CN110727631B - H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates - Google Patents
H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates Download PDFInfo
- Publication number
- CN110727631B CN110727631B CN201910863825.9A CN201910863825A CN110727631B CN 110727631 B CN110727631 B CN 110727631B CN 201910863825 A CN201910863825 A CN 201910863825A CN 110727631 B CN110727631 B CN 110727631B
- Authority
- CN
- China
- Prior art keywords
- middle plate
- ins
- plug
- left middle
- right middle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F15/00—Digital computers in general; Data processing equipment in general
- G06F15/08—Digital computers in general; Data processing equipment in general using a plugboard for programming
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multi Processors (AREA)
Abstract
The invention discloses an H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates. The method comprises the steps of vertically placing a left middle plate and a right middle plate, and approaching one ends of the left middle plate and the right middle plate to each other, so that the left middle plate and the right middle plate can form a surface; node plug-ins are respectively and horizontally placed in equal quantity on one side, far away from the right middle plate, of the front side and the rear side of the left middle plate and on one side, far away from the left middle plate, of the front side and the rear side of the right middle plate; a plurality of horizontal exchange plug-ins are horizontally placed on one side surface of the left middle plate and the right middle plate, which are close to each other, so that the horizontal exchange plug-ins can be connected with the left middle plate and the right middle plate at the same time, and a plurality of vertical exchange plug-ins are vertically placed on the other side surface of the left middle plate and the right middle plate, which are close to each other, so that vertical exchange plug-ins are arranged on the left middle plate and the right middle plate. The invention improves the assembly density of the nodes and the exchange chip, reduces the interconnection transmission distance between the nodes and the exchange chip, and improves the interconnection rate.
Description
Technical Field
The invention relates to the field of design of computing and storage systems, in particular to an H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates.
Background
The supercomputers, clusters, data centers and other large-scale computing and storage systems cannot realize the functions of the supercomputers, clusters, data centers and other large-scale computing and storage systems through a single node (node: CPU processor). With the development demands of higher and higher speed of computing and storage systems, the number of nodes is increased in an explosive manner. The high-efficiency high-speed high-density multi-node interconnection assembly technology is an important component in system design.
In general, one switching chip is responsible for external interconnection of a plurality of nodes, and thus, interconnection paths between nodes are divided into three types: the first is the same exchange chip interconnection, the second is the exchange chip interconnection on the same exchange plug-in, and the third is the exchange chip interconnection on different exchange plug-in. The point-to-point connection between nodes can be realized by the middle plate design meeting the three types of interconnection requirements.
However, the above-described existing horizontal and vertical interconnect methods do not meet the increasingly high interconnect density and signal transmission rate requirements. First, the number of interconnect nodes within a single midplane is limited. The number of node plug-ins and the number of nodes in the node plug-ins determine the total number of interconnected nodes in one middle plate. The size of the middle plate is limited by the printed board processing technology, and the length, width and thickness of the middle plate are required to be in a range suitable for mass production. The number of node connectors that a midplane can carry is limited by the size of the connector device, i.e., the device to which the connector is connected to the midplane. The size of the connector device is increased along with the increase of the signal rate, and the number of the connectors which can be carried by a single middle board is limited. Second, vertical orthogonality has the advantage of low loss relative to shorter horizontal interconnect distances because high-speed signal lines may not be routed within the midplane, but the interconnect density is lower than horizontal interconnects.
Disclosure of Invention
The invention aims to provide an H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates, which is applied to a large-scale computing or storage system.
According to a first aspect of the present invention, there is provided an H-type assembly method based on a dual midplane orthogonal and non-orthogonal heterogeneous interconnect, comprising:
the left middle plate and the right middle plate are vertically placed, and one ends of the left middle plate and the right middle plate are mutually close to each other, so that the left middle plate and the right middle plate can form a surface; node plug-ins are respectively and horizontally placed in equal quantity on one side, far away from the right middle plate, of the front side and the rear side of the left middle plate and on one side, far away from the left middle plate, of the front side and the rear side of the right middle plate; a plurality of horizontal exchange plug-ins are horizontally placed on one side surface of the left middle plate and the right middle plate, which are close to each other, so that the horizontal exchange plug-ins can be connected with the left middle plate and the right middle plate at the same time, and a plurality of vertical exchange plug-ins are vertically placed on the other side surface of the left middle plate and the right middle plate, which are close to each other, so that vertical exchange plug-ins are arranged on the left middle plate and the right middle plate.
Further, when the step S12 is executed, the number of node plug-ins is set to be 4n, and the node plug-ins set on each side surface of the left middle plate and the right middle plate are numbered 1, 2 and 3. When step S13 is performed, the number of horizontal switch cards is set to n, and the horizontal switch cards are numbered 1, 2, 3..n, respectively, from top to bottom.
Further, after executing step S13, the node plug-ins with the same number on the front and rear sides of the left middle board are in non-orthogonal interconnection with the switch chips on the horizontal switch plug-ins with the same number; the node plug-ins with the same number on the front side and the rear side of the right middle plate are in non-orthogonal interconnection with the exchange chips on the horizontal exchange plug-ins with the same number.
