Supercomputing device and data center

Abstract

The utility model provides supercomputing equipment and a data center. The supercomputing equipment comprises a shell and a supercomputing unit, the shell is provided with an air inlet and an air outlet which are oppositely arranged; the supercomputing unit is installed in the shell and located between the air inlet and the air outlet. The supercomputing unit comprises a computing power board, a first radiator and a sealing element, the computing power board comprises a substrate and a plurality of computing power chips, a first surface of the substrate is provided with a chip area and a bare board area along a first direction, the bare board area is closer to the air inlet than the chip area, and the plurality of computing power chips are arranged in the chip area; the first radiator is mounted on the first surface of the substrate, and the first radiator is provided with an extension structure opposite to the bare board area; the edge area, close to the air inlet, of the extension structure and the bare board area is provided with the sealing piece. Wherein the first direction is the opposite direction of the air inlet and the air outlet. According to the supercomputing device, dust and other foreign matters can be prevented from being deposited on the computing force plate and the computing force chip, so that the reliability and the computing efficiency of the whole device are improved.

Description

Technical Field

[0001] The utility model relates to the technical field of servers, in particular to a supercomputing device and a data center. Background Art

[0002] With the development of science and technology, the requirements for the computing power of servers are getting higher and higher, especially for supercomputing equipment used for virtual currency. Its computing power chips need to perform huge calculations. In order to ensure its normal operation, the heat dissipation requirements for the computing power chips are extremely high. In the existing technology, a heat sink is often attached to the computing board where the computing power chip is located, and the computing board and the heat sink are located in the entire air flow channel to take away the heat from the radiator as quickly as possible.

[0003] However, with the continuous improvement of heat dissipation requirements, while increasing the airflow in the airflow channel, more dust and other foreign matter will be brought into the airflow channel. These dusts are very easy to deposit on the computing board, especially on the computing chip. With the accumulation of dust and other foreign matter, it may cause faults such as short circuit of the computing board. The coverage of dust and other foreign matter also affects the heat dissipation effect of the computing chip, resulting in problems such as reduced computing efficiency. Utility Model Content

[0004] Based on the above situation, the main purpose of the utility model is to provide a supercomputing device and a data center to prevent dust and other foreign matter from being deposited on the computing board and computing chip as much as possible, thereby improving the reliability and computing efficiency of the entire device.

[0005] In order to achieve the above purpose, the technical solution adopted by the utility model is as follows:

[0006] A first aspect of the utility model provides a supercomputing device, comprising a housing and a supercomputing unit;

[0007] The housing has an air inlet and an air outlet arranged opposite to each other;

[0008] The supercomputing unit is installed in the housing and located between the air inlet and the air outlet;

[0009] The supercomputing unit includes a computing board, a first heat sink and a seal. The computing board includes a substrate and a plurality of computing chips. A chip area and a bare board area are provided on the first surface of the substrate along a first direction. The bare board area is closer to the air inlet than the chip area, and the plurality of computing chips are provided in the chip area. The first heat sink is installed on the first surface of the substrate, and the first heat sink has an extension structure opposite to the bare board area. The seal is provided on the edge area of ​​the extension structure and the bare board area close to the air inlet. The first direction is the relative direction of the air inlet and the air outlet.

[0010] Preferably, the sealing member is located between the bare board area and the extension structure.

[0011] Preferably, the sealing member is a strip-shaped plate structure, and two opposite surfaces thereof are respectively in contact with the first heat sink and the substrate.

[0012] Preferably, the extension structure is flush with the bare panel area on a side close to the air inlet; and the sealing member is attached to the extension structure and the side wall of the bare panel area close to the air inlet.

[0013] Preferably, the seal is a strip structure, including a first section, a second section and a third section, the first section is located at an edge area between the extension structure and the bare board area close to the air inlet, and extends along the edge of the edge area; the second section and the third section are respectively bent and connected to both ends of the first section, and are respectively located at two edge areas of the first heat sink and the bare board area that are opposite in a second direction; wherein the second direction is perpendicular to the first direction.

[0014] Preferably, when projected in the first direction, the projection of the chip area is located within the projection range of the sealing member.

