1361358 九、發明說明： 【發明所屬之技術領域】 本發明係一種對稱式多處理器系統，特別是一種具有星形互連裝置與 冷卻系統之對稱式多處理器電腦。 【先前技術】 目前對於建造一個大型對稱式多處理器系統的最大挑戰在於複雜且繁 多的互連。這些互連必須能在最短距離内，將每一個中央處理器與其他所 有中央處理器做電路連結，形成一個在三度空間中呈巢狀之互連；而且每 一翻互連的長度均必須小於依據電性要求所定出之最大容許線路長。 一種可將所有中央處理器連結起來之方法為製造一個背板（通常申央 處理器都是被安裝在數張中央處理器板或中央處理器卡上），然後將所有中 央處理器連結在此背板的同一面上。基於最大容許線路長的要求，中央處 理器(或中央處理器板或中央處理器卡)彼此的位置必須非常靠近。換句話 說，最大容許線路長決定了系統所能包含的中央處理器總數，此乃因每加 入一個額外的尹央處理器(或中央處理器板或中央處理器卡)勢必會造成處 理器之間的線路長度增加。 另-個為符合上述要求之先前技術為使用直角連結器來將兩個中央處 理器板以直角的方式連結起來。目前市場上所提供的此種連結器包括 Molexl-Trac以及Amphenol-TCS(Teradyne)的直角連結器。這類連結器不 僅比傳統的背板連結器昂貴’同時也連帶產生_個散熱方面的難題。當這 些板以直角方式配置時會形成十字形的氣體流道，導致冷空氣難以進入， 熱空氣難以導出的現象發生。 5 ^61358 【發明内容】 基此’本發明提出一種具有星形互連裳置及冷卻系統之對稱式多處理 器電腾’星形互連裝置包含：巾間板；第_處理器板與第二處理器板，垂 直叹置於板之;^同表面，且彼此呈交錯狀。冷卻祕包含1 一冷卻 模組與第二冷卻模組，分湘以產生第—處理器板與第二處理器板之第一 氣流與第二氣流，其中第—氣流與第二氣流之流動路徑彼此呈交錯狀且分 別位於_間板之不同表面。 此一新式互連設計使得成本較低之傳統背板連結器，得以被應用在性 能優異之交錯錢直角拉置上，_又兼具成本昂貴之直角連結器所具 有之較短線路長的優點。 本發明之特徵、概念與優勢在參照以下之說明、圖式及申請專利範圍 後’將更易於被理解。必須暸解的是，上述的概略說明以及以下搭配實施 例所做之詳細說明之用意乃為本發明之請求項提供更進一步的解釋。 【實施方式】 以下將針對本發明之較佳實施例，以圖式和文字做更詳盡的說明各 圖式或說明中所出現之相同標號代表相同或類似的物件。 請參照第1A圖與第m圖，為本發明第一實施例之對稱式多處理器電1361358 IX. Description of the Invention: [Technical Field] The present invention is a symmetrical multiprocessor system, and more particularly a symmetrical multiprocessor computer having a star interconnect and a cooling system. [Prior Art] The biggest challenge currently for building a large symmetric multiprocessor system is the complex and numerous interconnections. These interconnects must be able to circuit each central processor to all other central processors in the shortest distance to form a nested interconnect in three dimensions; and the length of each interconnect must be Less than the maximum allowable line length determined according to electrical requirements. One way to connect all the central processors is to create a backplane (usually the central processor is installed on several central processor boards or central processor cards) and then connect all the central processors here. The same side of the back panel. The central processor (or central processor board or central processing unit card) must be located very close to each other based on the maximum allowable line length requirements. In other words, the maximum allowable line length determines the total number of central processors that the system can contain. This is because each additional Yinyang processor (or central processor board or central processing unit card) is bound to cause the processor. The length of the line between the two increases. Another prior art that meets the above requirements is to use a right angle connector to join the two central processor boards at right angles. Such connectors currently available on the market include Molexl-Trac and Amphenol-TCS (Teradyne) right angle connectors. These types of connectors are not only more expensive than traditional backplane connectors, but they also create a heat dissipation problem. When these plates are arranged at right angles, a cross-shaped gas flow path is formed, which makes it difficult for cold air to enter and the hot air is difficult to be exported. 