Further, after step S13 is executed, the switch chips on the horizontal switch cards with different numbers are all orthogonally interconnected with the switch chips on the vertical switch card.
The beneficial effects of the invention are as follows: 1. through being H type overall arrangement with two medium plates, node plug-in components, horizontal exchange plug-in components, perpendicular exchange plug-in components, select type and pin distribution of all connector components well for the front and back node plug-in components that are located same medium plate height are favorable to saving wiring space, can also make horizontal exchange plug-in components and accurate alignment of two medium plates. 2. Because the dual middle plate design is adopted, the positions of node plug-ins are designed in the front and the back of each middle plate, the interconnection density is greatly improved, the node plug-ins and the switching chips positioned on the horizontal switching plug-ins realize non-orthogonal interconnection, the switching chips between the horizontal switching plug-ins and the vertical switching plug-ins realize orthogonal interconnection, the assembly density of the nodes and the switching chips is improved, the interconnection transmission distance between the nodes and the switching chips is reduced, and the interconnection rate is improved.
Drawings
FIG. 1 is a flow chart of an H-type assembly method based on a dual midplane orthogonal and non-orthogonal heterogeneous interconnect, according to one embodiment of the invention;
FIG. 2 is a schematic diagram of an H-type assembly method based on orthogonal and non-orthogonal heterogeneous interconnections of a dual midplane according to one embodiment of the invention;
wherein, 1-node plug-in, 2-horizontal exchange plug-in, 3-vertical exchange plug-in, 4-left middle plate, 5-right middle plate.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 shows a flow of an H-type assembly method based on a dual midplane orthogonal and non-orthogonal heterogeneous interconnect according to an embodiment of the invention, comprising:
s11, placing the left middle plate and the right middle plate vertically, and approaching one ends of the left middle plate and the right middle plate to each other, so that the left middle plate and the right middle plate can form a surface.
And S12, respectively horizontally and equally arranging node plug-ins on one side, far from the right middle plate, of the front side and the rear side of the left middle plate and on one side, far from the left middle plate, of the front side and the rear side of the right middle plate.
S13, horizontally placing a plurality of horizontal exchange plug-ins on one side surface of the left middle plate and the right middle plate, which are close to each other, so that the horizontal exchange plug-ins can be connected with the left middle plate and the right middle plate simultaneously, and vertically placing a plurality of vertical exchange plug-ins on the other side surface of the left middle plate and the right middle plate, which are close to each other, so that the vertical exchange plug-ins are arranged on the left middle plate and the right middle plate.
As a preferred embodiment, when step S12 is performed, the number of node plug-ins is set to 4n, and the node plug-ins set on each side of the left middle plate and the right middle plate are numbered 1, 2, 3..n from top to bottom respectively; when step S13 is performed, the number of horizontal switch cards is set to n, and the horizontal switch cards are numbered 1, 2, 3..n, respectively, from top to bottom.
As a preferred embodiment, after step S13 is performed, the node cards with the same number on the front and rear sides of the left middle board are in non-orthogonal interconnection with the switch chips on the horizontal switch cards with the same number; the node plug-ins with the same number on the front side and the rear side of the right middle plate are in non-orthogonal interconnection with the exchange chips on the horizontal exchange plug-ins with the same number.
As a preferred embodiment, after step S13 is performed, the switch chips on the horizontal switch cards of different numbers are all orthogonally interconnected with the switch chips on the vertical switch card.
Furthermore, those skilled in the art will appreciate that while some embodiments herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the claims, any of the claimed embodiments may be used in any combination.
Those of ordinary skill in the art will appreciate that: the above embodiments are only for illustrating the technical solution of the present invention, not for limiting the same, and although the present invention has been described in detail with reference to the above embodiments, it should be understood by those of ordinary skill in the art that: it is still possible to modify the technical solutions described in the foregoing embodiments or to make equivalent substitutions for some or all of the technical features thereof, without departing from the essence of the corresponding technical solutions from the scope of the present invention defined by the claims.
Claims (1)
1. An H-shaped assembly method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates is characterized by comprising the following steps:
s11, vertically placing a left middle plate and a right middle plate, and enabling one ends of the left middle plate and the right middle plate to be close to each other, so that the left middle plate and the right middle plate can form a surface;
s12, horizontally and equally arranging node plug-ins on one side, far away from the right middle plate, of the front side and the rear side of the left middle plate and one side, far away from the left middle plate, of the front side and the rear side of the right middle plate respectively;
s13, horizontally placing a plurality of horizontal exchange plug-ins on one side surface of the left middle plate and the right middle plate, which are close to each other, so that the horizontal exchange plug-ins can be connected with the left middle plate and the right middle plate at the same time, and vertically placing a plurality of vertical exchange plug-ins on the other side surface of the left middle plate and the right middle plate, which are close to each other, so that the vertical exchange plug-ins are arranged on the left middle plate and the right middle plate;
when the step S12 is executed, the number of the node plug-ins is set to be 4n, and the node plug-ins set on each side surface of the left middle plate and the right middle plate are numbered 1, 2 and 3. When executing the step S13, setting the number of the horizontal exchange plug-ins as n, and numbering the horizontal exchange plug-ins as 1, 2 and 3.
