High-speed data transmission GPU server cluster interconnection device

Abstract

The utility model provides a high-speed data transmission GPU server cluster interconnection device, and relates to the technical field of communication equipment, the high-speed data transmission GPU server cluster interconnection device comprises a control panel and an interconnection sphere, the control panel is electrically connected with the interconnection sphere, the surface of the interconnection sphere is provided with T-shaped ring grooves at equal intervals, a plurality of groups of T-shaped sliding blocks are arranged in the T-shaped ring grooves in a sliding manner, and the T-shaped sliding blocks are connected with the control panel. An interconnection base is connected to the surface of the T-shaped sliding block, a line interface is formed in the front face of the interconnection base, a wireless interface is formed in the top face of the interconnection base, a wireless receiver is inserted into the wireless interface, a mounting shaft seat is movably connected to the bottom of the interconnection ball through a shaft, and the interconnection base at multiple positions is arranged on the surface of the interconnection ball, so that multi-angle adaptive interconnection of the GPU cluster is achieved. The method effectively prevents the problem of line bending caused by inconsistent directions of the line and the GPU server, protects the line accessed to the GPU server, adapts the line direction, and adapts the position of the GPU cluster, so that the line can be quickly found, line accumulation can be prevented, and the line can be accurately positioned and disassembled.

Description

TECHNICAL FIELD

[0001] The utility model relates to communication equipment technical field especially relates to a high -speed data transmission GPU server cluster interconnection device. BACKGROUND

[0002] According to the server cluster of Chinese open number CN208656808U, set up internet access server, internet access server connects load balancing server, disaster recovery server and identity recognition server respectively, identity recognition server is connected with subsystem server cluster A, internet access server is provided with content distribution device, subsystem server cluster A is provided with storage cluster A, internet interface and third party system application interface, subsystem server cluster A contains several subsystem servers, the utility model discloses can identify user identity and push different application, and it is convenient and fast, internet access server sets up content distribution device, solves the crowded condition of Internet network, improves the response speed of user website access.

[0003] The above patent document and prior art have the following technical problems:

[0004] 1、In the traditional GPU server cluster interconnection equipment, because the line interface position is fixed, when the layout of GPU server and the line trend are inconsistent, the line bending situation is easy to appear, and the frequent bending can cause the wire breakage, insulation layer damage and other problems in the line, thereby causing signal transmission interruption or interference, and the previous equipment is in the face of multiple GPU server interconnection, the line often winds, piles up, causes the line layout to be in disorder.

[0005] 2、The traditional interconnection device can only provide limited line interface types, and cannot meet the diversified interface requirements of GPU server cluster, and the line interface is easy to be damaged in the traditional interconnection equipment, and the effective arrangement and state indication function of the line are lacked. Utility model content

[0006] The utility model aims at solving the problems of the line bending damage and line layout disorder in the GPU cluster interconnection device in the prior art, and the single interconnection function, and proposes a high-speed data transmission GPU server cluster interconnection device.

[0007] In order to achieve the above object, the utility model discloses the following technical scheme: a high -speed data transmission GPU server cluster interconnection device, including control panel and interconnection sphere, control panel and interconnection sphere electric connection, the surface of interconnection sphere is equipped with T type ring groove at equal interval, the inside of T type ring groove is equipped with a plurality of T type sliding blocks of sliding, the surface of T type sliding block is equipped with interconnection pedestal, the front of interconnection pedestal is equipped with line interface, the top of interconnection pedestal is equipped with wireless interface, the inside of wireless interface is equipped with wireless receiver of inserting, the bottom of interconnection sphere is equipped with mounting shaft seat of swing axle.

[0008] Preferably, the surface of the interconnection base is provided with a positioning shaft seat at both ends of one side, the surface of the positioning shaft seat is movably clamped with a torsional spring guide shaft, the surface of the torsional spring guide shaft is wound with a positioning rope, the end of the positioning rope is provided with a magic sticky hook surface, and the surface of the side away from the positioning shaft seat of the interconnection base is provided with a magic sticky hair surface.

[0009] Preferably, the inside of the interconnection base is provided with a control core plate, the bottom surface of the interconnection base is provided with a control switch, the line interface is connected with the control core plate through the control switch, the control switch is provided with an indicating ring lamp at the outer circumferential position, and the indicating ring lamp is electrically connected with the line interface.

