the virtual ground node 103 and through the power gate 101 to the real ground node 105. However, at the time of standby (i.e., when a voltage that is much less than the threshold voltage of the power gate 101 is applied to the control node 106), the power gate 101 cuts off the path from the virtual ground node 103 to the real ground node 105.

數種設計方法曾被用以提供電源閘單元,如以第2圖舉例顯示的一方法(「積體開關」)整合電源閘202與邏輯單元201。在此配置中,邏輯單元201依據慣常之標準單元設計被佈置成列。如第2圖顯示者,導體204a與204b為部分之提供電源電壓至開關單元的電源柵格。類似地,導體203a與203b為部分之提供真實接地基準至開關單元的接地基準柵格。虛擬接地節點被置於每一開關單元(如開關單元101)內。Several design methods have been used to provide the power gate unit, such as the method shown in Figure 2 ("integrated switch") to integrate the power gate 202 with the logic unit 201. In this configuration, logic unit 201 is arranged in columns according to conventional standard cell designs. As shown in Figure 2, conductors 204a and 204b are part of a power grid that provides a supply voltage to the switching unit. Similarly, conductors 203a and 203b are portions of a ground reference grid that provides a true ground reference to the switch unit. A virtual ground node is placed within each switching unit (e.g., switching unit 101).

第3圖顯示佈置電源閘之一第二方法(「凹洞開關」),其中一列邏輯單元(如邏輯單元301a、301b、301c與301d)共用在相鄰之專用列(如「電源開關凹區」302)中被提供的電源開關。導體303a與303b提供真實接地電壓基準及導體304a與304b提供電源電壓基準。Figure 3 shows a second method of arranging a power gate ("cavity switch") in which a column of logic cells (such as logic cells 301a, 301b, 301c, and 301d) are shared in adjacent dedicated columns (such as "power switch recesses" The power switch provided in 302). Conductors 303a and 303b provide a true ground voltage reference and conductors 304a and 304b provide a supply voltage reference.

第4圖顯示佈置電源閘之一第三方法(「環開關」),其中一組開關單元(如開關單元401)共用以圓圈環繞開關單元之圓形條帶(如電源閘區402)中被佈置的電源開關。電源閘區402中之電源開關典型上以並聯被連接。電源閘區402外之導體路由真實接地至電源閘區402中之電源閘。虛擬接地基準節點在電源閘區402與邏輯單元401間被提供。Figure 4 shows a third method of arranging a power gate ("ring switch") in which a group of switch units (such as switch unit 401) share a circular strip (such as power gate region 402) surrounded by a circle around the switch unit. The power switch is arranged. The power switches in the power gate region 402 are typically connected in parallel. The conductors outside the power gate region 402 are routed to the power gates in the power gate region 402. A virtual ground reference node is provided between the power gate region 402 and the logic unit 401.

第5圖顯示佈置電源閘之一第四方法(「柵格開關」),其中電源開關(如在電源開關區502a與502b中之電源開關)以規律之間隔被佈置於預設的位置以服務其鄰近之多列邏輯單元(如邏輯單元501a與501b被電源開關502a中之電源開關服務)。在預設之電源開關區中的電源開關可以並聯被連接,且與各列開關單元正交地延伸之導體(如導體505a、505b與505c)可提供虛擬接地基準柵格。Figure 5 shows a fourth method of arranging a power supply gate ("grid switch") in which power switches (such as power switches in power switch areas 502a and 502b) are arranged at regular intervals to serve at regular intervals. Its adjacent multi-column logic cells (e.g., logic cells 501a and 501b are serviced by power switches in power switch 502a). Power switches in the predetermined power switch zone can be connected in parallel, and conductors (e.g., conductors 505a, 505b, and 505c) extending orthogonally to the respective column of switch cells can provide a virtual ground reference grid.

發明概要Summary of invention

一種設計方法在邏輯單元列之未被佔用的位置佈置電源閘或開關單元。這些電源閘不須全面地以並聯被連接。在一實施例(「充填開關」)中,對電源閘之佈置沒有限制。在另一實施例(「密封開關」)中,電源閘以相鄰於同一睡眠信號域之一組連續地被佈置的邏輯單元地被佈置。由於慣常之標準單元設計與佈置技術只達成70-80%之佈置密度或運用(在每一邏輯單元中可得可用之空間20-30%間的未被佔用的空間),藉由在未被佔用的空間中佈置該等電源閘單元,該方法不會提高整體矽不動產需求,就算該等電源閘單元被引進該設計亦然。佈置最佳化技術可被施用以達成電源閘單元之適當大小與分佈,而避免因電源閘單元所致的效能處罰。在一實施例中,細粒狀之電源閘藉由選擇性地在電源閘邏輯單元中提供非電源閘邏輯單元而被達成。One design method places a power gate or switch unit in an unoccupied position of the logic cell column. These power gates do not have to be fully connected in parallel. In an embodiment ("fill switch"), there is no limit to the arrangement of the power gates. In another embodiment ("sealed switch"), the power gates are arranged in logic units that are successively arranged adjacent to one of the same sleep signal domains. Due to the customary standard cell design and layout technology, only 70-80% of the layout density or the use (the unoccupied space between 20-30% of the available space in each logical unit) is achieved. Arranging the power gate units in the occupied space does not increase the overall real estate demand, even if the power gate units are introduced into the design. Arrangement optimization techniques can be applied to achieve the proper size and distribution of the power gate unit while avoiding performance penalties due to the power gate unit. In one embodiment, the fine grained power gate is achieved by selectively providing a non-power gate logic unit in the power gate logic unit.

本發明可應用於佈置連接一虛擬接地電壓基準至真實接地之電源閘單元及佈置連接一虛擬電源電壓基準至真實電源電壓基準之電源閘單元。The invention can be applied to a power supply gate unit that connects a virtual ground voltage reference to a true ground and a power supply gate unit that connects a virtual power supply voltage reference to a true supply voltage reference.

在使用這些充填開關下,對邏輯單元之虛擬接地電壓基準可使用慣常的邏輯信號路由技術被路由。這些虛擬接地基準可使用路由通道、使用無通道路由技術(如在一個或多個信號路由層中之邏輯單元上)、或使用共用開關接地匯流排被路由。本發明之方法可以任何信號路由技術被使用。With these fill switches, the virtual ground voltage reference to the logic cells can be routed using conventional logic signal routing techniques. These virtual ground references can be routed using routing channels, using channelless routing techniques (such as on logical units in one or more signal routing layers), or using a common switch ground bus. The method of the present invention can be used with any signal routing technique.

依據本發明之一實施例,一積體電路可包括一電源電壓基準、一接地電壓基準、以線性組構被佈置之邏輯單元以形成多列之單元、與在各列單元內被提供之邏輯單元單元。在一實施例中,路由通道中之導體被用於以與邏輯信號在邏輯單元間依慣常之標準單元設計方法下被路由相同的方式來路由邏輯單元與電源閘間之虛擬接地信號。In accordance with an embodiment of the present invention, an integrated circuit can include a supply voltage reference, a ground voltage reference, logic cells arranged in a linear configuration to form a plurality of columns of cells, and logic provided in each column of cells Unit unit. In one embodiment, the conductors in the routing channel are used to route the virtual ground signal between the logic unit and the power gate in the same manner as the logic signal is routed between the logic cells in accordance with conventional standard cell design methods.

在一實施例中,每一具有不同個數之電晶體以並聯組構被連接之電源閘單元可在一設計館中被提供。適當地被選擇之電源閘單元限制所需之電路不動產與通過電源閘單元之電壓下降。In one embodiment, each of the power gate units having a different number of transistors connected in parallel can be provided in a design library. A properly selected power gate unit limits the required circuit real estate to the voltage drop through the power gate unit.