After executing the step S13, the node plug-ins with the same number on the front side surface and the rear side surface of the left middle plate are in non-orthogonal interconnection with the switching chips on the horizontal switching plug-ins with the same number; the node plug-ins with the same number on the front side surface and the rear side surface of the right middle plate are in non-orthogonal interconnection with the exchange chips on the horizontal exchange plug-ins with the same number;
after step S13 is executed, the switch chips on the horizontal switch cards with different numbers are orthogonally interconnected with the switch chips on the vertical switch card.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910863825.9A CN110727631B (en) | 2019-09-12 | 2019-09-12 | H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910863825.9A CN110727631B (en) | 2019-09-12 | 2019-09-12 | H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110727631A CN110727631A (en) | 2020-01-24 |
CN110727631B true CN110727631B (en) | 2023-08-08 |
Family
ID=69218105
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910863825.9A Active CN110727631B (en) | 2019-09-12 | 2019-09-12 | H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110727631B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002057829A1 (en) * | 2001-01-17 | 2002-07-25 | The Whitaker Corporation | Optical cross connect |
CN101083599A (en) * | 2007-07-13 | 2007-12-05 | 清华大学 | P2i interconnection structure based extendable router |
CN102695393A (en) * | 2012-05-04 | 2012-09-26 | 华为技术有限公司 | Perpendicular orthogonal interconnection system and communication device |
CN102882777A (en) * | 2012-09-28 | 2013-01-16 | 无锡江南计算技术研究所 | Full-crossover network interconnection assembly structure and full-crossover network interconnection assembly method |
CN105743819A (en) * | 2010-09-16 | 2016-07-06 | Iii控股第2有限责任公司 | Performance and power optimized computer system architectures and methods leveraging power optimized tree fabric interconnect |
-
2019
- 2019-09-12 CN CN201910863825.9A patent/CN110727631B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002057829A1 (en) * | 2001-01-17 | 2002-07-25 | The Whitaker Corporation | Optical cross connect |
CN101083599A (en) * | 2007-07-13 | 2007-12-05 | 清华大学 | P2i interconnection structure based extendable router |
CN105743819A (en) * | 2010-09-16 | 2016-07-06 | Iii控股第2有限责任公司 | Performance and power optimized computer system architectures and methods leveraging power optimized tree fabric interconnect |
CN102695393A (en) * | 2012-05-04 | 2012-09-26 | 华为技术有限公司 | Perpendicular orthogonal interconnection system and communication device |
CN102882777A (en) * | 2012-09-28 | 2013-01-16 | 无锡江南计算技术研究所 | Full-crossover network interconnection assembly structure and full-crossover network interconnection assembly method |
Non-Patent Citations (1)
Title |
---|
刘杰等.多核处理器存储体系分析.《第十五届计算机工程与工艺年会暨第一届微处理器技术论坛论文集》.2011,第48-52页. * |
Also Published As
Publication number | Publication date |
---|---|
CN110727631A (en) | 2020-01-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7103753B2 (en) | Backplane system having high-density electrical connectors | |
US8116332B2 (en) | Switch arbitration | |
JP2543312B2 (en) | Computer assembly | |
CN107766282B (en) | Design method of eight-path server back plate and double-buckle-plate interconnection system | |
US20070124529A1 (en) | Subrack with front and rear insertion of AMC modules | |
WO2007062236A2 (en) | Monolithic backplane having a first and second portion | |
CN110727631B (en) | H-shaped assembling method based on orthogonal and non-orthogonal heterogeneous interconnection of double middle plates | |
KR20040052223A (en) | Rack mounted routers | |
CN111258948B (en) | Novel GPU server system | |
CN112948316A (en) | AI edge computing all-in-one machine framework based on network interconnection | |
CN108183872B (en) | Switch system and construction method thereof | |
CN109739802B (en) | Computing cluster and computing cluster configuration method | |
EP2350719A1 (en) | Optical connector interconnection system and method | |
CN109213285A (en) | A kind of edge calculations server architecture | |
CN101741711A (en) | Fabric switch card independent from MCH and Micro Telecommunications Computing Architecture system | |
CN114817111A (en) | Embedded intelligent computing device supporting multi-card parallel | |
CN216647351U (en) | High-speed backboard interconnection system | |
CN218481828U (en) | Multi-node server and multi-node server system | |
CN219625889U (en) | Robot controller | |
CN205263720U (en) | Multi -path server extension module | |
CN217543768U (en) | Module interconnection backplate of integrated VPX trench and PCIe expansion slot | |
CN220367595U (en) | Mainboard of double-path server and server | |
WO2024140085A1 (en) | Switching network system and switching interconnection device | |
CN219145507U (en) | Modularized edge computing gateway suitable for indoor and outdoor scenes | |
CN112260969B (en) | Blade type edge computing equipment based on CPCI framework |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |