Energy-saving GPU (Graphic Processing Unit) server cluster heat dissipation server

Abstract

The utility model provides an energy-saving GPU (Graphic Processing Unit) server cluster heat dissipation server, which relates to the technical field of heat dissipation servers and comprises a case, a cover is arranged at the top of the case, eight pressing rods are arranged on the bottom surface of the cover, eight clamping grooves are formed in the top surface of the case, four GPU main bodies are arranged in the case, and shells are arranged outside the four GPU main bodies. Eight copper pipes are arranged in the shell, the shell is tightly pressed in a clamping groove of the case through a pressing rod on the bottom face of the case cover, when the GPU main body is used in a high-load mode, the copper pipes attached to the surface of the GPU main body can rapidly transfer heat of the GPU main body to one end with fins, and the heat of the GPU main body can be rapidly dissipated by opening a cooling fan on a mounting frame at the bottoms of the fins. The cooling fan rapidly blows air through the cooling fins in a forced convection mode, the heat transfer speed is increased, and therefore the overall cooling effect is improved, and the defect that when the GPU is in a high-load state, heat dissipation cannot be effectively conducted on the GPU only through the cooling fan in a case is overcome.

Description

TECHNICAL FIELD

[0001] The utility model relates to a heat dissipation server technical field especially relates to a kind of energy-saving GPU server cluster heat dissipation server. BACKGROUND

[0002] According to the GPU server of Chinese open (announcement) number CN207020594U, it includes: a case, which is set in the fan module near the middle part inside;At least eight GPU cards are evenly divided into two groups, and are respectively arranged on the two sides of the fan module;At least two PLX / PEX chips are arranged at the bottom of the case, each PLX / PEX chip is interconnected with at least four GPU cards, so that the interconnected GPU card is parallel operation;An OCUlink connector is arranged at one end of the case, and is electrically connected with the at least two PLX / PEX chips, and the GPU card is operated by distributing tasks to each GPU card through the PLX / PEX chip, then the GPU operation result is obtained, and data is input or output by using OCUlink connector.

[0003] The above-mentioned technology and prior art GPU usually use heat dissipation fan to dissipate heat, but in the state of high load of GPU, the heat dissipation fan inside the case cannot effectively dissipate heat of GPU, thereby affecting use and service life of GPU. UTILITY MODEL CONTENT

[0004] The utility model aims at solving the shortcoming that the heat dissipation fan inside the case cannot effectively dissipate heat of GPU in the state of high load of GPU in prior art, and provides an energy-saving GPU server cluster heat dissipation server.

[0005] In order to achieve the above-mentioned purpose, the utility model adopts the following technical scheme: an energy-saving GPU server cluster heat dissipation server, including the case, the top of the case is equipped with machine cover, the bottom surface of the machine cover is equipped with eight pressure rods, the top surface of the case is equipped with eight clamping grooves, the inside of the case is equipped with four GPU main bodies, the outside of four GPU main bodies is equipped with shell, the inside of shell is equipped with eight copper pipes, the surface of both sides of the case is equipped with mounting groove, one side of the case is equipped with mounting bracket, the inside of mounting bracket is equipped with four heat dissipation fans, the top of four heat dissipation fans is equipped with fourteen fins, one side of mounting bracket is equipped with two connecting brackets.

[0006] Preferably, the machine cover is installed on the top surface of the case, and the machine cover is bolted to the case, four GPU main bodies are arranged in the inside of four cases, and the four GPU main bodies are arranged at equal intervals.

[0007] Preferably, the eight slots are equally spaced on the edge of the chassis, the eight pressure rods are integrally formed with the cover, and the positions of the pressure rods correspond one-to-one with the positions of the slots. All eight pressure rods are installed in the slots of the chassis.

[0008] Preferably, the four housings and the eight copper tubes inside the housings are all U-shaped, and the four housings and the eight copper tubes inside the housings are welded together. Both ends of the housings and the copper tubes are installed in the slots of the chassis. The four GPU bodies are installed inside the housings, and the copper tubes inside the housings are in contact with the outer surface of the GPU bodies.

[0009] Preferably, the four cooling fans are parallel to the housing, and the fins on the top of the four cooling fans are all fitted onto eight copper pipes inside the housing, and the fins are welded to the copper pipes.