[0010] Preferably, the inside of the interconnection sphere is provided with a control cavity, the inside of the control cavity is provided with a control mainboard, the surface of the control mainboard is provided with a line main body, and the line main body is connected with the control core plate through the back surface of the T-shaped sliding block.

[0011] Preferably, the control cavity of the interconnection sphere and the inner wall of the T-shaped ring groove are provided with a line arc groove, and the line main body penetrates along the surface of the line ring groove.

[0012] Preferably, the T-shaped ring grooves are distributed at equal intervals along the surface of the interconnection sphere, the T-shaped sliding blocks and the interconnection bases are one-to-one corresponding, and the interface types of the line interfaces on the surfaces of the adjacent interconnection bases are different.

[0013] Preferably, the vertical center line of the mounting shaft seat and the vertical center line of the interconnection sphere coincide with each other, and the wireless receivers and the wireless interfaces are one-to-one corresponding.

[0014] Beneficial effects

[0015] The utility model discloses a line access control is carried out to the interconnection sphere and control panel connection, makes the position of the interconnection sphere surface through the adjustment of multiple T type ring groove adjustment T type sliding block, and then carries out the adjustment of interconnection base, makes the interconnection base surface connection GPU server line connection, realizes the multi -angle adaptation interconnection of GPU cluster through the interconnection base of multiple positions of interconnection sphere surface setting, effectively prevents the problem that the line is bent and causes the line and GPU server direction to be inconsistent, protects the line of GPU server access and adapts the line direction, and the position of GPU cluster is adapted simultaneously, and the line multidirectional design is convenient for line quick search, prevents line accumulation, and convenient for line accurate positioning and dismounting access, convenient for accurate positioning and dismounting access of line, in the construction and maintenance of large -scale GPU server cluster, operating personnel can easily identify and manage each line, greatly improve work efficiency, reduce maintenance cost and error probability.

[0016] The utility model discloses a line access control is carried out to the interconnection sphere and control panel connection, makes the position of the interconnection sphere surface through the adjustment of multiple T type ring groove adjustment T type sliding block, and then carries out the adjustment of interconnection base, makes the interconnection base surface connection GPU server line connection, realizes the multi -angle adaptation interconnection of GPU cluster through the interconnection base of multiple positions of interconnection sphere surface setting, effectively prevents the problem that the line is bent and causes the line and GPU server direction to be inconsistent, protects the line of GPU server access and adapts the line direction, and the position of GPU cluster is adapted simultaneously, and the line multidirectional design is convenient for line quick search, prevents line accumulation, and convenient for line accurate positioning and dismounting access, convenient for accurate positioning and dismounting access of line, in the construction and maintenance of large -scale GPU server cluster, operating personnel can easily identify and manage each line, greatly improve work efficiency, reduce maintenance cost and error probability. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is the three -dimensional structure diagram of the utility model;

[0018] Figure 2 It is the three -dimensional structure diagram of the utility model; Figure 1 It is the A place enlarged view of the utility model;

[0019] Figure 3 It is the front view of the interconnection sphere of the utility model;

[0020] Figure 4 It is the surface structure diagram of the interconnection sphere of the utility model;

[0021] Figure 5 is a mutual connection base structure diagram of the utility model;

[0022] Figure 6 is a mutual connection base isometric view of the utility model;

[0023] Figure 7 is a mutual connection base split structure diagram of the utility model.

[0024] Legend:

[0025] 1, control panel;2, mounting shaft seat;3, mutual connection sphere;4, control cavity;5, control mainboard;6, T type ring groove;7, T type sliding block;8, line arc groove;9, line main body;10, mutual connection base;11, control core plate;12, line interface;13, wireless interface;14, wireless receiver;15, control switch;16, indicating ring lamp;17, positioning shaft seat;18, torsional spring guide shaft;19, positioning rope;20, magic sticky wool surface;21, magic sticky hook surface. DETAILED DESCRIPTION

[0026] In order to make the technical means, creation features, purposes and effects realized by the utility model easy to understand, the utility model is further described below in combination with specific embodiments and drawings, but the following embodiments are only preferred embodiments of the utility model, not all. Based on the embodiments in the embodiments, other embodiments obtained by those skilled in the art without making creative efforts all belong to the protection scope of the utility model.