5 ^61358 [Summary of the Invention] Accordingly, the present invention provides a symmetric multiprocessor teleton 'star interconnect device having a star interconnected skirting and cooling system comprising: a towel board; a _ processor board and The second processor board is vertically slanted on the board; the same surface and staggered with each other. The cooling module comprises a cooling module and a second cooling module, and the first airflow and the second airflow of the first processor board and the second processor board are generated, wherein the flow paths of the first airflow and the second airflow They are staggered with each other and are located on different surfaces of the _ plate. This new interconnect design allows the lower cost of traditional backplane connectors to be used in high-performance, staggered, right-angled pulls, which combines the advantages of the shorter line lengths of costly right-angle connectors. . The features, concepts, and advantages of the present invention will be more readily understood by reference to the description, appended claims It is to be understood that the foregoing general description of the claims DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following, the preferred embodiments of the present invention will be described in more detail in the drawings and the drawings. Please refer to FIG. 1A and FIG. m to illustrate a symmetric multiprocessor power according to a first embodiment of the present invention.
腦100之示意圖。第1A圖為對稱式多處理器電腦100之前透視圖，第1B 圖為對稱式多處理器電腦100之前視圖。第2A圖為本發明第—實施例之對 稱式多處理器電腦1〇〇之後透視圖，第2B圖為第1A圖之對稱式多處理器 電腦100之後視圖。 6 1361358 對稱式多處理器電腦100主要包含星形互連裝置（未標號）與冷卻系 統（未標號）’設置於外殼140内。 星形互連裝置主要包含：中間板110,第一處理器板120與第二處理器 板130 ’依操作上需求’設置於外殼14〇内。外殼14〇提供支撐星形互連裝 置所必要之框架（未繪製於圖式中）。 s月參照第1A至第2B圖’其中中間板11〇，第一處理器板120與第二 處理器板130均為矩形印刷電路板，第一處理器板12〇彼此平行排列且垂 直設置於中間板110之前表面，且每一第一處理器板120設置有至少一第 一處理器（未繪製於圖式中）’同樣地，第二處理器板130彼此平行排列且 垂直設置於中間板110之後表面，且每一第二處理器板13〇設置有至少一 第二處理器（未繪製於圖式中）。設置於第一處理器板12〇或第二處理器板 130上之處理器之最佳數量可以是2、4、6、8或16…等。對稱式多處理器 電腦100之其中一個主要特徵為第一處理器板120與第二處理器板13〇彼 此呈交錯狀。 請參照第1B圖，每一第一處理器板120與位於中間板11〇另一面之每 一第二處理器板130構成一個X型的圖案。 請參照第1B圖及第2B圖，第一處理器板120與第二處理器板13〇分 別對中間板110之兩對角線方向平行設置，即第1B圖中之四個第一處理器 板120乃平行於中間板之右上角至左下角之對角線，相反地，四個第二處 理器板130乃平行於中間板左上角至右下角之對角線。另一種結構為第一 處理器板120垂直於第二處理器板130。 7 1361358 前連結器111 (第一實施例中有16個）設置於中間板110之前表面形 成一個前互連矩陣（如第1B圖與第2B圖所示，為4x4矩陣）以電性連結 第一處理器板120與中間板11〇。第一實施例中，兩個未繪製於圖式中之處 理器，南橋（SouthBridge)及基板管理控制器（baseboardmanagement controller) ’亦可設置於每一第一處理器板120上。四個排成一列之第一連 結器121設置於每一第一處理器板12〇之一緣，用以與同樣四個排成一列 之前連結器111連結，使訊號得以透過中間板110傳遞，四個第一處理器板 120總共需要16個第一連結器121來和中間板11〇上之16個相對應之前連 結器111互連。 類似地，後連結器112 (第一實施例中有16個）設置於中間板110之 後表面形成一個後互連矩陣（如第1B圖與第2B圖所示，為4x4矩陣）。兩 個未繪製於圖式中之處理器’南橋（SouthBridge)及基板管理控制器 (baseboardmanagement controller)’ 亦可設置於每一第二處理器板 130 上。 四個排成一列之第二連結器122設置於每一第二處理器板no之一緣，用 以與同樣四個排成一列之後連結器112連結，使訊號得以透過中間板11() 傳遞’四個第二處理器板130總共需要16個第二連結器122來和中間板η〇 上之16個相對應之後連結器112互連。 基本上，前互連矩陣和後互連矩陣在中間板11〇的兩邊形成一個互連 區’其中前連結器1Π和後連結器112被設置為一個星形直角圖案，成為星 形互連裝置巾之核心結構。為了驗線長及盡可能增加互連的數目，前連 8 1361358 結矩陣和後連結矩陣應該儘可能重疊。然而在本實施例中，任一前連結器 111與任一後連結器112均位於中間板110之不同表面且未重疊。 請參照第1A圖、第1B 0、第2A ®與第2B圖，為了解決不使用水冷 所造成的散熱問題，對稱式多處理器電腦100更包含一個冷卻系統（未標 號）’以提供星形互連裝置足夠的氣流。具有上述強化互連結構之星形互連 裝置亦提供氣流-個最佳推抽結構。冷卻系統包含：具有第一推進風扇151 與第一抽出風扇152之冷卻模組（未標號)，用以產生第一氣流153 ;及具 有第二推進風扇161與第二抽出風扇162之冷卻模組（未標號），用以產生 第一氣流163。 请參照第1A圖和第1B圖，位於對稱式多處理器電腦1〇〇左下角之第 -推進風扇151自對稱式多處理器電腦丨⑻之前方吸取冷風後，向上推送， 在第-處理器板12G之左下角位置產生第—氣流153，然後冷第—氣流153 流經平行的第-處理器板12〇且流經設置於其上之第一處理^。吸取熱量 後之熱第-氣流153亦在第一抽出風扇152的拉引下，自第一處理器板12〇 的右上方流出，接下來位於對稱式多處理器電腦1〇〇右上角之第一抽出風 扇152將熱第一氣流153抽出，並朝對稱式多處理器電腦100之後方排出。 為了保"^第―氣流如上述般流動，可以在第一推進風扇151和第一處理器 板120之左下方之間，或者是第一抽出風扇152和第一處理器板之右 上方之間，加裝一個或一個以上的風罩來導引所產生之氣流。 