[0015] Preferably, the sealing member is a silicone member or a sealing foam.

[0016] Preferably, the sealing member is a thermally conductive silicone member.

[0017] Preferably, the extension structure is arranged to be inclined toward the windward surface of the air inlet, and the end of the windward surface away from the air inlet is inclined toward the direction away from the computing board more than the end closer to the air inlet.

[0018] Preferably, a plurality of the first heat sinks are arranged on the same side of the same substrate along the second direction, and the plurality of the first heat sinks cover a plurality of the computing power chips in the second direction; wherein the second direction is perpendicular to the first direction; and the sealing member is arranged on the edge of each of the first heat sinks close to the air inlet;

[0019] A second heat sink is disposed on a second surface of the substrate facing away from the computing chip. The second heat sink is attached to and covers the substrate. The supercomputing unit is connected to the shell through the second heat sink.

[0020] A second aspect of the present invention provides a data center, comprising any of the supercomputing devices described above.

[0021] The supercomputing device of the utility model, by adding a seal between the computing board and the first radiator near the air inlet of the supercomputing unit, blocks the gap between the computing board and the first radiator at the air inlet end near the air inlet side. Therefore, even if the airflow in the airflow channel from the air inlet to the air outlet is large and there are relatively many foreign objects such as dust, these foreign objects such as dust are blocked by the seal when approaching the entire supercomputing unit, so that basically no foreign objects such as dust will enter between the computing board and the first radiator, greatly reducing the deposition of foreign objects such as dust on the computing board and even the computing chip, thereby reducing the probability of short circuit of the computing board due to foreign objects such as dust, and ensuring the heat dissipation effect of the computing chip. In addition, the heat dissipation area of ​​the first radiator is increased by setting an extension structure, thereby avoiding the influence of the heat dissipation effect of the entire computing board due to the airflow that basically does not directly flow through the computing chip.

[0022] Other beneficial effects of the utility model will be explained through the introduction of specific technical features and technical solutions in the specific implementation manner. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

[0024] Figure 1 A cross-sectional schematic diagram of a preferred implementation of the supercomputing device provided by the utility model;

[0025] Figure 2 A schematic structural diagram of a preferred implementation of a supercomputing unit in the supercomputing device provided by the utility model;

[0026] Figure 3 for Figure 2 A partial enlarged view at point I in the middle;

[0027] Figure 4 A schematic diagram of the relative positions of a power board and a sealing member in a preferred embodiment of the supercomputing device provided by the utility model;

[0028] Figure 5 A schematic diagram of the relative positions of another preferred embodiment of a computing board and a sealing member in the supercomputing device provided by the utility model;

[0029] Figure 6 The present invention is a schematic structural diagram of another preferred implementation of a supercomputing unit in the supercomputing device provided by the present invention.

[0030] In the figure:

[0031] 10. Shell; 11. Air inlet; 12. Air outlet;

[0032] 20. Supercomputing unit; 21. Calculation board; 211. Base plate; 2111. Bare board area; 2112. Chip area; 212. Calculation chip; 22. First radiator; 221. Heat sink; 222. Fin; 223. Extension structure; 2231. Windward surface; 23. Seal; 231. First section; 232. Second section; 233. Third section; 24. Second radiator. DETAILED DESCRIPTION

[0033] The present invention is described below based on embodiments, but the present invention is not limited to these embodiments. In the following detailed description of the present invention, some specific details are described in detail. In order to avoid confusing the essence of the present invention, known methods, processes, procedures, and components are not described in detail.

[0034] In addition, persons of ordinary skill in the art will appreciate that the drawings provided herein are for illustration purposes and are not necessarily drawn to scale.

[0035] Unless the context clearly requires otherwise, throughout the specification and claims, the words "include", "comprising" and similar words should be interpreted in an inclusive sense rather than an exclusive or exhaustive sense; that is, in the sense of "including but not limited to".

[0036] In the description of the present invention, it should be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "plurality" is two or more.