[0027] The specific embodiments of the utility model are described below in combination with the drawings. Specific embodiment one:

[0029] Reference Figures 1 to 7The application discloses a high-speed data transmission GPU server cluster interconnection device which comprises a control panel 1 and an interconnection sphere 3, wherein the control panel 1 is electrically connected with the interconnection sphere 3, T-shaped ring grooves 6 are arranged on the surface of the interconnection sphere 3 at equal intervals, a plurality of groups of T-shaped sliding blocks 7 are slidably arranged in the T-shaped ring grooves 6, interconnection bases 10 are connected to the surfaces of the T-shaped sliding blocks 7, line interfaces 12 are arranged on the front surfaces of the interconnection bases 10, wireless interfaces 13 are arranged on the top surfaces of the interconnection bases 10, wireless receivers 14 are inserted into the wireless interfaces 13, a mounting shaft seat 2 is movably connected to the bottom of the interconnection sphere 3, the control panel 1 is electrically connected with the interconnection sphere 3 and is used for controlling and monitoring the whole device, the T-shaped ring grooves on the surface of the interconnection sphere 3 provide sliding tracks for the T-shaped sliding blocks, the T-shaped sliding blocks can drive the interconnection bases 10 to move on the ring grooves so as to adjust the positions of the interconnection bases 10, the line interfaces 12 on the interconnection bases 10 are used for connecting the lines of GPU servers, the wireless interfaces 13 can be inserted into the wireless receivers 14 to realize wireless access functions, and the mounting shaft seat 2 facilitates the mounting and fixing of the whole interconnection sphere 3; after the device is mounted, according to the actual layout and line connection requirements of the GPU servers, the T-shaped sliding blocks are manually pushed to slide in the T-shaped ring grooves, the interconnection bases 10 are moved to appropriate positions so as to be connected with the line interfaces 12 of the GPU servers, and simultaneously, if the wireless access functions are needed to be used, the wireless receivers 14 can be inserted into the wireless interfaces 13 to realize wireless data transmission; the structural design makes the interconnection device capable of flexibly adapting to the layouts of different GPU server clusters, the positions of the interconnection bases 10 can be adjusted, the line bending problem caused by the inconsistent directions of the lines and the GPU servers can be effectively avoided, the lines are protected, the connection and management of the lines are facilitated, and the stability and reliability of the whole system are improved.

[0030] The surface side of the interconnection base 10 is provided with a positioning shaft seat 17 at both ends, the surface of the positioning shaft seat 17 is movably clamped with a torsion spring guide shaft 18, the surface of the torsion spring guide shaft 18 is wound with a positioning rope 19, the end of the positioning rope 19 is provided with a magic sticky hook surface 21, the surface of the side of the interconnection base 10 away from the positioning shaft seat 17 is provided with a magic sticky hair surface 20, the positioning shaft seat 17 is used for installing the torsion spring guide shaft 18, one end of the positioning rope 19 wound on the torsion spring guide shaft 18 is provided with a magic sticky hook surface 21, when it is needed to tighten the line, the positioning rope 19 is wound around the line, and then the magic sticky hook surface 21 and the magic sticky hair surface 20 on the other side of the interconnection base 10 are pasted together, so that the line is fixed. The torsion spring guide shaft 18 can ensure the tension of the positioning rope 19 to a certain extent, and prevent it from loosening. After the line of the GPU server is connected to the line interface 12 of the interconnection base 10, the operator pulls out the positioning rope 19 from the torsion spring guide shaft 18, winds it around the line, and then pastes the magic sticky hook surface 21 on the magic sticky hair surface 20, completing the tightening and positioning of the line. If it is needed to replace or maintain the line, the magic sticky hook surface 21 can be easily operated by being unwound, so that the same type or the access GPU line processing the same area and function can be easily operated, the line is more orderly, maintenance and searching are facilitated, the efficiency of line maintenance is improved, and the troubleshooting time and difficulty caused by line disorder are reduced.

[0031] The inside of the interconnection base 10 is provided with a control core plate 11, the bottom surface of the interconnection base 10 is provided with a control switch 15, the line interface 12 is connected with the control core plate 11 through the control switch 15, the outside circumferential position of the control switch 15 is provided with an indicating ring lamp 16, the indicating ring lamp 16 is electrically connected with the line interface 12, the control core plate 11 is the core control component of the interconnection base 10, the line interface 12 is connected with the control core plate 11 through the control switch 15, and the control switch 15 is used for controlling the on-off of the line interface 12. When the line interface 12 has data transmission, the indicating ring lamp 16 is lit to display the working state of the interface. In the normal use process, when it is needed to connect or disconnect a GPU server line, the control switch 15 can be operated to realize the connection or disconnection without directly plugging or unplugging the line interface 12, so that the plugging and unplugging times of the line interface 12 are reduced, and the service life is prolonged. At the same time, the operator can observe the state of the indicating ring lamp 16 to quickly understand the working conditions of the line interface 12, so that problems can be found and solved in time. The control switch 15 controls the on-off of the line interface 12, the plugging and unplugging times are reduced, the risk of damage of the line interface 12 caused by frequent plugging and unplugging is reduced, and the reliability of the device is improved. The setting of the indicating ring lamp 16 facilitates the operator to monitor the working state of the line interface 12, and improves the maintainability and troubleshooting efficiency of the system.

[0032] The control cavity 4 inside the interconnected sphere 3 is provided with a control main plate 5, and the surface of the control main plate 5 is provided with a line main body 9. The line main body 9 is connected to the control core plate 11 through the back of the T-shaped sliding block 7. The control main plate 5 inside the control cavity 4 is the control center of the entire interconnected device, responsible for coordinating and managing the data transmission of each interconnected base 10. The line main body 9 is arranged on the control main plate 5 and connected to the control core plate 11 inside the interconnected base 10 through the back of the T-shaped sliding block, realizing the transmission of data from the control main plate 5 to each interconnected base 10. When the GPU server transmits data through the line interface 12 or the wireless interface 13 of the interconnected base 10, the data first reaches the control main plate 5 inside the interconnected sphere 3, and then is transmitted to the control core plate 11 of the corresponding interconnected base 10 through the line main body 9 and the T-shaped sliding block. The control core plate 11 further processes and distributes the data, ensuring that the data can be accurately transmitted between the GPU server clusters. This internal structure design ensures the stability and reliability of data transmission, reduces interference and loss in the data transmission process through reasonable wiring and connection method, improves the data transmission efficiency of the entire GPU server cluster, and provides strong support for high-speed data transmission.

[0033] The line arc groove 8 is provided between the control cavity 4 of the interconnected sphere 3 and the inner wall of the T-shaped ring groove 6, and the line main body 9 penetrates along the surface of the line ring groove. The design of the line arc groove 8 enables the line main body 9 to penetrate along the inner wall of the T-shaped ring groove, avoiding the disorderly winding of the line inside the interconnected sphere 3. The line main body 9 extends to the position of each T-shaped sliding block through the line arc groove 8, and then connects with the control core plate 11 of the interconnected base 10, realizing the transmission of data. In the assembly process of the device, the line main body 9 is pre-laid along the line arc groove 8. When the T-shaped sliding block moves the interconnected base 10, the line main body 9 can stretch and bend in the arc groove accordingly with the movement of the sliding block, always maintaining the connection with the interconnected base 10, ensuring the continuity of data transmission. The arrangement of the line arc groove 8 makes the arrangement of the line more regular, reduces the mutual interference between the lines, and improves the safety and reliability of the line. At the same time, it also facilitates the assembly and maintenance of the device, and reduces the failure rate caused by line problems.

[0034] The T-shaped ring grooves 6 are distributed equidistantly along the surface of the interconnected sphere 3, and the T-shaped sliders 7 correspond to the interconnected bases 10 one by one. The interface types of the line interfaces 12 on the surfaces of adjacent interconnected bases 10 are different. The plurality of equidistantly distributed T-shaped ring grooves can install a plurality of interconnected bases 10, and each interconnected base 10 can be provided with different types of line interfaces 12, such as HDMI, USB, Ethernet, etc., according to actual needs, to meet the connection requirements of different devices in the GPU server cluster. In actual application, according to the interface type and layout of the GPU server, the interconnected base 10 at a suitable position is selected, and the corresponding line is connected to the corresponding interface. Since the interconnected base 10 can be flexibly moved on the T-shaped ring groove, the positions of the various interfaces can be conveniently adjusted to accurately dock with the interfaces of the GPU server, realizing efficient data transmission. By providing a plurality of types of line interfaces 12 and flexible layout of the interconnected base 10, the diversified connection requirements of different GPU server clusters can be met, without the need for additional switching equipment, reducing cost, improving the integration and adaptability of the system, and making the construction of the entire GPU server cluster more convenient and efficient.

