經濟部中央標準局員工消費合作社印製 A7 B7五、發明説明(i ) 本發明偽有開一種動態記億體,特別是有關於分辨快 頁模式動態記億體(Fast Page node DRAM,簡稱FPD)以及 延伸输出動態記億體(Extended Data-Out DRAM ,簡稱EDO )兩者,當使用者同時應用快頁模式動態記億醱以及延伸 输出動態記億體於同一電腦条統中,条統能夠輕易檢驗出 來,以便設定不同的操作組態,提昇条统整體之效能。 動態記億體是目前最普遍採用的電腦糸统主記億體, 整値結構是矩陣形式的排/列記億單元所組成。因此,操 作動態記億體時必須利用列位址及行位址指定某一待處理 記億單元,再進行讀出或寫入資料的動作。亦即,一般動 態記億體進行存取資料步驟時.首先將列位址閃控信號ΰ! 降為低電位用以觸發列位址的讀取,再將行位址閃控信號 降為低電位用以觸發行位址的纊取,如此指定出所需 處理的記億單元,再利用輸出/輸入資料線DQ存取資料即 可。然而,以注動態記憶體製造廠商為了提高操作速度, 另外提供一種頁模式的存取方式,操作方式則限於由列位 址所界定的頁邊界為限,稱之為快頁模式。快頁模式動態 記億體的操作方式與習知不同之處,在於當列位址閃控倍 號^降為低電位以觸發列位址的讀取後,持缠性保持低 霣位的狀態,而只改變行位址閃控信號為低電位狀態 以便讀取相同列位址的資料。在列位址閃控信號回愎 高電位時,快頁模式卽告結束。 雖然快頁模式動態記億體已為目前市面上檫準的動態 記億體,但是在速度方面的表現一直無法令人滿意,特別 ----i--------------訂--\-----線 (請先閲讀背面之注意事^P填寫本頁) 本紙張尺度適用中國國家標準(CNS ) A4規格(2丨OX297公釐) 經濟部中央標準局員工消費合作社印製 A7 B7五、發明説明(2 ) 在新一代艟人電腦中,如Intel Pentium ,更為顯箸。因 此部份的動態記億體廠商.例如Micron, HEC, Sansung. Toshiba等,卽制定一種新的頁模式動態記億體,稱之為 延伸输出動態記億體。延伸输出動態記億體與快頁模式動 態記億髖在接脚型態以及絕大部份時序待性均相同,差異 點在於輪出/輸人資料線DQ上資料所持缅的時間不同。以 往快頁模式動態記億體的輸出/輸入資料線DQ上資料會随 箸行位址閃控信號的上昇邊緣而關閉其緩衝器,但延 伸输出動態記億體則不會。第1圖與第2_分別為快頁模 式動態記億體以及延伸输出動態記億體讀取資料之時序圖 。在第1圖,其中,Clock為中央處理機時脈信號,READ 為讅取信號。由於行位址閃控信號ϋΙΥ的預充電時間ten 必須在輸入/输出資料線DQ為無效資料時進行,因此每個 讀取動作間隔3値中央處理機時脈週期;在第2圖,由行 位址閃控信號ΰΥ的預充電時間tc 〃不霈在輸入/輸出資 料線DQ為無效資料時進行,因此每嫡讀取動作只需間隔2 傾中央處理機時脈週期即可。如此將導致動態記億體的操 作速度大幅提昇。 然而,使用者有可能在同一塊主機板上,同時蓮用快 頁模式動態記億體以及延伸輸出動態記億體。對於II腦条 統而言,只要兩者置於不同的記億體組(bank)中,目前糸 統已有能力根據動態記億體的時序待性,給予最佳的操作 模式。否則,即使電腦糸統内插有操作速度更快的延伸輸 出動態記億體,為了確保資料的正確性,仍必須將整體速 ^ ^ u n n ^ n J n 線 (請先閲讀背面之注意事填寫本頁) 本紙張尺度適用中國國家標準(CNS ) A4規格(210 X 297公釐) A7 B7 五、發明説明(3 ) 度以快頁模式動態記憧體為基準。因此,只要兩種動態記 億醱是設置於不同的記億體组内,並且条統能夠分辨,即 可充分發揮条统資源的效能,提昇整體速度。 直接分辨快頁模式及延伸輸出動態記億體的方法,是 將一筆測試資料寫入待測試的動態記億體中,随後利用延 伸输出動態記億體的較快時序讀取,假如待測動態記億體 為延伸輸出動態記億體,諛取資料與寫入資料會完全相同 ,若為快頁模式動態記憶體,則會因速度過快而謓取錯誤 。利用軟體,即可輕易達成上述方法。然而,習知方法在 實際應用上,極易發生判別錯誤的情況。原因在於動態記 億體的輸入/輸出資料線DQ的對外導線上,具有随機性的 雜散電容,當快頁模式動態記億體的DQ在行位址閃控信號 上昇邊緣後消失,但由於雜散電容的緣故,有可能輸 入/輸出資料線DQ上資料不會完全消失,仍然會持缠一段 相當長的時間(20 ns以上)。拙使以延伸输出動態記億體 的時序讀取,仍然有可能寫入與讀取資料相符,而造成誤 I —Ί I Ί— I I I I I I .^1 I I n 訂 I. —Ji I ~ 線 (請先閲讀背面之注意事寫本頁) 經濟部中央樣準局員工消费合作社印製 条 能 ·統 測或 腦 ,時糸 體體 電 法體起 億億 的 方億引 記記 後 試記, 態態 判 測態況 動動 誤。體動情 種式 此態億出的 一模 因狀記輸判 供頁 ,的態伸誤 提快 性定動延成 先為 特穩種及造 首體 機不 I 體. 明億 随度供億應 發記 有高提記效 本態 具種明態散 ,動 是一 發動雜 的 ~ 容於本式的 目別 電處 ,模板 的判 散是此頁機 述以 雜上於快主 C 上用 於際鑑別為定據 , 由實有判因穩根法 0 , 在免不 方 判統 夠避的 試 本紙張尺度適用中國國家標準(CNS ) A4说格(210X297公釐) 經濟部中夬標準局員工消費合作社印製 A7 ^___ B7 五、發明説明(4 ) 是延伸輸出動態記億體,該测試方法包括步驟:寫入至少 一測試資料於該動態記億體中;進行讀取動作中,當該快 頁模式的輸入輸出資料線送出該測試資料後,將該輸入輸 出資料線進行放電;讀取該輸入輸出資,料線之線上資料; 藉此,當該線上資料與該測試資料一致時,該動態記億體 為延伸輸出動態記憶體,當該線上資料與該測試資料不一 致時,該動態記億體為快頁模式動態記億體。 另外,本發明又提供一種動態記億體測試方法,用以 判別一動態記億體為快頁模式動態記億體或是延伸輸出動 態記億體,該测試方法包括步驟:寫入至少一測試資料於 該動態記億體中;進行讀取動作中,當該動態記億體的输 入输出資料線送出該測試資料後,延遲一適當時間;讀取 該輸入輸出資料線之線上資料;藉此,當該線上資料舆該 测試資料一致時,該動態記億體為延伸輸出動態記億體, 當該線上資料與該測試資料不一致時,該動態記億體為快 頁模式動態記憶體。 為使本發明之上述目的、特徴、和優點能更明顯易懂 ,下文特舉一較佳實施例,並配合所附圖式,作詳細說明 如下: 圔式的簡單說明: 第1圖為快頁模式動態記億體謓取動作之時序圖。 第2圖為延伸輸出動態記億體謓取動作之時序圔。 第3圖為本發明第一實施例之時序圖。 第4圖為本發明第二實施例之時序圖。 -6 - 本紙張尺度適用中國國家標隼(CNS ) A4規格(210X297公釐) . I I 裝 H I I I 訂— —I I 線 (請先閱讀背面之注意事填寫本頁) B7 1、發明説明(5 ) 第一實施例: 本實施例判別動態記億體是快頁模式動態記億醱或者 延伸输出動態記億體的方法,首先,寫入一組测試資料到 待判別的動態記億體中。接著,進行讀取資料判別的動作 。謫參考第3圖,其表示第一實施例中讀取測試資料判別 的時序圖。當行位址閃控信號在高霣位降至低電位的 下降邊緣時,將已存入測試資料之動態記億體的位址 讀入;當行位址閃控信號在低電位昇至高電位的上昇 邊緣時,即由位址ADR所指定的動態記億體中輸入輸出資 料線DQ送出正確的測試資料。快頁模式動態記億體的DQ在 經過一段時間後,即轉變為無效資料,延伸輸出動態記億 體則能維持有效資料直到下一値讀取遇期。第一實施例所 採用的方法,稱之為***等待法,亦即當行位址閃控信號 經濟部中央標準局員工消費合作社印裝 (請先閲讀背面之注意事填寫本頁) 上昇時,讀取信號^必須等待的時間,使 得快頁模式動態記億體受到主機板上雜散電容的影堪而延 伸的資料,轉變成無效資料為止。例如一般快頁模式動態 記億體及延伸輪出動態記億體的讀取信號***為X-3-3-3 及X-2-2-2 ,表示每2及3艇週期謓取一次資料,在本實 施例中測試階段,讀取信號***可設定為X-7-7-7或更大 。因此,本實施例中的***等待方式可以藉由軟醴達成。 另外,延遲時間〃在不同主機板上必須有不同設定, 以因應不同主機板所造成雜散電容之效應。 第二實施例: 本實施例所採用的方法,稱之為強迫放電法。同樣地 本紙張尺度適用中國國家揉準(CNS ) A4規格(210X297公釐) A7 B7 經濟部中央橾準局—工消費合作社印製 i、發明説明(6 ) ,對待測動態記億體寫入一組測試資料,再謓取出來進行 比對。當行位址閃控倍號在低電位舁至高電位的上昇 邊緣時,動態記億體的输入输出資料線DQ為正確的測試資 料。