請參照第2A圖和第2B圖，位於對稱式多處理器電腦1〇〇左上角之第 一推進風扇161自對稱式多處理器電腦1〇〇之前方吸取冷風後，向下推送， 9 1361358 在第二處理器板13G之左上角位置產生第二氣流163，然後冷第二氣流163 流經平行的第4理器板⑽且祕設置於其上之第三處理器。吸取熱量 後之熱第二氣流163亦在第二抽出風扇162的拉引下自第二處理器板13〇 的右下方流出’接下來位於對稱式多處理器電腦⑽右下角之第二抽出風 扇162將熱第二氣流163抽出，並朝對稱式多處理器電腦之後方排出。 為了保證第二氣流如上述般流動，可以在第二推進風扇i6i和第二處理器 板130之左上方之間’或者是在帛二抽丨風扇162和第二處理器板13〇之 右上方之間’加裝-個或—個以上的風罩來導引所產生之氣流。 因為第-氣流153和第二氣流163之流動路徑分別與第一處理器板12〇 和第二纽器板130之方向一致，因此第一氣流153和第三氣流163之流 動路徑在巾間板11G的不同表面卿成交錯狀，此乃應用在具有星形互連 裝置之對稱式多處理器電腦1〇〇之冷卻系統之主要特徵。諸如風扇之樣式、 尺寸轉速’抑或疋氣流流入與流出風扇之方向不應偈限在本實施例於上 述或如後所揭露之範圍。 月參…、第3Α、3Β、4Α與4Β®，為本發明之第二實施例。對稱式多處 理器電滕2GG之星形互連裝置主要包含:_間板2m;第—處理器板娜與 第二處理器板230,第-處理器板22〇與第二處理器板23〇設置於外殼24〇 内之方式與第一實施例相同。對稱式多處理之冷卻系統設置方 式亦與第-實__。第二實補和第—實_之主要差異為第二實施 例具有兩個T型板213及214，分觀於t間板⑽之上側與下側，或者是 其他可行之位置4型板213及214皆為τ型印刷電路板，兩者皆可被做 10 1361358 為一獨立的零件，然後利用邊對邊的連結方式與中間板210互連，抑或是 成為中間板210之延伸部位。τ型板213及214乃用以提供系統和儲存器、 電源、擴充卡、友網路介面彼此間的互連。此外，由於T型板213及214 在風扇的位置處形成四個開口，因此並不會阻擋氣流的流動。對稱式多處 理器電腦200之T型板的總數與實際實施方式有關，在有些情況下只要一 個便已足夠。此外，T型板213及214亦可以僅做為支撐用，而不含任何電 路。 請參照第5圖與第6圖，為本發明之第三實施例。對稱式多處理器電 腦300之星形互連裝置主要包含：中間板31〇,第一處理器板32〇，第二處 理器板330與兩個T型板313及314,設置於外殼340的方式與第二實施例 相同。對稱式多處理器電腦300之冷卻系統之設置方式亦與第二實施例相 同。在第5圖和第6圖中，兩個硬碟模組351及352，兩個電源供應模組 353及354以及兩個擴充卡模組355及356，設置於星形互連裝置上。硬碟 模組351及352 ’電源供應模組353及354以及擴充卡模，组355及356均被 足夠的支撐配件所支撐。此外，每一位於硬碟模組351及议内之硬碟彼 此間均具有足夠的通道’以避免擒住氣流。同理，每—位於擴充卡模組355 及356之擴充卡亦然。 雖然本發明的技術β容已經以較佳實施例揭露如上然其並非用以限 定本發明，任何熟習此技藝者，在;^脫離本發日狀精神所作些許之更動與 濁飾’皆應涵蓋於本發_料内，因此本發a月之保護範圍當視後附之申 請專利範圍所界定者為準》 11 丄妁1358 【圖式簡單說明】 第1A圓為本發明第一實施例之具有星形互連裝置及冷卻系統之對稱 式多處理器電腦之前透視圖。 第1B圖為第1A圖之前視圖。 第2A圖為本發明第一實施例之對稱式多處理器電腦之後透視圖。 第2B圖為本發明第一實施例之對稱式多處理器電腦之後視圖。 第3A圖為本發明第二實施例之具有星形互連裝置及冷卻系統之對稱 式多處理器電腦之前透視圖。 第3B圖為第3A圖之前視圖。 第4A圖為本發明第二實施例之對稱式多處理器電腦之後透視圖。 第4B圖為本發明第二實施例之對稱式多處理器電腦之後視圖。 第5圖為本發明第三實施例之具有星形互連裝置及冷卻系統之對稱式 多處理器電腦之前透視圖。 •第6'圖為本發明第三實施例之後透視圖。 【主要元件符號說明】 100.........對稱式多處理器電腦 110 .........中間板 111 .........前連結器 112 .........後連結器 120 .........第一處理器板 121 .........第一連結器 122 .........第二連結器 12 1361358 130.........第二處理器板 140.........外殼 151 .........第一推進風扇 152 .........第一抽出風扇 153 .........第一氣流 161 .........第二推進風扇 162 ....... · ·第二抽出風扇 163 .........第二氣流 200 .........對稱式多處理器電腦(本發明第二實施例） 210.........中間板 213 .........T形板 214 .........T形板 220 .........第一處理器板 230 .........第二處理器板 240 .........外殼 300 .........對稱式多處理器電腦（本發明第三實施例） 310.........中間板 313 .........T形板 314 .........T形板 320 .........第一處理器板 330 .........第二處理器板 13 1361358 340 .........外殼 351 .........硬碟模組 352 .........硬碟模組 353 .........電源模組 354 .........電源模組 355 .........擴充卡模組 356 .........擴充卡模齒·Schematic diagram of the brain 100. 1A is a front perspective view of a symmetric multiprocessor computer 100, and FIG. 1B is a front view of a symmetric multiprocessor computer 100. Fig. 2A is a perspective view of the symmetrical multiprocessor computer of the first embodiment of the present invention, and Fig. 2B is a rear view of the symmetrical multiprocessor computer 100 of Fig. 1A. 6 1361358 The symmetrical multiprocessor computer 100 mainly includes star interconnects (not labeled) and a cooling system (not labeled) disposed within the outer casing 140. The star interconnect device mainly includes an intermediate board 110, and the first processor board 120 and the second processor board 130' are disposed in the housing 14" according to operational requirements. The housing 14A provides the necessary frame to support the star interconnect (not shown in the drawings). Referring to FIGS. 1A to 2B, in which the intermediate board 11 is, the first processor board 120 and