[0035] The vertical center line of the mounting shaft seat 2 coincides with the vertical center line of the interconnected sphere 3, and the wireless receiver 14 corresponds to the wireless interface 13 one by one. The vertical center line of the mounting shaft seat 2 coincides with the vertical center line of the interconnected sphere 3, which can ensure that the interconnected sphere 3 is in a stable balanced state after installation, reducing the shaking and damage of the equipment caused by unbalanced installation. At the same time, each wireless interface 13 corresponds to a wireless receiver 14, ensuring the accuracy and stability of wireless data transmission.

[0036] In summary, the high-speed data transmission GPU server cluster interconnection device has remarkable advantages in improving data transmission efficiency, protecting lines, adapting to different server layouts, and facilitating maintenance and management, etc. through ingenious structural design and functional configuration, and can meet the needs of GPU server clusters for high-speed, stable and reliable interconnection, and has high practical value and market prospect. Specific embodiment two:

[0038] Reference Figures 1 to 7 According to the contents in the above specific embodiments, the following contents are further disclosed:

[0039] The data signal transmission relationship between the control mainboard 5 and each component is as follows:

[0040] The relationship between the control mainboard 5 and the line main body 9 and the control core board 11: when the GPU server transmits data through the interconnection base 10, the data first enters the control core board 11 of the interconnection base 10 from the external line interface 12 or the wireless interface 13, and then is transmitted to the control mainboard 5 through the line main body 9, or data is transmitted from the control mainboard 5 to the control core board 11 of each interconnection base 10, so as to realize the exchange and transmission of data between the GPU server clusters.

[0041] The relationship between the control core board 11 and the line interface 12, the control switch 15 and the indicating ring lamp 16: inside the interconnection base 10, the control core board 11 is the core control component. The line interface 12 is connected with the control core board 11 through the control switch 15, and the control switch 15 is used for controlling the on-off of the line interface 12. When the line interface 12 has data transmission, the indicating ring lamp 16 electrically connected with the line interface 12 is lighted, and the data from the line interface 12 reaches the control core board 11 through the control switch 15, or is transmitted from the control core board 11 to the line interface 12 through the control switch 15, and the indicating ring lamp 16 displays according to the data transmission state of the line interface 12, so that the user can intuitively understand the working condition of the line interface 12.

[0042] The relationship between the wireless interface 13 and the wireless receiver 14 and the control core board 11: the wireless interface 13 is arranged on the front face of the interconnection base 10, and the wireless receiver 14 is plugged into the inside of the wireless interface 13. The wireless receiver 14 receives external wireless signals and transmits the signals to the control core board 11, and the control core board 11 transmits the data to the control mainboard 5 through the line main body 9 for processing; conversely, the data transmitted from the control mainboard 5 can also be sent to the wireless receiver 14 through the control core board 11, and the wireless receiver 14 sends out the data through the wireless mode, so as to realize the wireless data transmission between the GPU server clusters.

[0043] The control panel 1 displays and controls the state of the interconnection base 10 on the surface of the interconnection sphere 3.

[0044] Connection state display: the control panel 1 is electrically connected with the interconnection sphere 3, and can receive the state information from each interconnection base 10. For example, through the communication with the control core board 11, whether the connection state of the line interface 12 and the wireless interface 13 has a device connected and whether the data transmission state is in data transmission is obtained. For the line interface 12, the control panel 1 can determine the connection and working condition according to the state of the control switch 15 and the indicating ring lamp 16 fed back by the control core board 11; for the wireless interface 13, whether it is normally connected and transmits data can be determined according to the working state of the wireless receiver 14 and the data transmission condition of the control core board 11. Then, the connection state of each interconnection base 10 is displayed on the display screen of the control panel 1 in the form of graphics, characters or indicating lamps, so that the user can understand at a glance.

[0045] Control operation: The control panel 1 can control the interconnected base 10 on the surface of the interconnected sphere 3. For example, by sending instructions to the control core board 11 to control the on-off of the control switch 15, remote control of the line interface 12 is realized, and the user does not need to directly operate the control switch 15 on the interconnected base 10, at the same time, the control panel 1 can also configure and manage the wireless interface 13 and the wireless receiver 14, such as setting the wireless communication parameters, to ensure the stability and reliability of wireless data transmission.

[0046] In actual use, the control mainboard 5 and the control core board 11 can have the following structure:

[0047] In actual use, for the control mainboard 5 of such a high-speed data transmission GPU server cluster interconnection device, a mainboard similar to Intel C621 chipset can be selected. The Intel C621 chipset supports high-speed PCI-Express interface, which can meet the bandwidth demand of large amount of data transmission between GPU servers. It has multiple PCI-Express channels, which can connect multiple devices and provide sufficient interfaces for communication with multiple interconnected bases 10.