在一實施例中,電源閘式邏輯單元(即直接被連接至真實接地電壓基準之開關單元)被佈置。非電源閘式邏輯單元可在被需要之處達成較高之電路效能。在此方式下,高電路效能在以非常小量之額外洩漏電流的權衡下被達成。In an embodiment, the power gate logic unit (ie, the switch unit that is directly connected to the true ground voltage reference) is arranged. Non-power gate logic cells can achieve higher circuit performance where needed. In this manner, high circuit performance is achieved with a trade-off of a very small amount of additional leakage current.

在一實施例中,電源閘單元被佈置於其內之未被佔用的空間以多少為隨機之方式發生,使得電源閘單元以多少為不規律之方式被佈置。最佳化技術可被施用以降低電源閘單元分佈之不規律性。此類最佳化技術可包括調整要使用之電源閘單元的數目、每一電源閘單元之大小與電源閘單元的被選擇之位置。In an embodiment, the unoccupied space in which the power gate unit is disposed is generated in a random manner such that the power gate unit is arranged in an irregular manner. Optimization techniques can be applied to reduce irregularities in the distribution of the power supply gate units. Such optimization techniques may include adjusting the number of power gate units to be used, the size of each power gate unit, and the selected location of the power gate unit.

依據本發明之一實施例,一種用於佈置電源閘單元之方法包括首先以列佈置邏輯單元。在這些被佈置之邏輯單元中為將被作成電源閘之邏輯單元,其依據個具有一接頭用於耦合一虛擬接地電壓基準。然後電源閘單元以列被佈置於未被佔用的空間內及虛擬接地電壓節點依據所採用之路由技術在邏輯單元與電源閘單元間被路由。用於承載虛擬接地電壓基準之導體的寬度依據配線電阻之估計被提供。In accordance with an embodiment of the present invention, a method for arranging power supply gate units includes first arranging logic units in columns. Among the arranged logic units is a logic unit to be a power gate, which has a connector for coupling a virtual ground voltage reference. The power gate unit is then arranged in columns in the unoccupied space and the virtual ground voltage node is routed between the logic unit and the power gate unit in accordance with the routing technique employed. The width of the conductor used to carry the virtual ground voltage reference is provided in accordance with an estimate of the wiring resistance.

在一實施例中,在未被佔用的空間被佈置之每一電源閘單元的大小依據電源閘單元將被連接之邏輯單元中被估計的電流被計算。該等電流使用靜態或動態技術被估計。在另一實施例中,被佈置之每一電源閘單元的大小可使用將被連接之邏輯單元中的電晶體數目被決定。In an embodiment, the size of each of the power gate units that are disposed in the unoccupied space is calculated based on the estimated current in the logic unit to which the power gate unit is to be connected. These currents are estimated using static or dynamic techniques. In another embodiment, the size of each power gate unit that is arranged can be determined using the number of transistors in the logic unit to be connected.

在一實施例中,在佈置電源閘單元前,一佈置最佳化步驟被執行以重新決定未被佔用的空間之大小,例如藉由沿著橫向方向移除每一列內之邏輯單元,或藉由以縱向方向移除穿過各列之邏輯單元。In an embodiment, an arrangement optimization step is performed to re-determine the size of the unoccupied space before arranging the power gate unit, for example by removing the logic unit in each column along the lateral direction, or by The logic cells passing through the columns are removed by the longitudinal direction.

在一實施例中,在佈置電源閘單元前,一佈置最佳化步驟被執行以重新安排邏輯單元而達成未被佔用的空間之特定分佈。該特定分佈可藉由定出介於二電源閘單元間之每一列的特定目標之距離而被達成。In an embodiment, prior to arranging the power gate unit, an arrangement optimization step is performed to rearrange the logic unit to achieve a particular distribution of unoccupied space. This particular distribution can be achieved by determining the distance to a particular target for each column between the two power gate units.

本發明在對下面詳細之描述與附圖考量下被較佳地了解。The invention will be better understood from the following detailed description and drawings.

圖式簡單說明Simple illustration

第1圖示意地顯示控制之邏輯單元102對接地的洩漏電流路徑之電源閘或開關單元101。Figure 1 schematically shows the power gate or switching unit 101 of the leakage current path of the controlled logic unit 102 to ground.

第2圖顯示集積邏輯單元202之邏輯單元201,邏輯單元201可使用慣常的標準單元設計方法被佈置。Figure 2 shows the logic unit 201 of the accumulation logic unit 202, which can be arranged using conventional standard cell design methods.

第3圖顯示用於佈置邏輯單元之一第二方法,其中一列邏輯單元(如邏輯單元301a,301b,301c與301d)共用在相鄰專用列(「開關凹區」,如電源開關列302)中被提供之電源開關。Figure 3 shows a second method for arranging logic cells in which a column of logic cells (e.g., logic cells 301a, 301b, 301c, and 301d) are shared in adjacent dedicated columns ("switch recesses", such as power switch column 302). The power switch is provided in the middle.

第4圖顯示用於佈置邏輯單元之一第三方法,其中一組邏輯單元(如邏輯單元401)共用以圓圈圍繞邏輯單元之圓形條帶(如邏輯單元區402)中被佈置的電源開關。Figure 4 shows a third method for arranging logic cells, wherein a group of logic cells (e.g., logic cells 401) share a power switch disposed in a circular strip (e.g., logic cell region 402) surrounding the logic cells in a circle. .

第5圖顯示用於佈置邏輯單元之一第四方法(「柵格開關」),其中電源開關(如電源開關區502a與502b中之電源開關)以相對於各列開關單元之規律間隔在預設位置被佈置。Figure 5 shows a fourth method ("grid switch") for arranging logic cells in which power switches (such as power switches in power switch zones 502a and 502b) are spaced at regular intervals relative to the columns of switch cells. Let the position be arranged.

第6a與6b圖顯示依據本發明之一實施例的一設計方法(「充填開關」)。Figures 6a and 6b show a design method ("fill switch") in accordance with an embodiment of the present invention.

第6c圖顯示依據本發明之一實施例的例如邏輯單元601a與充填開關單元602a之結構。Figure 6c shows the structure of, for example, logic unit 601a and fill switch unit 602a in accordance with an embodiment of the present invention.

第7圖顯示使用依據本發明之一實施例的密封開關。Figure 7 shows the use of a sealed switch in accordance with an embodiment of the present invention.

較佳實施例之詳細說明Detailed description of the preferred embodiment

本發明提供一種設計方法,其中邏輯單元或開關單元在邏輯單元列之未被佔用的位置被佈置。本發明人觀察到使用慣常之標準單元設計與佈置技術,佈置密度或運用典型上可被達成70-80%之間(即未被佔用的空間構成每一列之邏輯單元的20-30%間之可得可用的空間)。本發明在未被佔用的空間中佈置電源閘單元。可得可用之未被佔用的空間典型上提供足夠的空間以容納在一列邏輯單元中所需之全部電源閘單元。因而,在邏輯電路中納入電源閘不會提高矽之不動產需求。The present invention provides a design method in which a logic unit or a switching unit is arranged at an unoccupied position of a column of logic cells. The inventors have observed that using conventional standard cell design and placement techniques, the placement density or utilization can typically be achieved between 70-80% (i.e., unoccupied space constitutes between 20-30% of the logic cells of each column). Available space). The present invention arranges the power gate unit in an unoccupied space. The available unoccupied space typically provides sufficient space to accommodate all of the power gate units required in a column of logic cells. Therefore, the inclusion of a power gate in a logic circuit does not increase the real estate demand.