the second processor board 130 are rectangular printed circuit boards, and the first processor boards 12 are arranged in parallel with each other and vertically disposed on The front surface of the intermediate plate 110, and each of the first processor boards 120 is provided with at least one first processor (not shown in the drawings). Similarly, the second processor boards 130 are arranged in parallel with each other and vertically disposed on the intermediate board. The rear surface of 110, and each second processor board 13 is provided with at least one second processor (not shown in the drawings). The optimal number of processors disposed on the first processor board 12 or the second processor board 130 may be 2, 4, 6, 8, or 16...etc. One of the main features of the symmetrical multiprocessor computer 100 is that the first processor board 120 and the second processor board 13 are staggered with each other. Referring to Figure 1B, each of the first processor boards 120 and each of the second processor boards 130 on the other side of the intermediate board 11 form an X-shaped pattern. Referring to FIGS. 1B and 2B, the first processor board 120 and the second processor board 13 are respectively disposed in parallel with respect to the two diagonal directions of the intermediate board 110, that is, the four first processors in FIG. 1B. The plate 120 is parallel to the diagonal line from the upper right corner to the lower left corner of the intermediate plate. Conversely, the four second processor plates 130 are parallel to the diagonal line from the upper left corner to the lower right corner of the intermediate plate. Another configuration is that the first processor board 120 is perpendicular to the second processor board 130. 7 1361358 The front connector 111 (16 in the first embodiment) is disposed on the front surface of the intermediate plate 110 to form a front interconnection matrix (as shown in FIGS. 1B and 2B, which is a 4×4 matrix) to electrically connect. A processor board 120 and an intermediate board 11 are. In the first embodiment, two processors not shown in the figure, a South Bridge and a baseboard management controller may be disposed on each of the first processor boards 120. Four first connectors 121 arranged in a row are disposed on one edge of each of the first processor boards 12 to be connected to the same four arrays of front connectors 111 so that signals can be transmitted through the intermediate board 110. The four first processor boards 120 require a total of 16 first connectors 121 to interconnect with the 16 corresponding front connectors 111 on the intermediate board 11''. Similarly, the rear connector 112 (16 in the first embodiment) is disposed on the rear surface of the intermediate plate 110 to form a rear interconnection matrix (as shown in Figs. 1B and 2B, which is a 4x4 matrix). Two processors "SouthBridge" and a baseboard management controller (not shown) may also be disposed on each of the second processor boards 130. Four second connectors 122 arranged in a row are disposed on one edge of each second processor board no for connecting with the same four rows of connectors 112 so that signals can be transmitted through the intermediate board 11 () The 'four second processor boards 130 require a total of 16 second connectors 122 to interconnect with the 16 corresponding post connectors 112 on the intermediate board n. Basically, the front interconnect matrix and the rear interconnect matrix form an interconnect region on both sides of the intermediate plate 11', wherein the front connector 1 and the rear connector 112 are arranged in a star-shaped right-angle pattern to become a star interconnect device. The core structure of the towel. In order to check the line length and increase the number of interconnections as much as possible, the front junction 8 1361358 junction matrix and the post-join matrix should overlap as much as possible. In the present embodiment, however, either the front connector 111 and any of the rear connectors 112 are located on different surfaces of the intermediate plate 110 and do not overlap. Referring to FIG. 1A, FIG. 1B 0, 2A ® and 2B, in order to solve the heat dissipation problem caused by not using water cooling, the symmetric multiprocessor computer 100 further includes a cooling system (not labeled) to provide a star. Interconnect the device with sufficient airflow. The star interconnect device having the above-described reinforced interconnect structure also provides an air flow - an optimum push structure. The cooling system includes: a cooling module (not labeled) having a first propulsion fan 151 and a first extraction fan 152 for generating a first airflow 153; and a cooling module having a second propulsion fan 161 and a second extraction fan 162 (not numbered) to generate the first gas stream 163. Referring to FIG. 1A and FIG. 1B, the first-pushing fan 151 located in the lower left corner of the symmetric multiprocessor computer picks up the cold air before the symmetrical multiprocessor computer 丨 (8), and then pushes up, in the first processing. The lower left corner of the plate 12G produces a first airflow 153, and then the cold first airflow 153 flows through the parallel first processor board 12 and flows through the first process disposed thereon. The heat first airflow 153 is also drawn from the upper right side of the first processor board 12〇 under the pull of the first extraction fan 152, and then located in the upper right corner of the symmetric multiprocessor computer. An extraction fan 152 draws the hot first airflow 153 and discharges it toward the symmetrical multiprocessor computer 100. In order to ensure that the "flow" flows as described above, between the first propulsion fan 151 and the lower left side of the first processor board 120, or the first extraction fan 152 and the upper right side of the first processor board Between, one or more hoods are added to guide the generated airflow. Referring to FIG. 2A and FIG. 2B, the first propulsion fan 161 located in the upper left corner of the symmetric multiprocessor computer is sucked down by the cold air before the symmetrical multiprocessor computer 1 ,, 9 1361358 A second airflow 163 is generated at a position in the upper left corner of the second processor board 13G, and then the cold second airflow 163 flows through the parallel fourth processor board (10) and is disposed on the third processor. After the heat is extracted, the second second airflow 163 is also discharged from the lower right side of the second processor board 13A under the pull of the second extraction fan 162. The second extraction fan is located in the lower right corner of the symmetric multiprocessor computer (10). 162 draws the hot second gas stream 163 and discharges it toward the symmetrical multiprocessor computer. In order to ensure that the second airflow flows as described above, it may be between the second propulsion fan i6i and the upper left of the second processor board 130' or on the upper right of the second and second processor fans 162 and 13 Between the installation of one or more hoods to guide the resulting airflow. Since the flow paths of the first airflow 153 and the second airflow 163 are respectively aligned with the directions of the first processor board 12 and the second new board 130, the flow paths of the first airflow 153 and the third airflow 163 are in the inter-board The different surfaces of the 11G are staggered, which is the main feature of the cooling system applied to a symmetrical multiprocessor computer with a star interconnect. The direction such as the fan pattern, the sizing speed, or the direction in which the turbulent airflow flows into and out of the fan should not be limited to the scope of the present embodiment as described above or as disclosed hereinafter. Months, ..., 3, 3, 4, and 4 are the second embodiment of the present invention. The symmetric multiprocessor electric Teng 2GG star interconnection device mainly comprises: _ interboard 2m; first processor board and second processor board 230, first processor board 22 〇 and second processor board 23 The manner in which the crucible is disposed in the outer casing 24 is the same as that of the first embodiment. The symmetrical multi-processing cooling system setting method is also related to the first-real __. The main difference between the second real complement and the first real one is that the second embodiment has two T-shaped plates 213 and 214, which are viewed from the upper side and the lower side of the t-board (10), or other feasible positions of the 4-type plate 213. And 214 are all τ type printed circuit boards, both of which can be made into a separate part of 10 1361358, and then interconnected with the intermediate plate 210 by the side-to-side connection, or become an extension of the intermediate plate 210. The τ-type boards 213 and 214 are used to provide interconnection between the system and the storage, the power supply, the expansion card, and the friend network interface. Further, since the T-shaped plates 213 and 214 form four openings at the position of the fan, the flow of the airflow is not blocked. The total number of T-plates of the symmetrical multiprocessor computer 200 is related to the actual implementation, and in some cases only one is sufficient. In addition, the T-plates 213 and 214 may also be used only for support without any circuit. Please refer to FIG. 5 and FIG. 6 for a third embodiment of the present invention. The star interconnect device of the symmetric multiprocessor computer 300 mainly includes: an intermediate board 31〇, a first processor board 32〇, a second processor board 330 and two T-shaped boards 313 and 314 disposed on the outer casing 340. The mode is the same as that of the second embodiment. The cooling system of the symmetrical multiprocessor computer 300 is also arranged in the same manner as the second embodiment. In Figures 5 and 6, two hard disk modules 351 and 352, two power supply modules 353 and 354, and two expansion card modules 355 and 356 are disposed on the star