[0010] Preferably, the two mounting slots are mirror images of the chassis, and the two connecting brackets on the surface of the mounting bracket correspond one-to-one with the mounting slots. Both connecting brackets of the mounting bracket are installed in the mounting slots of the chassis, and both connecting brackets are bolted to the chassis.

[0011] Preferably, all four cooling fans are mounted on the mounting bracket, and the cooling fans are plugged into the mounting bracket.

[0012] Beneficial effects

[0013] In this invention, a housing with copper pipes is installed on a slot. During installation, the housing is fitted over the GPU body, and the copper pipes inside the housing are made to fit against the outer surface of the GPU body. After fitting, the cover is installed on the top surface of the chassis with fixing bolts, and the housing is pressed tightly into the slot of the chassis by the pressure bar on the bottom of the cover. When the GPU body is under high load, the copper pipes attached to the surface of the GPU body will quickly transfer the heat of the GPU body to the end with fins. By turning on the cooling fan on the mounting bracket at the bottom of the fins, the cooling fan will quickly blow air over the cooling fins through forced convection, enhancing the heat transfer speed and improving the overall heat dissipation effect. This solves the problem that the cooling fan inside the chassis alone cannot effectively dissipate heat from the GPU under high load. Attached Figure Description

[0014] Figure 1 This is an isometric drawing of the present invention;

[0015] Figure 2 This is a front view of the present invention;

[0016] Figure 3 For the present utility model Figure 2 Sectional view at point AA;

[0017] Figure 4 This is a partial isometric drawing of the present invention;

[0018] Figure 5 This is a partial perspective view of the present invention;

[0019] Figure 6 This is an isometric drawing of a partial part of this utility model;

[0020] Figure 7 This is a top view of a partial part of the present invention;

[0021] Figure 8 For the present utility model Figure 7 Sectional view at BB.

[0022] Legend:

[0023] 1. Chassis; 2. Cover; 3. Mounting slot; 4. Connector; 5. Mounting bracket; 6. Cooling fan; 7. Card slot; 8. Pressure bar; 9. Housing; 10. Copper pipe; 11. Fins; 12. GPU body. Detailed Implementation

[0024] To make the technical means, creative features, and achieved objectives and effects of this utility model easier to understand, the present utility model is further described below with reference to specific embodiments and accompanying drawings. However, the following embodiments are merely preferred embodiments of this utility model and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments described in the implementation plan without creative effort are all within the protection scope of this utility model.

[0025] The specific embodiments of this utility model are described below with reference to the accompanying drawings. Specific Implementation Example 1:

[0027] Reference Figures 1-8An energy-saving GPU server cluster cooling server includes a chassis 1, a cover 2 on the top of the chassis 1, eight pressure bars 8 on the bottom surface of the cover 2, eight slots 7 on the top surface of the chassis 1, four GPU bodies 12 inside the chassis 1, and a shell 9 on the outside of each of the four GPU bodies 12. Each shell 9 has eight copper pipes 10 inside. Mounting slots 3 are provided on both sides of the chassis 1. A mounting bracket 5 is provided on one side of the chassis 1, with four cooling fans 6 inside the mounting bracket 5. Each of the four cooling fans 6 has fourteen fins 11 on its top. Two connecting brackets 4 are provided on one side of the mounting bracket 5. The cover 2 is mounted on the top surface of the chassis 1 and is bolted to the chassis 1. The four GPU bodies 12 are arranged in an equally spaced array inside the four chassis 1. The eight slots 7 are equally spaced along the edge of the chassis 1. The eight pressure bars 8 are integrally formed with the cover 2, and the positions of the pressure bars 8 and the slots 7 are aligned. The components are arranged in a one-to-one correspondence. All eight pressure rods 8 are installed in the slots 7 of the chassis 1. The four housings 9 and the eight copper pipes 10 inside the housings 9 are all U-shaped and welded together. Both ends of the housings 9 and the copper pipes 10 are installed in the slots 7 of the chassis 1. The four GPU bodies 12 are installed inside the housings 9, and the copper pipes 10 inside the housings 9 are in contact with the outer surface of the GPU bodies 12. The four cooling fans 6 are parallel to the housings 9. The fins 11 on the top of the four cooling fans 6 are all fitted onto the eight copper pipes 10 inside the housings 9, and the fins 11 are welded to the copper pipes 10. The two mounting slots 3 are mirrored with the chassis 1 as the center, and the two connecting brackets 4 on the surface of the mounting bracket 5 correspond one-to-one with the mounting slots 3. The two connecting brackets 4 of the mounting bracket 5 are installed in the mounting slots 3 of the chassis 1, and the two connecting brackets 4 are bolted to the chassis 1. The four cooling fans 6 are all installed on the mounting bracket 5, and the cooling fans 6 are plugged into the mounting bracket 5.