在此詳細之描述中,本發明用電源閘單元連接邏輯單元至真實接地電壓基準之實施例被說明。然而,本發明可相等地應用於連接邏輯單元至真實電源電壓基準之電源閘單元(在此情形中,連接電源閘至邏輯單元之接頭被稱為「虛擬電源電壓基準」)。本發明亦可應用於連接邏輯單元至一真實接地電壓基準之邏輯單元與連接邏輯單元至真實電源電壓基準之邏輯單元的設計二者均被使用。In this detailed description, the invention is illustrated with an embodiment in which the power supply gate unit is connected to the logic unit to a true ground voltage reference. However, the present invention is equally applicable to a power supply gate unit that connects a logic unit to a real power supply voltage reference (in this case, a connector that connects the power supply gate to the logic unit is referred to as a "virtual power supply voltage reference"). The invention can also be applied to both the logic unit connecting the logic unit to a true ground voltage reference and the design of the logic unit connecting the logic unit to the true supply voltage reference.

第6a與6b圖顯示依據本發明之一實施例的一設計方法(「充填開關」)。在充填開關設計方法下,充填開關(即電源閘單元)被設計以具有與設計館中之邏輯單元相同的高度與外形,使得充填開關單元以與邏輯單元被佈置於此列中相同之方式被佈置於慣常之邏輯單元列中。如第6a圖中顯示者,組構600代表為慣常之標準單元設計技術被設計的邏輯單元間路由信號被保留之區域的空間(「通道」)隔離的三列邏輯單元。Figures 6a and 6b show a design method ("fill switch") in accordance with an embodiment of the present invention. In the filling switch design method, the filling switch (ie, the power gate unit) is designed to have the same height and shape as the logic unit in the design library, such that the filling switch unit is in the same manner as the logic unit is arranged in this column. Arranged in the usual logical unit column. As shown in Figure 6a, fabric 600 represents a three-column logic unit isolated by the space ("channel") of the area in which the inter-logic routing signals are reserved for conventional standard cell design techniques.

組構600使用慣常之佈置技術但未佈置任何充填開關地被達成。在一實施例中,邏輯單元與充填開關之佈置使用分離的工具發生。在此實施例中,邏輯單元之佈置用不知道提供該等充填開關之後續步驟的一佈置工具被實施。在另一實施例中,邏輯單元與充填開關之佈置可使用同一工具被實施,但以邏輯單元之佈置在充填開關被排除的一第一階段被實施。排除充填開關可藉由在佈置成本函數中提供適當之權重給充填開關或藉由在此特別佈置步驟被使用之設計館排除充填開關而被達成。在初始佈置中,邏輯單元可依據其「睡眠信號域」(即被同一控制信號控制之將被連接至充填開關單元的邏輯單元)而被佈置。一邏輯單元之睡眠域可藉由從開關單元之控制信號向後透過緩衝器開始並序列地連接各對反相器至控制信號之源頭地橫越連線清單(Netlist)而被定出。Fabric 600 is achieved using conventional placement techniques without the placement of any fill switches. In an embodiment, the arrangement of the logic unit and the fill switch occurs using a separate tool. In this embodiment, the arrangement of the logic units is implemented with an arrangement tool that is unaware of the subsequent steps of providing the fill switches. In another embodiment, the arrangement of the logic unit and the fill switch can be implemented using the same tool, but with the arrangement of the logic units being implemented in a first stage in which the fill switch is eliminated. Excluding the fill switch can be achieved by providing the appropriate weights in the placement cost function to the fill switch or by eliminating the fill switch by the design hall in which the particular arrangement step is used. In the initial arrangement, the logic unit can be arranged according to its "sleep signal domain" (ie, the logic unit that is controlled by the same control signal to be connected to the fill switch unit). The sleep domain of a logic unit can be determined by traversing the list of nets from the control signal of the switching unit to the back through the buffer and serially connecting the pairs of inverters to the source of the control signal.

如第6a圖顯示者,導體603a與603b以慣常之方式提供電源電壓基準。類似地,導體604a與604b為所顯示之各列邏輯單元以慣常方式提供真實接地電壓基準。典型地如第6a圖顯示者,該佈置法則僅能在每一列邏輯單元充填70-80%之可得可用的空間,留下如間隙605與606之間隙。在佈置法則能達成較高運用率而就佈置充填開關留下之未被佔用的空間為不足夠的稀有情形中,該佈置法則可被導向於不以較低運用程度之可得可用的空間為目標,而為充填開關單元允許最少量之未被佔用的空間。As shown in Figure 6a, conductors 603a and 603b provide a supply voltage reference in a conventional manner. Similarly, conductors 604a and 604b provide a true ground voltage reference in a conventional manner for each column of logic cells shown. Typically, as shown in Figure 6a, the arrangement can only fill 70-80% of the available space in each column of logic, leaving gaps such as gaps 605 and 606. In the rare case where the arrangement law can achieve a higher utilization rate and the unoccupied space left by the placement of the filling switch is insufficient, the arrangement can be directed to a space that is not available at a lower level of operation. The target, while filling the switch unit allows a minimum amount of unoccupied space.

第6c圖例如顯示依據本發明之一實施例邏輯單元601a與充填開關單元602a的結構。如在第6a圖顯示者,邏輯單元601a與充填開關單元602a二者均為真實接地與真實電源電壓基準以延伸穿過其各別之頂端與底部部位(即部位624與626)。導體623在邏輯單元601a中被提供用於連接一邏輯輸入信號。類似地,充填開關單元602a中之導體625被提供用於連接電源閘之睡眠控制信號。導體622與導體620在邏輯單元601a與充填開關單元602a中分別被提供以連接虛擬接地電壓基準。該連接例如可被導體624使用適合之路由技術方法加以生效。Figure 6c shows, for example, the structure of logic unit 601a and fill switch unit 602a in accordance with one embodiment of the present invention. As shown in Figure 6a, both logic unit 601a and fill switch unit 602a are both true ground and true supply voltage references to extend through their respective top and bottom portions (i.e., locations 624 and 626). Conductor 623 is provided in logic unit 601a for connecting a logic input signal. Similarly, conductor 625 in fill switch unit 602a is provided with a sleep control signal for connecting the power gate. Conductor 622 and conductor 620 are respectively provided in logic unit 601a and fill switch unit 602a to connect a virtual ground voltage reference. This connection can be effected, for example, by conductor 624 using a suitable routing technique.

依據本發明之一實施例,該設計館包括不同大小與不同定向之數個充填開關單元。在下一步驟中,邏輯單元列中未被佔用的空間(「間隙」)被定出。就每一間隙中,一個或多個充填開關單元被佈置。被佈置之充填開關單元的大小例如可由被指定給充填開關之邏輯單元的電流估計被決定。被指定給充填開關之邏輯單元的電流可使用靜態或動態估計技術被估計。適合的電流估計技術包括電流平均或在2003年12月17日申請之美國專利申請案第10/739,659號的“Current Scheduling System and Method For Optimizing Multi-Threshold CMOS Design”與在2002年10月1日申請及在2004年10月19日被發給之美國專利第6,807,660號的“Vectorless Instantaneous Current Estimation”中所揭示之任一動態電流估計技術。被佈置之充填開關單元的大小亦可用充填開關單元可被指定至附近的邏輯單元電晶體的個數被決定。使用任何這些與其他適合之方法,為每一組邏輯單元之電源閘用的所需之必要的累積裝置大小乃被決定。根據此所需之大小,適當的充填開關單元被選擇及在間隙中被佈置。被佈置之充填開關單元的定向〔註1〕被選擇以使虛擬接地淨阻抗最小化。第6b圖顯示結果之組構610。例如,第6a圖之間隙605與606分別用充填開關單元602a與602b被替換。According to an embodiment of the invention, the design hall comprises a plurality of filling switch units of different sizes and different orientations. In the next step, the unoccupied space ("gap") in the logical unit column is determined. In each gap, one or more fill switch units are arranged. The size of the arranged fill switch unit can be determined, for example, by a current estimate assigned to the logic unit filling the switch. The current assigned to the logic unit filling the switch can be estimated using static or dynamic estimation techniques. Suitable current estimation techniques include current averaging or "Current Scheduling System and Method For Optimizing Multi-Threshold CMOS Design" of U.S. Patent Application Serial No. 10/739,659, filed on Dec. 17, 2003. Any of the dynamic current estimation techniques disclosed in "Vectorless Instantaneous Current Estimation" of U.S. Patent No. 6,807,660, issued Oct. 19, 2004. The size of the arranged fill switch unit can also be determined by the number of transistors that can be assigned to nearby logic cells by the fill switch unit. The use of any of these and other suitable methods is necessary to determine the necessary cumulative device size for each group of logic cells. Depending on the required size, a suitable filling switch unit is selected and arranged in the gap. The orientation of the placed fill switch unit (Note 1) is selected to minimize the virtual ground net impedance. Figure 6b shows the resulting composition 610. For example, gaps 605 and 606 of Figure 6a are replaced with fill switch units 602a and 602b, respectively.

[註1]具有某一特定電流能力之充填開關單元的多種版本可被提供(如一充填開關單元與其鏡影像)。在一些實施例中,一版本針對另一版本之選擇可形成虛擬接地信號中較低電阻之結果。該差異例如可因用於連接虛擬接地電壓基準之導體位置所致。[Note 1] Various versions of the filling switch unit with a certain current capability can be provided (such as a filling switch unit and its mirror image). In some embodiments, the selection of one version for another version may result in a lower resistance in the virtual ground signal. This difference can be due, for example, to the location of the conductor used to connect the virtual ground voltage reference.

用於根據一虛擬接地網所服務之充填開關單元的位置與邏輯單元之位置用於估計該虛擬接地網的配線電阻之一精確的全面路由模型為用於指導佈置的一較佳工具。在一實施例中,在最佳化之際,充填開關以電流需求之下降順序被指定充填開關(即具有最高電流需求之邏輯單元首先被指定讓充填開關),使得具有最高電流需求之邏輯單元最靠近其被佈置。服務同一邏輯單元或同組邏輯單元之充填開關以並聯被連接以提供單一虛擬接地網,而符合具有較高電流需求之邏輯單元的虛擬接地臨界電壓及使用於充填開關單元佈置之可得可用的空間運用最大化(即服務同一邏輯單元或同組邏輯單元之充填開關可在同一鄰近區中之數個小間隙上分佈)。同樣地,充填開關單元佈置亦可被配線界限規格限制而限制在邏輯單元之真實接地電壓基準與真實接地基準間的電壓下降。該配線界限規格亦可使對結果所得之虛擬接地網的期望電遷移影響最小化。A comprehensive overall routing model for estimating the wiring resistance of the virtual ground grid based on the position of the filling switch unit served by a virtual grounding grid and the location of the logic unit is a preferred tool for guiding the arrangement. In one embodiment, at the time of optimization, the fill switch is assigned a fill switch in the order of decreasing current demand (ie, the logic unit with the highest current demand is first designated to fill the switch), so that the logic unit has the highest current demand. It is placed closest to it. Filling switches that serve the same logical unit or groups of logical units are connected in parallel to provide a single virtual grounding grid, while meeting the virtual grounding threshold voltage of the logic unit with higher current requirements and available for use in the filling switch unit arrangement Maximizing space utilization (ie, filling switches that serve the same logical unit or groups of logical units can be distributed over several small gaps in the same neighborhood). Similarly, the fill switch unit arrangement can also be limited by the wiring limit specification to limit the voltage drop between the true ground voltage reference of the logic unit and the true ground reference. This wiring limit specification also minimizes the effect of the desired electromigration of the resulting virtual ground grid.

本發明亦可以「局部電源閘」技術被使用。在此技術中,充填開關單元就算共用同一睡眠信號域或同一真實接地連接,不會以並聯被連接。The invention can also be used with "local power gate" technology. In this technique, the filling switch units do not share in parallel if they share the same sleep signal domain or the same true ground connection.

例如包括藉由以橫向方向移除邏輯單元來重新調整間隙大小之佈置最佳化技術可被用以允許被要求之充填開關配適於間隙內。在適當的情形中,邏輯單元亦可在列間被移動或被交換以提供充填開關所用之間隙或使之最佳化。該等間隙亦可被重新定大小或重新定位,使得替換間隙之充填開關單元可被限制為特定大小或以目標分佈方式分佈。允許在一列內重新安排邏輯單元佈置或在列間移動邏輯單元之最佳化步驟可被使用以允許充填開關單元之適當分佈。充填單元之此重新分佈例如可藉由避免指定充填開關單元以比充填開關單元為此被指定地服務較多個數之開關單元電晶體及因而影響時機來使效能最佳化。一般而言,其欲使充填開關單元儘可能接近其服務的邏輯單元地被佈置。用於指定邏輯單元至充填單元之一適合的成本函數可考慮虛擬接地電壓基準與真實接地基準間之使用指定的距離限制、配線電阻、估計之電流、期望之電壓下降與關切之信號整合性。其亦欲限制在充填開關單元佈置之際的邏輯單元之運動。在此實施例中,由於充填開關單元之佈置係依邏輯單元之初始佈置而定,充填開關單元因而以相當隨機地及以不規律之方式被佈置。替選的是,邏輯單元之佈置可在決定充填開關單元之大小後被調整,使得充填開關單元以較規律之方式被佈置。在充填開關單元被佈置後,其運動應被限制。An arrangement optimization technique including, for example, re-adjusting the gap size by removing the logic cells in a lateral direction can be used to allow the required fill switch to fit within the gap. Where appropriate, the logic unit can also be moved or exchanged between columns to provide or optimize the gap used to fill the switch. The gaps may also be resized or repositioned such that the fill switch units that replace the gaps may be limited to a particular size or distributed in a target distribution. An optimization step that allows rearrangement of logical unit arrangements within a column or movement of logical units between columns can be used to allow for proper distribution of the filled switch units. This redistribution of the filling unit can be optimized, for example, by avoiding the designation of the filling switch unit to serve more than a certain number of switching unit transistors for this purpose and thus affecting the timing. In general, it is intended to arrange the filling switch unit as close as possible to the logic unit it serves. A suitable cost function for specifying one of the logic cells to the fill cells may take into account the distance limit specified by the use of the virtual ground voltage reference and the true ground reference, the wiring resistance, the estimated current, the desired voltage drop, and the signal integrity of the concern. It also intends to limit the movement of the logic unit at the time of filling the switching unit arrangement. In this embodiment, since the arrangement of the filling switch unit is dependent on the initial arrangement of the logic unit, the filling switch unit is thus arranged in a relatively random and irregular manner. Alternatively, the arrangement of the logic units can be adjusted after determining the size of the filling switch unit such that the filling switch unit is arranged in a more regular manner. After the filling switch unit is placed, its movement should be limited.

該佈置在可能時可藉由重新定被佈置之充填開關單元的大小以補償因充填開關單元與邏輯單元之運動所形成的被估計之互連電阻與真實互連電阻間的任何變異而進一步地被最佳化。The arrangement may further compensate for any variation between the estimated interconnect resistance formed by the movement of the fill switch unit and the logic unit and the true interconnect resistance by reconfiguring the size of the filled switch unit when possible Be optimized.