[0037] The utility model provides a supercomputing device for processing data, information, etc., such as processing data and information in virtual currency. Figure 1-Figure 6The supercomputing device includes a shell 10 and a supercomputing unit 20. The shell 10 has an air inlet 11 and an air outlet 12 that are relatively arranged. Specifically, the shell 10 can be a rectangular structure or a quasi-rectangular structure. The air inlet 11 is arranged on a first end surface of the shell 10, and the air outlet 12 can be arranged on a second end surface of the shell 10, so that the air inlet 11, the inner cavity of the shell 10, and the air outlet 12 form an airflow channel, and the airflow is blown out from the air inlet 11 through the inner cavity of the shell 10 to the air outlet 12. The supercomputing unit 20 is installed in the housing 10 and is located between the air inlet 11 and the air outlet 12. The supercomputing unit 20 includes a computing board 21, a first radiator 22 and a seal 23. The computing board 21 includes a substrate 211 and a plurality of computing chips 212. A first surface of the substrate 211 is provided with a chip area 2112 and a bare board area 2111 along a first direction. The bare board area 2111 is closer to the air inlet 11 than the chip area 2112. The plurality of computing chips 212 are arranged in the chip area 2112. That is, the substrate 211 has two opposite surfaces, namely a first surface and a second surface. The plurality of computing chips 212 are arranged on the same surface of the substrate 211. That is, on the first surface; the first heat sink 22 is installed on the first surface of the substrate 211, that is, installed on the side of the substrate 211 where the computing chip 212 is set. The first heat sink 22 has an extension structure 223 opposite to the bare board area 2111, that is, the extension structure 223 at least partially extends out of the chip area 2112. It can be understood that there is a gap between the substrate 211 and the first heat sink 22. The extension structure 223 and the bare board area 2111 are provided with a seal 23 at the edge area close to the air inlet 11. That is, the seal 23 is provided on the side of the supercomputing unit 20 close to the air inlet 11 to seal the air inlet end of the gap close to the air inlet 11. Among them, the first direction X is the relative direction of the air inlet 11 and the air outlet 12.

[0038] In the supercomputing device of the present invention, a sealing member 23 is added between the computing board 21 and the first radiator 22 near the air inlet 11 of the supercomputing unit 20, so that the gap between the computing board 21 and the first radiator 22 is blocked at the air inlet end near the air inlet 11, so that each computing chip 212 is located on the side of the sealing member 23 away from the air inlet 11. Therefore, even if the airflow on the airflow channel from the air inlet 11 to the air outlet 12 is large and there are relatively many foreign objects such as dust, these foreign objects are also close to the entire The supercomputing unit 20 is blocked by the seal 23, so that basically no foreign matter such as dust will enter between the computing board 21 and the first radiator 22, which greatly reduces the deposition of foreign matter such as dust on the computing board 21 and even the computing chip 212, thereby reducing the probability of short circuit of the computing board 21 due to foreign matter such as dust, and alleviates the corrosion of the computing board 212 caused by the deposition of dust, etc., and because the coverage of foreign matter such as dust is reduced, the heat dissipation effect of the computing chip 212 and even the entire computing board 21 is guaranteed. And by setting the extension structure, the heat dissipation area of ​​the first radiator is increased, thereby avoiding the influence of the heat dissipation effect of the entire computing board due to the airflow on the substrate 211 that does not directly flow through the computing chip.

[0039] Among them, the shell 10 can be provided with an air inlet fan at the air inlet 11, or an air outlet fan at the air outlet 12, or an air inlet fan and an air outlet fan can be provided at the same time. Grilles can also be provided at the air inlet 11 and the air outlet 12 to reduce the entry of foreign matter such as dust into the air flow channel, and further reduce the probability of foreign matter such as dust being deposited on the computing power chip 212.