此時,条统在输入输出資料線DQ的另一端改變狀態為 0,假若動態記億體是快頁模式動態記億體時,由於输出 缓衝器已關閉,因此雜散電容上的電位會下降,成為無效 資料。另一方面,假若動態記億體為延伸輸出動態記億體 時,由於輸出缓衝器仍未關閉,因此能夠維持有效資料。 最後,讀取線上資料之後,與原測試資料比對即可判別動 態記億體的類型。同樣地,本實施例亦可透過軟體的方式 完成。 本發明之待點如下所述: 1. 本發明所採用的***等待法及強迫放電法,能夠相 當有效的防止雜散電容影堪判別結果。 2. 實施方式可以藉由軟體或BIOS程式執行,適用於一 般電腦条統中改動測試,,自動建立動態記億體的正確組態 〇 雖然本發明已以較佳實施例掲露如上,然其並非用以 限定本發明,任何熟習此項技藝者,在不脱離本發明之精 神和範圍内.當可作些許之更動與潤飾,因此本發明之保 護範圍當視後附之申誚專利範圍所界定者為準。 ----:--.-----裝 I .1 (請先閲讀背面之注意事填寫本頁) 訂 線 本紙張尺度適用中國國家標準(CNS ) A4规格(210X2.97公釐)Printed by the Consumers' Cooperative of the Central Standards Bureau of the Ministry of Economic Affairs A7 B7 V. Description of the Invention (i) The present invention pretends to open a dynamic memory memory, especially about the fast page mode dynamic memory memory (Fast Page node DRAM, FPD for short). ) And Extended Data-Out DRAM (EDO), when the user simultaneously applies the fast page mode to dynamically record 100 million bytes and the extended output dynamic memory 100 million in the same computer system, the system can Easily check it out in order to set different operating configurations and improve the overall performance of the system. The dynamic memory system is currently the most commonly used computer system in the system, and the entire structure is composed of rows / columns of 100 million units in the form of a matrix. Therefore, it is necessary to use a column address and a row address to designate a certain to-be-recorded 100 million cell when operating dynamic memory, and then read or write data. That is, when the general dynamic memory is used to access the data, the column address flash control signal ΰ! Is first lowered to trigger the reading of the column address, and then the row address flash control signal is reduced to low. The potential is used to trigger the capture of the row address, so that the hundred million units to be processed are specified, and then the data can be accessed using the output / input data line DQ. However, in order to improve the operation speed, the dynamic memory manufacturer also provides a page mode access mode. The operation mode is limited to the page boundary defined by the column address, which is called the fast page mode. The operation mode of dynamic page memory in the fast page mode is different from the conventional one. After the column address flash control multiplying ^ drops to a low potential to trigger the reading of the column address, the entanglement remains low. , And only change the flash address signal of the row address to a low state so as to read the data of the same column address. When the column address flash control signal returns to high level, the fast page mode ends. Although the dynamic page-recording mode of the fast page mode has been the current standard dynamic-recording mode, the performance in terms of speed has been unsatisfactory, especially ---------------- --- Order-\ ----- line (please read the notes on the back ^ P to fill out this page) This paper size applies to China National Standard (CNS) A4 (2 丨 OX297 mm) Central Standard of the Ministry of Economic Affairs A7 B7 printed by the Bureau ’s Consumer Cooperatives V. Invention Description (2) The new generation of human computers, such as Intel Pentium, is even more prominent. Therefore, some manufacturers of dynamic memory memory, such as Micron, HEC, Sansung, Toshiba, etc., have formulated a new page model of dynamic memory memory, which is called extended output dynamic memory memory. The extended output dynamic memory model and the fast page model dynamic memory model have the same pin configuration and most of the timing characteristics. The difference lies in the time held by the data on the rotation / input data line DQ. In the fast page mode, the data on the output / input data line DQ of dynamic memory will close its buffer with the rising edge of the address flash control signal, but the extended output dynamic memory will not. Figures 1 and 2_ are the timing diagrams for reading data in the fast page mode for dynamic memory counting and extended output dynamic memory counting, respectively. In Figure 1, Clock is the central processing unit clock signal and READ is the capture signal. Since the pre-charging time ten of the row address flash control signal ϋΙ 必须 must be performed when the input / output data line DQ is invalid, each reading action interval is 3 値 the central processor clock cycle; in Figure 2, the line The pre-charging time tc of the address flash control signal 〃 is not performed when the input / output data line DQ is invalid data, so each read operation only needs to be spaced 2 clock cycles from the central processing unit. This will lead to a significant increase in the operating speed of dynamic memory. However, the user may be on the same motherboard, while using the fast page mode to dynamically record billions of characters and extended output to dynamically record billions of characters. As far as the II brain system is concerned, as long as the two are placed in different banks, the current system has the ability to give the best mode of operation based on the timing of dynamic banks. Otherwise, even if the computer system is inserted with a faster extended output dynamic memory, in order to ensure the accuracy of the data, the overall speed must be ^ ^ unn ^ n J n line (please read the notes on the back first and fill in (This page) This paper size is in accordance with Chinese National Standard (CNS) A4 (210 X 297 mm) A7 B7 5. Description of the invention (3) The speed is based on the dynamic recording of the fast page mode. Therefore, as long as the two types of dynamic memory are set up in different memory groups, and the system can be distinguished, the efficiency of system resources can be fully exerted and the overall speed can be improved. The method of directly distinguishing the fast page mode and the extended output dynamic memory is to write a piece of test data into the dynamic memory to be tested, and then use the extended output dynamic memory to record the faster time series. The dynamic memory is the extended output dynamic memory, and the fetching data and writing data will be exactly the same. If it is fast page mode dynamic memory, it will fetch errors due to too fast speed. The above methods can be easily achieved using software. However, in practice, the conventional method is prone to discriminating errors. The reason is that the stray capacitance on the external wire of the input / output data line DQ of the dynamic memory register is lost. When the DQ of the dynamic memory register of the fast processor in the fast page mode disappears after the rising edge of the row address flash control signal, but Due to stray capacitance, it is possible that the data on the input / output data line DQ will not completely disappear, and it will still linger for a considerable time (more than 20 ns). Even if the timing of the dynamic output of the billion-byte reading is extended, it is still possible to write in accordance with the read data and cause errors. I —Ί I Ί — IIIIII. ^ 1 II n Order I. —Ji I ~ Line (Please First read the notes on the back page and write this page) The Central Consumers ’Bureau of the Ministry of Economic Affairs, the Consumer Cooperatives, printed bar energy, system test, or brain, and then recorded the test results with billions of billions of quotations. Judging the state of movement. This kind of physical and emotional situation is a meme-like form of memorizing