然後適合之信號路由技術可被使用以在通道中為充填開關路由虛擬接地電壓基準、邏輯單元間之電源電壓基準信號與控制信號。因而,依據本發明之一實施例,就虛擬接地電壓基準而言不需有專用之路由通道。被用以路由虛擬接地電壓基準之導體的寬度可依據被估計之電流與配線電阻被調整,而使在有作用之邏輯電路作業之際對切換速度的影響最小化。為使虛擬接地網中之電壓下降最小化,寬的配線可就虛擬接地網被使用。如上面提及者,本發明不專用於任一路由技術之使用。因而,可被施用於使用無通道路由技術(如在單元上路由)之技術。在另一實施例中,共用之虛擬接地匯流排可被提供。A suitable signal routing technique can then be used to route the virtual ground voltage reference, the supply voltage reference signal and the control signal between the logic cells for the fill switch in the channel. Thus, in accordance with an embodiment of the present invention, there is no need for a dedicated routing channel for a virtual ground voltage reference. The width of the conductor used to route the virtual ground voltage reference can be adjusted based on the estimated current and wiring resistance to minimize the effects on switching speed when the active logic circuit is operating. To minimize the voltage drop in the virtual ground grid, a wide wiring can be used for the virtual ground grid. As mentioned above, the invention is not specific to the use of any routing technology. Thus, techniques can be applied to the use of channelless routing techniques, such as routing on a cell. In another embodiment, a shared virtual ground bus can be provided.

雖然充填開關單元之佈置未被限制,本發明亦提供一種方法,其對何處可佈置電源閘單元加以要求。此方法(「密封開關」)要求電源閘單元(「密封開關」)在同一睡眠域之一組連續地被佈置的電源閘式邏輯單元的一或二端部被佈置。第7圖顯示依據本發明之一實施例使用密封開關單元。如第7圖顯示者,組構700包括電源閘式邏輯單元(如邏輯單元701與702)與非電源閘式邏輯單元(如邏輯單元707)二者。與充填開關方法對照下,密封開關單元在同一睡眠域之一組連續地被佈置的電源閘式邏輯單元的一或二端部被佈置。在此例中,密封開關單元706a與706b鄰接邏輯單元701與702被提供。若整列邏輯單元(如邏輯單元列703)在同一睡眠信號域中,密封開關可在該列之一或二端部被佈置(如密封開關單元705a與705b在邏輯單元列703之端部被佈置)。Although the arrangement of the filling switch unit is not limited, the present invention also provides a method for where a power supply gate unit can be placed. This method ("sealed switch") requires that the power lock unit ("sealed switch") be placed at one or both ends of a power gate logic unit that is continuously arranged in one of the same sleep domains. Figure 7 shows the use of a sealed switch unit in accordance with an embodiment of the present invention. As shown in FIG. 7, fabric 700 includes both power gate logic cells (such as logic cells 701 and 702) and non-power gate logic cells (such as logic cell 707). In contrast to the filling switch method, the sealed switch unit is arranged at one or both ends of a power gate logic unit that is continuously arranged in one of the same sleep domains. In this example, sealed switch units 706a and 706b are provided adjacent to logic units 701 and 702. If the entire column of logic cells (e.g., logic cell column 703) is in the same sleep signal domain, the sealed switch can be placed at one or both ends of the column (e.g., sealed switch cells 705a and 705b are arranged at the end of logic cell column 703) ).

第7圖亦顯示列703之邏輯單元703-n與密封開關單元705b之內部結構。如第7圖中顯示者,導體(如導體723與725)被提供用於分別連接邏輯信號與一睡眠控制信號至邏輯單元703-n與密封開關單元705b。不像充填開關單元的是,虛擬接地電壓基準721以與真實接地及真實電源電壓基準相同之方式被提供。在此方式中,一邏輯單元與其關聯之密封開關單元間的連接在佈置之際生效,而迴避後續之路由步驟。上面針對充填開關單元被討論之最佳化技術可被用以重新安排邏輯單元與間隙而為密封開關單元作出空間。Fig. 7 also shows the internal structure of the logic unit 703-n of column 703 and the sealed switch unit 705b. As shown in Figure 7, conductors (e.g., conductors 723 and 725) are provided for connecting logic signals and a sleep control signal to logic unit 703-n and sealed switch unit 705b, respectively. Unlike the filling switch unit, the virtual ground voltage reference 721 is provided in the same manner as the true ground and true supply voltage reference. In this manner, the connection between a logic unit and its associated sealed switch unit takes effect at the time of placement, bypassing the subsequent routing steps. The optimization techniques discussed above for the fill switch unit can be used to rearrange the logic cells and gaps to make room for the sealed switch unit.

除了對密封開關相對於其關聯之邏輯單元的佈置限制外,密封開關達成與上面討論之充填開關實質相同的彈性與益處。In addition to the arrangement constraints on the sealed switch relative to its associated logic unit, the sealed switch achieves substantially the same resiliency and benefits as the fill switch discussed above.

依據本發明之一實施例的用於佈置充填開關與密封開關及其最佳化技術可在信號路由前、信號路由後、或二者均是地被實施。The technique for arranging the fill switch and the sealed switch and its optimization according to an embodiment of the present invention can be implemented before signal routing, after signal routing, or both.

下列之表比較在此說明書所討論的用於MTCMOS之各種設計技術的特徵: The following table compares the features of the various design techniques discussed in this specification for MTCMOS:

根據上表所評估之特徵,本發明人得到充填開關方法比起習知技藝具有下列優點之結論: Based on the characteristics evaluated in the above table, the inventors have come to the conclusion that the filling switch method has the following advantages over the prior art:

上面之詳細描述被提供以說明本發明之特定實施例且不欲被限制。在本發明之領域內的很多修改與變化為可能的。本發明在下列之如申請專利範圍中被設立。The above detailed description is provided to illustrate the specific embodiments of the invention Many modifications and variations are possible in the field of the invention. The invention is set forth in the following patent claims.

101...電源閘,開關單元101. . . Power brake

102...開關單元102. . . Switch unit

103...虛擬接地節點103. . . Virtual ground node

104...電源節點104. . . Power node

105...真實接地節點105. . . Real ground node

106...控制節點106. . . Control node

201...開關單元201. . . Switch unit

202...電源閘202. . . Power brake

203a...接地基準柵格203a. . . Ground reference grid

203b...接地基準柵格203b. . . Ground reference grid

204a...導體204a. . . conductor

204b...導體204b. . . conductor

301a-301d...邏輯單元301a-301d. . . Logical unit

302...電源開關凹區302. . . Power switch recess

303a...導體303a. . . conductor

303b...導體303b. . . conductor

304a...導體304a. . . conductor

304b...導體304b. . . conductor

401...邏輯單元401. . . Logical unit

402...電源閘區402. . . Power gate

501a...邏輯單元501a. . . Logical unit

501b...邏輯單元501b. . . Logical unit

502a...電源開關區502a. . . Power switch area

502b...電源開關區502b. . . Power switch area

503a-503d...導體503a-503d. . . conductor

505a-505d...導體505a-505d. . . conductor

600...組構600. . . Fabrication

601a-601c...邏輯單元601a-601c. . . Logical unit

602a-602b...充填開關單元602a-602b. . . Filling switch unit

603a-603b...導體603a-603b. . . conductor

604a-604b...導體604a-604b. . . conductor

605...間隙605. . . gap

606...間隙606. . . gap

610...組構610. . . Fabrication

620...導體620. . . conductor

621...虛擬接地621. . . Virtual ground

622...導體622. . . conductor

623...導體623. . . conductor

624...部位624. . . Part

625...導體625. . . conductor

626...部位626. . . Part

700...組構700. . . Fabrication

701...邏輯單元701. . . Logical unit

702...邏輯單元702. . . Logical unit

703...邏輯單元列703. . . Logical unit column

703a,703b...導體703a, 703b. . . conductor

704a,704b...導體704a, 704b. . . conductor

705a,705b...密封開關單元705a, 705b. . . Sealed switch unit

706a,706b...密封開關單元706a, 706b. . . Sealed switch unit

707...邏輯單元707. . . Logical unit

721...虛擬接地電壓基準721. . . Virtual ground voltage reference

723...導體723. . . conductor

725...導體725. . . conductor

第1圖示意地顯示控制之邏輯單元102對接地的洩漏電流路徑之電源閘或開關單元101。Figure 1 schematically shows the power gate or switching unit 101 of the leakage current path of the controlled logic unit 102 to ground.

第2圖顯示集積邏輯單元202之邏輯單元201,邏輯單元201可使用慣常的標準單元設計方法被佈置。Figure 2 shows the logic unit 201 of the accumulation logic unit 202, which can be arranged using conventional standard cell design methods.

第3圖顯示用於佈置邏輯單元之一第二方法,其中一列邏輯單元(如邏輯單元301a,301b,301c與301d)共用在相鄰專用列(「開關凹區」,如電源開關列302)中被提供之電源開關。Figure 3 shows a second method for arranging logic cells in which a column of logic cells (e.g., logic cells 301a, 301b, 301c, and 301d) are shared in adjacent dedicated columns ("switch recesses", such as power switch column 302). The power switch is provided in the middle.

第4圖顯示用於佈置邏輯單元之一第三方法,其中一組邏輯單元(如邏輯單元401)共用以圓圈圍繞邏輯單元之圓形條帶(如邏輯單元區402)中被佈置的電源開關。Figure 4 shows a third method for arranging logic cells, wherein a group of logic cells (e.g., logic cells 401) share a power switch disposed in a circular strip (e.g., logic cell region 402) surrounding the logic cells in a circle. .

第5圖顯示用於佈置邏輯單元之一第四方法(「柵格開關」),其中電源開關(如電源開關區502a與502b中之電源開關)以相對於各列開關單元之規律間隔在預設位置被佈置。Figure 5 shows a fourth method ("grid switch") for arranging logic cells in which power switches (such as power switches in power switch zones 502a and 502b) are spaced at regular intervals relative to the columns of switch cells. Let the position be arranged.

第6a與6b圖顯示依據本發明之一實施例的一設計方法(「充填開關」)。Figures 6a and 6b show a design method ("fill switch") in accordance with an embodiment of the present invention.

第6c圖顯示依據本發明之一實施例的例如邏輯單元601a與充填開關單元602a之結構。Figure 6c shows the structure of, for example, logic unit 601a and fill switch unit 602a in accordance with an embodiment of the present invention.

第7圖顯示使用依據本發明之一實施例的密封開關。Figure 7 shows the use of a sealed switch in accordance with an embodiment of the present invention.

600...組構600. . . Fabrication

601a-601c...邏輯單元601a-601c. . . Logical unit

603a-603b...導體603a-603b. . . conductor

604a-604b...導體604a-604b. . . conductor

605...間隙605. . . gap

606...間隙606. . . gap

Claims (51)