[0048] Processor slot: There is one or more high-performance processor slots on the control mainboard 5 for installing server-level CPUs, which are responsible for processing various control instructions in the data transmission process, such as coordinating data flow between different interconnected bases 10 and processing network protocols.

[0049] Memory slot and cache system: equipped with multiple memory slots for inserting large-capacity and high-frequency server memories. Memory is used to temporarily store data being processed, and the cache system accelerates data reading and writing to ensure that the control mainboard 5 can quickly respond to data requests from each interconnected base 10.

[0050] High-speed communication interface: with multiple high-speed PCI-Express interfaces connected to the line body 9, which transmits data to the control core board 11 of each interconnected base 10 through the line body 9. In addition, it may also include Ethernet interface, USB interface, etc. for management and configuration connection with external devices.

[0051] Storage interface: usually has SATA interface or NVMe interface for connecting storage devices. These storage devices can store control programs, configuration files and data cache, etc. to ensure data integrity and recoverability during system operation.

[0052] Power Management Module: Contains complex power management circuitry to provide a stable power supply to various components on the control motherboard 5. It also intelligently manages the power supply, adjusting the power output of each component according to the system load to achieve energy saving and stable operation.

[0053] The control chip 11 can be a customized control chip 11 based on an FPGA (Field Programmable Gate Array), such as a control chip 11 based on a Xilinx Kintex-7 series FPGA chip. The FPGA chip has high programmability and can be flexibly designed according to the specific functional requirements of the interconnect base 10.

[0054] Structural features:

[0055] FPGA Core Chip: The FPGA chip is the core of the control chip board 11. It contains a large number of programmable logic units, lookup tables (LUTs), flip-flops, and other basic circuit elements. By programming these elements, various custom logic functions can be implemented, such as data reception, processing, transmission, and coordination with components such as control switch 15 and indicator ring light 16.

[0056] Interface circuitry: Features multiple interface circuits for connecting the line interface 12, wireless interface 13, control switch 15, indicator ring light 16, and connecting to the line body 9 via a T-slider. These interface circuits enable signal level conversion, match the electrical characteristics of different interfaces, and ensure correct data transmission.

[0057] Local storage units: These include local storage units of a certain capacity, such as SRAM (Static Random Access Memory). These storage units are used to store temporary data, such as configuration parameters of line interface 12 and data cache, facilitating fast access by the FPGA chip during data processing.

[0058] Clock Management Module: Equipped with a clock management module, this module generates and manages various clock signals. Different interfaces and internal logic units may require clock signals of different frequencies to ensure synchronous operation. The clock management module can provide stable clock signals according to system requirements, ensuring the timing accuracy of data processing.

[0059] Configuration Interface: This interface is used to configure the FPGA chip. A pre-designed logic configuration file can be loaded into the FPGA chip, enabling it to operate according to specific functional logic. The logic functionality of the FPGA chip can be updated through this interface during system startup or function upgrades.

[0060] In summary:

[0061] 1、Adopt interconnection sphere 3 with control panel 1 connection carries out line access control, make interconnection sphere 3 surface pass through the adjustment T type slider 7 of multiple T type ring groove 6 of being set, further carry out the adjustment of interconnection base 10, make interconnection base 10 surface connection GPU server line and carry out connection, through the interconnection base 10 of multiple positions of interconnection sphere 3 surface setting, realize the multi-angle adaptation interconnection of GPU cluster, effectively prevent the problem of line bending caused by the inconsistency of line and GPU server direction, protect the line of GPU server access and adapt the direction of line, simultaneously adapt the position of GPU cluster, line multidirectional design, facilitate line quick search, prevent line accumulation, facilitate line accurate positioning and disassembly access, facilitate accurate positioning and disassembly access of line, in the construction and maintenance of large GPU server cluster, operating personnel can easily identify and manage each line, greatly improve work efficiency, reduce maintenance cost and error probability;