the submission of judgment pages. The state of error and speed will be extended to the first stable species and the first machine. It should be noted that there is a high-level effect in this state. There is a kind of clear state, and the action is a complex one. Contained in the eye of this formula, the discriminant of the template is used on this page to add to the fast main C. The international identification is the basis, and the actual root cause method is 0. The paper size of the test paper that can be avoided by all means is subject to the Chinese National Standard (CNS) A4 standard (210X297 mm). A7 ^ ___ B7 printed by the Bureau ’s Consumer Cooperatives V. The invention description (4) is an extended output of the dynamic memory of billions, the test method includes the steps of: writing at least one test data in the dynamic memory of billions; In the fast page mode, the input and output data line sends the test data, and then the input and output data line is discharged; the input and output data and online data of the material line are read; thereby, when the online data and the test When the data is consistent, the dynamic record of the billion body is the extended output dynamic Recalling the body, when the line data and the test data are inconsistent, the body is a dynamic memories Fast Page Mode Dynamic memories body. In addition, the present invention also provides a dynamic memory counting method for determining whether a dynamic memory recording is a fast page mode dynamic memory recording or extended output dynamic memory recording. The test method includes the steps of writing at least one The test data is stored in the dynamic recorder; in the reading operation, when the input and output data lines of the dynamic recorder are sent out, the test data is delayed for an appropriate time; the online data of the input and output data lines are read; Therefore, when the online data is consistent with the test data, the dynamic memory is the extended output dynamic memory. When the online data is inconsistent with the test data, the dynamic memory is the fast page mode dynamic memory. . In order to make the above-mentioned objects, features, and advantages of the present invention more comprehensible, a preferred embodiment is given below in conjunction with the accompanying drawings to make a detailed description as follows: A brief description of the formula: FIG. 1 is fast Timing chart of page mode dynamic recording of billion-body grabbing action. Fig. 2 is the timing sequence of the extended output dynamic counting operation. FIG. 3 is a timing chart of the first embodiment of the present invention. FIG. 4 is a timing chart of the second embodiment of the present invention. -6-The size of this paper is applicable to China National Standard (CNS) A4 (210X297mm). II Pack HIII Order-—II Line (Please read the notes on the back first and fill in this page) B7 1. Description of Invention (5) First embodiment: In this embodiment, it is determined that the dynamic memory counting is a method of dynamically recording the billion memory in the fast page mode or extending the output of the dynamic memory recording. First, a set of test data is written into the dynamic memory recording to be determined. Next, read the data to determine the operation.