interconnect. Hard disk modules 351 and 352' power supply modules 353 and 354 and expansion card modules, groups 355 and 356 are supported by sufficient support members. In addition, each hard disk module located in the hard disk module 351 and the forum has sufficient passages to avoid snagging the airflow. Similarly, each expansion card located in the expansion card modules 355 and 356 is also the same. Although the technology of the present invention has been disclosed in the preferred embodiments as above, it is not intended to limit the present invention, and any skilled person skilled in the art will be able to cover some of the changes and opacity that are removed from the spirit of the present invention. In the present invention, the scope of protection of the current month is subject to the definition of the patent application scope. 11 丄妁 1358 [Simple Description of the Drawing] The 1A circle is the first embodiment of the present invention. A front perspective view of a symmetric multiprocessor computer with a star interconnect and cooling system. Figure 1B is a front view of Figure 1A. Fig. 2A is a rear perspective view of the symmetric multiprocessor computer of the first embodiment of the present invention. Fig. 2B is a rear view of the symmetric multiprocessor computer of the first embodiment of the present invention. Figure 3A is a front perspective view of a symmetrical multiprocessor computer having a star interconnect and a cooling system in accordance with a second embodiment of the present invention. Figure 3B is a front view of Figure 3A. Figure 4A is a rear perspective view of a symmetric multiprocessor computer in accordance with a second embodiment of the present invention. Figure 4B is a rear elevational view of a symmetric multiprocessor computer in accordance with a second embodiment of the present invention. Figure 5 is a front perspective view of a symmetrical multiprocessor computer having a star interconnect and a cooling system in accordance with a third embodiment of the present invention. • Figure 6' is a perspective view of the third embodiment of the present invention. [Description of main component symbols] 100.........symmetric multiprocessor computer 110 .........intermediate board 111 ......... front connector 112 .. . . . after connector 120 ... ... first processor board 121 ... ... first connector 122 ... ... The second connector 12 1361358 130...the second processor board 140...the outer casing 151 ....the first propulsion fan 152 .... ..... first extraction fan 153 ... ... first air flow 161 ... ... second propulsion fan 162 ......... · · second extraction fan 163 .........the second airflow 200 ...the symmetric multiprocessor computer (the second embodiment of the present invention) 210...the middle plate 213 .........T-shaped plate 214 .... T-shaped plate 220 .... First processor board 230 ......... Second processor board 240 ... ... housing 300 ... symmetrical multiprocessor computer (third embodiment of the invention) 310 .... Intermediate plate 313 ... T-shaped plate 314 ... ... T-shaped plate 320 ... ... first processor board 330 ... ...second processor board 13 1361358 340 ......... housing 351 ......... hard disk module 352 ......... hard disk module 353 .........power supply The power module 355 sets 354 ......... ......... ......... expansion card module 356 tooth-mold expansion card