[0062] 2、Adopt interconnection sphere 3 surface connection multiple interconnection base 10 carry out the line access of GPU server, the interconnection base 10 of interconnection sphere 3 surface is equipped with multiple different types of line interface 12, can satisfy the diversified line connection demand of GPU server cluster, without additional switching equipment.Simultaneously, still support wireless receiver 14 and wireless interface 13 wireless access mode, provide more flexible selection for the interconnection of GPU cluster, enhance the compatibility and expansibility of entire system, can adapt to different configuration and application scene GPU server cluster, utilize the cooperation of positioning rope 19 and magic sticky hook surface 21, can carry out line bundling positioning to the same type or handle the same area, the function of access GPU line, facilitate maintenance and search, in addition, control switch 15 controls the on-off of line interface 12, reduces the plug and unplug frequency of line interface 12, reduces the risk of interface damage, and indication ring lamp 16 directly shows the running state of line interface 12, so that the operating personnel can understand the line connection condition and troubleshoot in time.These designs greatly improve the ease of use and maintainability of the device, reduce the downtime caused by line problems, and improve the overall availability and operating efficiency of the GPU server cluster.

[0063] In the present application, unless otherwise expressly specified and limited, the first feature is "on" or "under" the second feature, which can include direct contact between the first and second features, or indirect contact between the first and second features through another feature therebetween. Moreover, the first feature is "on", "above" and "on" the second feature, which includes the first feature being directly above and obliquely above the second feature, or only indicating that the first feature is higher than the second feature in horizontal height. The first feature is "under", "below" and "under" the second feature, which includes the first feature being directly below and obliquely below the second feature, or only indicating that the first feature is lower than the second feature in horizontal height.

[0064] The basic principle, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above embodiments, and the above embodiments and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A high-speed data transmission GPU server cluster interconnection device, comprising a control panel (1) and an interconnection sphere (3), characterized in that: The control panel (1) is electrically connected to the interconnect sphere (3). The surface of the interconnect sphere (3) is provided with T-shaped annular grooves (6) at equal intervals. Multiple sets of T-shaped sliders (7) are slidably arranged inside the T-shaped annular grooves (6). An interconnect base (10) is connected to the surface of the T-shaped sliders (7). A line interface (12) is provided on the front of the interconnect base (10). A wireless interface (13) is provided on the top surface of the interconnect base (10). A wireless receiver (14) is inserted into the wireless interface (13). A mounting shaft (2) is movably connected to the bottom of the interconnect sphere (3).

2. The high-speed data transmission GPU server cluster interconnection device according to claim 1, characterized in that: The interconnect base (10) has positioning shaft seats (17) at both ends on one side of its surface. A torsion spring guide shaft (18) is movably engaged on the surface of the positioning shaft seat (17). A positioning rope (19) is wound around the surface of the torsion spring guide shaft (18). A Velcro hook surface (21) is provided at the end of the positioning rope (19). A Velcro hook surface (20) is provided on the side of the interconnect base (10) away from the positioning shaft seat (17).

3. The high-speed data transmission GPU server cluster interconnection device according to claim 2, characterized in that: The interconnect base (10) is provided with a control core board (11) inside. The bottom surface of the interconnect base (10) is provided with a control switch (15). The line interface (12) is connected to the control core board (11) through the control switch (15). An indicator ring light (16) is provided on the outer circumference of the control switch (15). The indicator ring light (16) is electrically connected to the line interface (12).

4. The high-speed data transmission GPU server cluster interconnection device according to claim 3, characterized in that: The interconnected sphere (3) has a control cavity (4) inside, and a control motherboard (5) is provided inside the control cavity (4). The control motherboard (5) has a circuit body (9) on its surface. The circuit body (9) is connected to the control core board (11) through the back of the T-shaped slider (7).

5. The high-speed data transmission GPU server cluster interconnection device according to claim 4, characterized in that: A line arc groove (8) is provided between the control cavity (4) of the interconnected sphere (3) and the inner wall of the T-shaped annular groove (6), and the line body (9) penetrates along the surface of the line annular groove.

6. The high-speed data transmission GPU server cluster interconnection device according to claim 1, characterized in that: The T-shaped annular grooves (6) are evenly distributed along the surface of the interconnected sphere (3), and the T-shaped sliders (7) correspond one-to-one with the interconnected bases (10). The interface types of the line interfaces (12) on the surfaces of adjacent interconnected bases (10) are different.

7. The high-speed data transmission GPU server cluster interconnection device according to claim 1, characterized in that: The vertical centerline of the mounting bracket (2) coincides with the vertical centerline of the interconnecting sphere (3), and the wireless receiver (14) corresponds one-to-one with the wireless interface (13).

Images