谪 Refer to FIG. 3, which shows a timing chart for discriminating the reading test data in the first embodiment. When the row address flash control signal falls to the falling edge of the low potential at the high threshold, read the address of the dynamic memory that has been stored in the test data; when the row address flash control signal rises to the high potential at the low potential When the rising edge of the input, the input and output data lines DQ in the dynamic digitizer specified by the address ADR send the correct test data. After a period of time, the DQ of dynamic memory of billions of pages in the fast page mode is changed to invalid data. The extended output of dynamic memory of billions of hours can maintain valid data until the next reading period. The method used in the first embodiment is called the plug-in waiting method, that is, when the consumer address of the Central Standards Bureau of the Ministry of Economic Affairs of the Ministry of Economic Affairs of the Ministry of Economic Affairs of the Bank of China is printed (please read the notice on the back and fill in this page). The time that the signal must wait for reading, so that the fast page mode dynamically records the data that is affected by the stray capacitance of the motherboard and becomes invalid data. For example, in the general fast page mode, the dynamic signals of the 100 million body and the extended round-out of the 100 million body are inserted as X-3-3-3 and X-2-2-2, which means that data is captured every 2 and 3 boat cycles. In the test phase in this embodiment, the read signal insertion can be set to X-7-7-7 or greater. Therefore, the insertion waiting mode in this embodiment can be achieved by soft palate. In addition, the delay time 必须 must be set differently on different motherboards to account for the effects of stray capacitance caused by different motherboards. Second embodiment: The method used in this embodiment is called the forced discharge method. In the same way, this paper size applies to the Chinese National Standard (CNS) A4 (210X297 mm) A7 B7 Printed by the Central Bureau of Standards of the Ministry of Economic Affairs-Industrial and Consumer Cooperatives i. Invention Description (6) A set of test data, and then taken out for comparison. When the row address flash control multiplier is at the rising edge of low potential to high potential, the input and output data line DQ of the dynamic memory of the billion body is the correct test data. At this time, the system changes the state to 0 on the other end of the input and output data line DQ. If the dynamic memory is in fast page mode, the output buffer is closed, so the potential on the stray capacitance will be Decline and become invalid. On the other hand, if the dynamic digitizer is the extended output dynamic digitizer, the effective data can be maintained because the output buffer has not been closed. Finally, after reading the online data, compare with the original test data to determine the type of dynamic memory. Similarly, this embodiment can also be implemented by software. The main points of the present invention are as follows: 1. The insertion waiting method and the forced discharge method used in the present invention can effectively prevent the stray capacitance from being judged. 2. The implementation mode can be executed by software or BIOS program, suitable for modification test in general computer system, and automatically establish the correct configuration of dynamic memory. Although the present invention has been disclosed as above in the preferred embodiment, but its It is not intended to limit the present invention. Any person skilled in the art will not deviate from the spirit and scope of the present invention. Some modifications and retouching can be made. Therefore, the scope of protection of the present invention shall be regarded as the scope of the attached patent. The ones defined shall prevail. ----: --.----- Packing I .1 (Please read the notes on the back to fill in this page first) Threading This paper size applies the Chinese National Standard (CNS) A4 specification (210X2.97 mm)