一種積體電路,其包含:一導體用於提供一電壓基準;一列或多列邏輯單元,包括一第一團之邏輯單元,其每一邏輯單元具有一接頭用於耦合一虛擬電壓基準;以及一個或多個電源閘單元,其每一個具有一接頭用於耦合該虛擬電壓基準與一接頭用於耦合用於提供一電壓基準之該導體,該等數個電源閘單元在該等邏輯單元間被佈置,其中用於耦合該第一團之邏輯單元的虛擬電壓基準之該等接頭被連接至用於耦合該等電源閘單元之虛擬電壓基準的接頭。 An integrated circuit comprising: a conductor for providing a voltage reference; a column or columns of logic cells comprising a first group of logic cells, each logic cell having a connector for coupling a virtual voltage reference; One or more power supply gate units each having a connector for coupling the virtual voltage reference and a connector for coupling the conductor for providing a voltage reference between the plurality of power supply gate units Arranged wherein the joints of the virtual voltage references for coupling the logic elements of the first group are connected to a joint for coupling the virtual voltage references of the power gate units. 如申請專利範圍第1項所述積體電路,其中該第一團之邏輯單元連續地在一列中被佈置,及其中該等電源閘單元之一鄰接該第一團之邏輯單元中的該等邏輯單元之一被佈置。 The integrated circuit of claim 1, wherein the logic unit of the first group is continuously arranged in a column, and wherein one of the power gate units is adjacent to the logic unit of the first group One of the logical units is arranged. 如申請專利範圍第2項所述積體電路,其中在一預設位置之導體在鄰近的邏輯單元與電源閘單元二者均被提供,使得用於耦合該邏輯單元中之虛擬電壓基準的接頭因該佈置之性質被連接至用於耦合該電源閘單元之虛擬電壓基準的接頭。 The integrated circuit of claim 2, wherein the conductor at a predetermined position is provided in both the adjacent logic unit and the power gate unit, such that the connector for coupling the virtual voltage reference in the logic unit Because of the nature of this arrangement, it is connected to a connector for coupling the virtual voltage reference of the power gate unit. 如申請專利範圍第2項所述積體電路,其中該積體電路包含以數列被配置之邏輯單元,每一列包含位於該列之端部的電源閘單元。The integrated circuit of claim 2, wherein the integrated circuit comprises logic cells arranged in a series, each column comprising a power gate unit at an end of the column. 如申請專利範圍第2項所述積體電路,其中該積體電路包含以第一列與第二列被配置之邏輯單元,其中電源閘單元在該第一列之端部被提供,及邏輯單元在該第二列之端部被提供。The integrated circuit of claim 2, wherein the integrated circuit comprises a logic unit configured in a first column and a second column, wherein a power gate unit is provided at an end of the first column, and logic The unit is provided at the end of the second column. 如申請專利範圍第5項所述積體電路,其中該第一團之邏輯單元位於該第二列內。The integrated circuit of claim 5, wherein the logic unit of the first group is located in the second column. 如申請專利範圍第5項所述積體電路,進一步包含一第三列僅包含被連接至用於提供該虛擬電壓基準之導體。The integrated circuit of claim 5, further comprising a third column comprising only conductors connected to provide the virtual voltage reference. 如申請專利範圍第1項所述積體電路,其中該電壓基準為一接地電壓基準。The integrated circuit of claim 1, wherein the voltage reference is a ground voltage reference. 如申請專利範圍第1項所述積體電路,其中該電壓基準為一電源電壓基準。The integrated circuit of claim 1, wherein the voltage reference is a power supply voltage reference. 如申請專利範圍第1項所述積體電路,其中用於耦合該第一團之邏輯單元的虛擬電壓基準之接頭用在路由通道中之導體被連接至用於耦合該等電源閘單元之虛擬電壓基準的接頭。The integrated circuit of claim 1, wherein the connector for the virtual voltage reference for coupling the logic unit of the first group is used in a conductor in the routing channel to be connected to a virtual circuit for coupling the power gate units Voltage reference connector. 如申請專利範圍第1項所述積體電路,其中用於耦合該第一團之邏輯單元的虛擬電壓基準之接頭使用無通道路由技術被連接至用於耦合該等電源閘單元之虛擬電壓基準的接頭。The integrated circuit of claim 1, wherein the dummy voltage reference connector for coupling the logic unit of the first group is connected to a virtual voltage reference for coupling the power gate units using a channelless routing technique. Connector. 如申請專利範圍第1項所述積體電路,其中用於耦合該第一團之邏輯單元的虛擬電壓基準之接頭用共用路由通道中之導體被連接至用於耦合該等電源閘單元之虛擬電壓基準的接頭之虛擬電壓基準匯流排。The integrated circuit of claim 1, wherein the connector for the virtual voltage reference for coupling the logic unit of the first group is connected to the dummy for coupling the power gate unit with a conductor in the shared routing channel The virtual voltage reference bus of the voltage reference connector. 如申請專利範圍第1項所述積體電路,其中每一電源閘單元包含數個以並聯組構被連接之電晶體。 The integrated circuit of claim 1, wherein each of the power supply gate units comprises a plurality of transistors connected in a parallel configuration. 如申請專利範圍第1項所述積體電路,其中數個電源閘單元以並聯組構被連接。 The integrated circuit of claim 1, wherein the plurality of power supply gate units are connected in a parallel configuration. 如申請專利範圍第1項所述積體電路,其中一第二團之邏輯單元的每一個被連接至提供一電壓基準之導體。 The integrated circuit of claim 1, wherein each of the logic cells of the second group is connected to a conductor that provides a voltage reference. 如申請專利範圍第1項所述積體電路,其中該等電源閘單元以不規律方式在列中被佈置。 The integrated circuit of claim 1, wherein the power supply gate units are arranged in a column in an irregular manner. 如申請專利範圍第1項所述積體電路,其中該等電源閘在由該等邏輯單元之佈置結果所得的列內之間隙中被佈置。 The integrated circuit of claim 1, wherein the power gates are arranged in a gap in a column resulting from the arrangement of the logic cells. 一種用於佈置電源閘單元之方法,其包含:提供一導體用於承載一電壓基準;佈置一列或多列之數個邏輯單元,其中該等邏輯單元包括一第一團之邏輯單元,其每一邏輯單元具有一接頭用於耦合一虛擬電壓基準;以及佈置多個電源閘單元,其每一個具有一接頭用於耦合該虛擬電壓基準與一接頭用於耦合用於提供一電壓基準之該導體,該等數個電源閘單元在該等邏輯單元間被佈置。 A method for arranging a power gate unit, comprising: providing a conductor for carrying a voltage reference; arranging one or more columns of logic cells, wherein the logic cells comprise a first group of logic cells, each a logic unit having a connector for coupling a virtual voltage reference; and a plurality of power supply gate units each having a connector for coupling the virtual voltage reference to a connector for coupling the conductor for providing a voltage reference The plurality of power gate units are arranged between the logic units. 如申請專利範圍第18項所述方法,其中該第一團之邏輯單元連續地在一列中被佈置,及其中該等電源閘單元之一鄰接該第一團之邏輯單元中的該等邏輯單元之一被佈置。 The method of claim 18, wherein the logical unit of the first group is continuously arranged in a column, and wherein one of the power gate units is adjacent to the logic unit in the logic unit of the first group One is arranged. 如申請專利範圍第18項所述方法,其中在一預設位置之導體在鄰近的邏輯單元與電源閘單元二者均被提供,使得用於耦合該邏輯單元中之虛擬電壓基準的接頭因該佈置之性質被連接至用於耦合該電源閘單元之虛擬電壓基準的接頭。 The method of claim 18, wherein the conductor at a predetermined location is provided in both the adjacent logic unit and the power gate unit such that the connector for coupling the virtual voltage reference in the logic unit is The nature of the arrangement is connected to a joint for coupling the virtual voltage reference of the power gate unit. 如申請專利範圍第18項所述方法,其中該電壓基準為一接地電壓基準。 The method of claim 18, wherein the voltage reference is a ground voltage reference. 如申請專利範圍第18項所述方法,其中該電壓基準為一電源電壓基準。 The method of claim 18, wherein the voltage reference is a power supply voltage reference. 如申請專利範圍第18項所述方法,其中該等數個電源閘單元在由佈置該等邏輯單元結果所得之邏輯單元間的未被佔用的空間中被佈置。 The method of claim 18, wherein the plurality of power gate units are arranged in an unoccupied space between logical units resulting from the arrangement of the logic units. 如申請專利範圍第23項所述方法,其中在該未被佔用的空間中被佈置之每一電源閘單元的大小依據將被連接至該電源閘單元之該等邏輯單元中的電流之估計被計算。 The method of claim 23, wherein the size of each power gate unit disposed in the unoccupied space is determined by an estimate of the current to be connected to the logic cells of the power gate unit. Calculation. 如申請專利範圍第24項所述方法,其中該電流之估計使用靜態估計技術被實施。 The method of claim 24, wherein the estimation of the current is performed using a static estimation technique. 如申請專利範圍第24項所述方法,其中該電流之估計使用動靜態估計技術被實施。 The method of claim 24, wherein the estimation of the current is performed using a dynamic and static estimation technique. 如申請專利範圍第23項所述方法,其中在該未被佔用的空間中被佈置之每一電源閘單元的大小使用此電源閘單元被連接的該等邏輯單元中之電晶體的個數被計算。 The method of claim 23, wherein the size of each of the power gate units arranged in the unoccupied space is the number of transistors in the logic units to which the power gate unit is connected Calculation. 如申請專利範圍第18項所述方法,進一步包含路由該虛 擬電壓基準以連接用於耦合該第一團之邏輯單元的虛擬電壓基準之接頭至用於耦合該等電源閘單元之虛擬電壓基準的接頭。 The method of claim 18, further comprising routing the virtual A pseudo voltage reference is coupled to the connector of the virtual voltage reference for coupling the logic cells of the first group to a connector for coupling the virtual voltage references of the power gate units. 如申請專利範圍第28項所述方法,其中該路由係使用路由通道中之導體被實施。 The method of claim 28, wherein the routing is performed using a conductor in the routing channel. 如申請專利範圍第28項所述方法,其中該路由係使用非路由通道中之導體被實施。 The method of claim 28, wherein the routing is performed using a conductor in the non-routing channel. 如申請專利範圍第28項所述方法,其中該路由係使用共用虛擬電壓基準匯流排導體被實施。 The method of claim 28, wherein the routing is implemented using a shared virtual voltage reference busbar conductor. 如申請專利範圍第18項所述方法,進一步包含藉由沿著橫向方向移動每一列內之該等邏輯單元來重新決定未被佔用空間的大小。 The method of claim 18, further comprising re-determining the size of the unoccupied space by moving the logical units within each column in a lateral direction. 如申請專利範圍第18項所述方法,進一步包含藉由跨過列移動該等邏輯單元來重新決定未被佔用空間的大小。 The method of claim 18, further comprising re-determining the size of the unoccupied space by moving the logical units across the columns. 如申請專利範圍第18項所述方法,其中在佈置該等電源閘單元前,一佈置最佳化步驟被執行以重新安排該等邏輯單元而達成該一或多列中之未被佔用空間之一特定分佈。 The method of claim 18, wherein prior to arranging the power gate units, an arrangement optimization step is performed to rearrange the logic units to achieve unoccupied space in the one or more columns A specific distribution. 如申請專利範圍第34項所述方法,其中該特定分佈藉由以沿著二電源閘單元間之每一列的一特定距離為目標地被達成。 The method of claim 34, wherein the specific distribution is achieved by targeting a specific distance along each of the columns between the two power gate units. 如申請專利範圍第18項所述方法,進一步包含直接連接一第二團邏輯單元至用於提供該電壓基準之導體。 The method of claim 18, further comprising directly connecting a second group logic unit to the conductor for providing the voltage reference. 如申請專利範圍第18項所述方法,其中被用於該等邏輯 單元與該等電源閘單元間之連接的導體寬度係依據配線電阻之估計被提供。 Such as the method described in claim 18, which is used for such logic The conductor width of the connection between the unit and the power supply gate units is provided in accordance with an estimate of the wiring resistance. 如申請專利範圍第18項所述方法,其中具有共同睡眠域之該等邏輯單元以彼此緊密相鄰地被佈置。 The method of claim 18, wherein the logical units having a common sleep domain are arranged in close proximity to each other. 如申請專利範圍第38項所述方法,進一步包含橫越一連線清單(Netlist)以定出與每一邏輯單元相關聯之電源閘的一控制信號。 The method of claim 38, further comprising traversing a list of netlists to determine a control signal for the power gate associated with each of the logic units. 如申請專利範圍第39項所述方法,其中該橫越包括通過多對之緩衝器與反相器向後追蹤該控制信號。 The method of claim 39, wherein the traversing comprises tracking the control signal backwards through a plurality of pairs of buffers and inverters. 如申請專利範圍第18項所述方法,其中該等邏輯單元與該等電源閘單元係依據有關使用者所指定之距離的限制被佈置。 The method of claim 18, wherein the logic units and the power gate units are arranged in accordance with a limit on a distance specified by the user. 如申請專利範圍第18項所述方法,其中該等邏輯單元與該等電源閘單元係依據有關被估計之電壓下降的限制被佈置。 The method of claim 18, wherein the logic units and the power gate units are arranged in accordance with a limit on the estimated voltage drop. 如申請專利範圍第18項所述方法,其中該等邏輯單元與該等電源閘單元係依據有關所關切之信號整合性的限制被佈置。 The method of claim 18, wherein the logic units and the power gate units are arranged in accordance with restrictions on signal integration of concern. 如申請專利範圍第18項所述方法,其中每一電源閘單元依據其根據該電源閘單元之佈置與該電源閘之關聯的邏輯單元的佈置估計配線電阻之一全面路由模型被佈置。 The method of claim 18, wherein each of the power gate units is arranged in accordance with an arrangement of logic units in accordance with the arrangement of the power gate unit and the logic unit associated with the power gate. 如申請專利範圍第18項所述方法,其中每一電源閘單元依據其根據該電源閘單元之佈置與該電源閘之關聯的 邏輯單元的佈置估計電流之一全面路由模型被佈置。 The method of claim 18, wherein each of the power gate units is associated with the power gate according to an arrangement of the power gate unit The arrangement of the logic cells estimates that one of the overall routing models is arranged. 如申請專利範圍第18項所述方法,其中該等電源閘單元依據電流需求之下降順序被指定至邏輯單元。 The method of claim 18, wherein the power gate units are assigned to the logic unit in descending order of current demand. 如申請專利範圍第18項所述方法,其中該電源閘單元以受到電流限制而被佈置。 The method of claim 18, wherein the power gate unit is arranged to be limited by current. 如申請專利範圍第18項所述方法,其中該等邏輯單元使用一第一佈置步驟,及該等電源閘單元使用第二佈置步驟被佈置,其中該第一佈置步驟以不關切該等電源閘單元之佈置地被執行。 The method of claim 18, wherein the logic unit uses a first arrangement step, and the power gate units are arranged using a second arrangement step, wherein the first arrangement step is not concerned with the power gates The arrangement of the units is performed. 如申請專利範圍第18項所述方法,其中該等邏輯單元與該等電源閘單元使用多佈置步驟被佈置,其中該等佈置步驟之一以該等電源閘單元被排除地被實施。 The method of claim 18, wherein the logic units are arranged with the power supply gate units using a plurality of arrangement steps, wherein one of the arrangement steps is performed with the power supply gate units excluded. 如申請專利範圍第49項所述方法,進一步包含為該等電源閘單元指定高的成本函數以在此佈置步驟中排除該等電源閘單元。 The method of claim 49, further comprising assigning a high cost function to the power gate units to exclude the power gate units in the step of arranging. 如申請專利範圍第49項所述方法,其中該等電源閘單元由在此佈置步驟中被使用之設計館被排除。The method of claim 49, wherein the power gate units are excluded from the design hall used in the arrangement step.
TW94129877A 2004-11-26 2005-08-31 Design method and architecture for power gate switch placement TWI393189B (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US10/998,204 US7590962B2 (en) 2003-12-17 2004-11-26 Design method and architecture for power gate switch placement