[0028] The server consists of components including a chassis 1, a cover 2, a GPU main body 12, a housing 9, copper pipes 10, a cooling fan 6, fins 11, mounting slots 3, mounting brackets 5, and connecting brackets 4. Its design aims to effectively address heat dissipation issues when the GPU main body 12 is under high load. The chassis 1, as the outer shell of the entire server, provides installation space and structural support for the internal components. The cover 2 is connected to the chassis 1 by bolts, and eight pressure bars 8 on its bottom surface engage with eight slots 7 on the top surface of the chassis 1. The main function of the pressure bars 8 is to secure the housing 9 within the slots 7 of the chassis 1, ensuring it does not move during operation. The GPU main body 12 is the core computing component of the server and is hot-swappable and mounted on the motherboard inside the chassis 1. The housing 9 is U-shaped and wraps around the GPU main body 12. The eight copper pipes 10 inside the housing 9 are attached to the outer surface of the GPU main body 12 to conduct heat generated by the GPU main body 12. The copper pipe 10 utilizes copper's excellent thermal conductivity to quickly transfer the heat generated by the GPU body 12 to the end with the fins 11. Four sets of fins 11 are positioned on top of the four cooling fans 6, and each set of fins 11 contains fourteen fins. The fins 11 are fitted onto the copper pipe 10 inside the housing 9 and welded together. The function of the fins 11 is to increase the heat dissipation area of ​​the copper pipe 10, allowing for more effective heat dissipation when air flows over them. The cooling fan 6 is installed inside the mounting bracket 5, promoting airflow through forced convection. When the cooling fan 6 is activated, it blows air over the fins 11. The mounting bracket 5 is connected to the chassis 1 via a connecting bracket 4. The mounting bracket 5 secures the cooling fan 6 and ensures it is parallel to the housing 9, allowing the cooling fan 6 to effectively blow air over the fins 11. The connecting bracket 4 is installed in the mounting slot 3 of the chassis 1 using fixing bolts, ensuring the stability of the mounting bracket 5 and preventing it from shaking when the cooling fan 6 is running. It should be noted that this application innovates on energy-saving GPU server clusters. Energy-saving GPU server clusters can flexibly allocate resources according to the dynamic changes of tasks to prevent resource waste. In addition, energy-saving GPU server clusters should also have components such as CPU, multiple memory modules, fan groups and power supply groups, but the above-mentioned components are existing public technologies and are not the main content protected by this application, so they do not need to be described in detail. Specific Implementation Example 2:

[0030] Reference Figures 1-8An energy-saving GPU server cluster cooling server, further based on the basic structure in Specific Embodiment 1, operates as follows: When the GPU main body 12 is under high load, it generates a large amount of heat. This heat is conducted through the copper pipe 10, which is attached to the outer surface of the GPU main body 12, to the end of the copper pipe 10 with fins 11, and then transferred to the fins 11. Subsequently, the cooling fan 6 on the mounting bracket 5 is turned on. The cooling fan 6 uses forced convection to quickly blow air over the fins 11 fitted onto the copper pipe 10. Because the fins 11 have a large heat dissipation area, the airflow can quickly carry away the heat from the fins 11, achieving efficient heat dissipation for the GPU main body 12 through the cooperation of the copper pipe 10 and the fins 11. It should be noted that the copper pipe 10 is hollow and contains a small amount of water or other chemical substances. When the GPU main body 12 exceeds a critical temperature, the water vapor in the cooling copper pipe 10 will carry away the heat from the GPU main body 12 along the capillary structure; after the water vapor cools and liquefies, it begins to circulate back, thus cyclically carrying away the heat from the GPU main body 12.

[0031] In summary:

[0032] 1. The housing 9 with copper pipes 10 is installed on the slot 7. During installation, the housing 9 covers the GPU body 12, and the copper pipes 10 inside the housing 9 are in contact with the outer surface of the GPU body 12. After the contact is made, the cover 2 is installed on the top surface of the chassis 1 with fixing bolts, and the housing 9 is pressed tightly into the slot 7 of the chassis 1 by the pressure bar 8 on the bottom surface of the cover 2. When the GPU body 12 is used under high load, the copper pipes 10 in contact with the surface of the GPU body 12 will quickly transfer the heat of the GPU body 12 to the end with fins 11. By turning on the cooling fan 6 on the mounting bracket 5 at the bottom of the fins 11, the cooling fan 6 blows the air quickly across the cooling fins 11 through forced convection, which enhances the heat transfer speed and improves the overall heat dissipation effect. This solves the problem that the cooling fan 6 inside the chassis 1 cannot effectively dissipate heat from the GPU under high load.

[0033] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0034] The foregoing has shown and described the basic principles, main features, and advantages of this utility model. Those skilled in the art should understand that this utility model is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the utility model. Various changes and modifications can be made to this utility model without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed utility model. The scope of protection of this utility model is defined by the appended claims and their equivalents.

Claims

1. An energy-saving GPU server cluster heat dissipation server, comprising a chassis (1), characterized in that: The top of the chassis (1) is provided with a cover (2), the bottom surface of the cover (2) is provided with eight pressure bars (8), the top surface of the chassis (1) is provided with eight slots (7), the inside of the chassis (1) is provided with four GPU bodies (12), the outside of the four GPU bodies (12) is provided with a shell (9), the inside of the shell (9) is provided with eight copper pipes (10), the surfaces on both sides of the chassis (1) are provided with mounting slots (3), one side of the chassis (1) is provided with a mounting bracket (5), the inside of the mounting bracket (5) is provided with four cooling fans (6), the top of the four cooling fans (6) is provided with fourteen fins (11), and one side of the mounting bracket (5) is provided with two connecting brackets (4).

2. The energy-saving GPU server cluster cooling server according to claim 1, characterized in that: The cover (2) is installed on the top surface of the chassis (1) and the cover (2) is bolted to the chassis (1). The four chassis (1) are equipped with four GPU bodies (12) and the four GPU bodies (12) are arranged in an array at equal intervals.

3. The energy-saving GPU server cluster cooling server according to claim 1, characterized in that: The eight slots (7) are equally spaced on the edge of the chassis (1), and the eight pressure rods (8) are integrally formed with the cover (2). The positions of the pressure rods (8) correspond one-to-one with the positions of the slots (7). All eight pressure rods (8) are installed in the slots (7) of the chassis (1).

4. The energy-saving GPU server cluster cooling server according to claim 1, characterized in that: The four housings (9) and the eight copper tubes (10) inside the housings (9) are all U-shaped, and the four housings (9) and the eight copper tubes (10) inside the housings (9) are welded together. The two ends of the housings (9) and the copper tubes (10) are installed in the slots (7) of the chassis (1). The four GPU bodies (12) are installed inside the housings (9), and the copper tubes (10) inside the housings (9) are attached to the outer surface of the GPU bodies (12).

5. The energy-saving GPU server cluster cooling server according to claim 1, characterized in that: The four cooling fans (6) are parallel to the housing (9). The fins (11) on the top of the four cooling fans (6) are all fitted onto the eight copper pipes (10) inside the housing (9), and the fins (11) are welded to the copper pipes (10).

6. The energy-saving GPU server cluster cooling server according to claim 1, characterized in that: The two mounting slots (3) are mirror images of the chassis (1), and the two connecting brackets (4) on the surface of the mounting bracket (5) correspond one-to-one with the mounting slots (3). The two connecting brackets (4) of the mounting bracket (5) are installed in the mounting slots (3) of the chassis (1), and the two connecting brackets (4) are bolted to the chassis (1).

7. The energy-saving GPU server cluster cooling server according to claim 1, characterized in that: All four cooling fans (6) are mounted on the mounting bracket (5), and the cooling fans (6) are plugged into the mounting bracket (5).

Images