Publications (2)

Publication Number Publication Date
TW200620486A TW200620486A (en) 2006-06-16
TWI393189B true TWI393189B (en) 2013-04-11

Family

ID=48803329

Family Applications (1)

Application Number Title Priority Date Filing Date
TW94129877A TWI393189B (en) 2004-11-26 2005-08-31 Design method and architecture for power gate switch placement

Country Status (1)

Country Link
TW (1) TWI393189B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10204920B2 (en) 2015-04-09 2019-02-12 Samsung Electronics Co., Ltd. Semiconductor device including polygon-shaped standard cell
KR102321605B1 (en) * 2015-04-09 2021-11-08 삼성전자주식회사 Method for designing layout of semiconductor device and method for manufacturing semiconductor device using the same
KR102541506B1 (en) * 2018-01-17 2023-06-08 삼성전자주식회사 Semiconductor device including switch cells

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030218478A1 (en) * 2002-05-24 2003-11-27 Sani Mehdi Hamidi Regulation of crowbar current in circuits employing footswitches/headswitches
US20040210728A1 (en) * 2003-04-10 2004-10-21 Krisztian Flautner Data processor memory circuit

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030218478A1 (en) * 2002-05-24 2003-11-27 Sani Mehdi Hamidi Regulation of crowbar current in circuits employing footswitches/headswitches
US20040210728A1 (en) * 2003-04-10 2004-10-21 Krisztian Flautner Data processor memory circuit

Also Published As

Publication number Publication date
TW200620486A (en) 2006-06-16

Similar Documents

Publication Publication Date Title
US7987441B2 (en) Method and architecture for power gate switch placement
US7509613B2 (en) Design method and architecture for power gate switch placement and interconnection using tapless libraries
US5384472A (en) Symmetrical multi-layer metal logic array with continuous substrate taps and extension portions for increased gate density
US7943436B2 (en) Integrated circuit devices and methods and apparatuses for designing integrated circuit devices
US7129562B1 (en) Dual-height cell with variable width power rail architecture
KR101163891B1 (en) Supplying Power to Integrated Circuits Using a Grid Matrix Formed of Through-Silicon Vias
KR101611888B1 (en) Semiconductor integrated circuit
US9018046B2 (en) Area-efficient distributed device structure for integrated voltage regulators
WO2004077556A1 (en) Semiconductor integrated circuit device and its power supply wiring method
US20040005738A1 (en) Sea-of-cells array of transistors
US20020153574A1 (en) Sea-of-cells array of transistors
TWI393189B (en) Design method and architecture for power gate switch placement
JPS63120439A (en) Circuit and system of synchronous array logic
JPH04116951A (en) Semiconductor integrated circuit
US5635737A (en) Symmetrical multi-layer metal logic array with extension portions for increased gate density and a testability area
US7002253B2 (en) Semiconductor device and design method thereof
JPH0349214B2 (en)
US5404034A (en) Symmetrical multi-layer metal logic array with continuous substrate taps
Lin et al. A low-power field-programmable gate array routing fabric
Chen et al. Crosstalk minimization in three-layer HVH channel routing
JPH07273208A (en) Polycell integrated circuit
Zhang et al. Stochastic wire-length and delay distributions of 3-dimensional circuits
EP1743374A1 (en) Integrated circuit layout for virtual power supply
KR19980032570A (en) Cells Used To Form A Composite Integrated Circuit And Method Of Forming Them
JP2005251862A (en) Semiconductor integrated circuit and method of designing its layout