[0040] Continue to refer to Figure 2 , Figure 4 and Figure 5, multiple computing chips 212 can be arranged along the second direction Y on the substrate 211 to form multiple chipsets, each chipset includes multiple computing chips 212 arranged along the first direction X. The second direction Y and the first direction X are substantially parallel to the substrate 211 and are two directions perpendicular to each other. Specifically, the second direction Y, i.e., the arrangement direction hereinafter, can be consistent with the height direction of the housing 10, and the first direction X is consistent with the depth direction of the housing 10, and is also the relative direction of the air inlet 11 and the air outlet 12. Specifically, the substrate 211 has a chip area 2112 and a bare board area 2111, the bare board area 2111 is located at the edge area of ​​the substrate 211 close to the air inlet 11, and the chip area 2112 is located on the side of the bare board area 2111 away from the air inlet 11, that is, the bare board area 2111 and the chip area 2112 are arranged along the first direction X, and the bare board area 2111 is closer to the air inlet 11 than the chip area 2112. In this embodiment, a plurality of computing chips 212 are arranged in the chip area 2112, and no computing chip 212 is arranged on the bare board area 2111. It can be understood that other components are also arranged on the substrate 211, such as other chips, wiring, interface terminals, etc. These components may be arranged in the chip area 2112 or may not be arranged in the chip area 2112.

[0041] The first heat sink 22 includes a heat sink 221 and fins 222. The first surface of the heat sink 221 faces the computing board 21, and can be specifically attached to the computing chip 212 to increase the efficiency of heat conduction. The second surface of the heat sink 221 is provided with a plurality of fins 222, each fin 222 extends along the first direction X, that is, each fin 222 extends along the relative direction of the air inlet 11 and the air outlet 12, and the plurality of fins 222 are arranged at intervals along the second direction Y, that is, there is space between adjacent fins 222, so that the airflow can better take away the heat thereon. In order to increase the heat dissipation effect of the first heat sink 22, the thickness of the first heat sink 22 can be set thicker, and specifically, the size of the fins 222 extending out of the heat sink 221 can be set larger.

[0042] The sealing member 23 may be a strip-shaped structure, extending along the arrangement direction, that is, the second direction Y, to seal the air inlet end of the gap between the entire first heat sink 23 and the substrate 211. Specifically, the sealing member 23 may only extend along the second direction Y. In a preferred embodiment, Figure 5As shown, the seal 23 includes a first section 231, a second section 232 and a third section 233. The first section 231 is located at the edge area of ​​the extension structure 223 and the bare board area 2111 close to the air inlet 11, and extends along the edge of the edge area, that is, the first section 231 basically extends along the second direction Y; the second section 232 and the third section 233 are respectively bent and connected to the two ends of the first section 231, and are respectively located at two edge areas of the first heat sink 22 and the bare board area 2111 opposite to each other in the second direction Y, that is, the second section 232 and the third section 233 basically extend along the first direction X, and can be arranged only in the bare board area 2111, or can continue to extend to the side where the chip area 2112 is located, such as basically surrounding the chip area 2112 in the entire seal 23. By providing the seal 23 with this structure, it is possible to further reduce the deposition of foreign matter such as dust near the computing chip 212.

[0043] Further preferably, the seal 23 may be a strip plate structure, and two opposite surfaces of the seal 23 are in contact with the first heat sink 22 and the substrate 211 to improve the sealing between the first heat sink 22 and the substrate 211 on the air inlet 11 side.

[0044] In one embodiment, the seal 23 may be a component with thermal conductivity, such as a thermally conductive rubber strip, to increase the heat conduction between the computing board 21 and the first heat sink 22 while sealing. In another embodiment, the seal 23 is a component with certain compression performance, which may be a silicone member or a sealing foam, to increase the fit between the seal 23 and the substrate 211 and the first heat sink 22, especially when the contact parts of the substrate 211, the first heat sink 22 and the seal 23 are not flat enough, the sealing between the substrate 211 and the first heat sink 22 can be greatly improved, especially in the implementation where the seal 23 is arranged between the bare board area 2111 and the extension structure (described in detail below), such a seal 23 can better increase the sealing of the air inlet end through the squeezing force of the substrate 211 and the first heat sink 22 during assembly, and can also improve the reliability of the installation of the seal 23. In a preferred embodiment, the seal 23 is a thermally conductive silicone member, so as to simultaneously improve the sealing of the air inlet end and the heat dissipation efficiency of the computing board 21.

[0045] In one embodiment, on the side where the air inlet 11 is located, the first radiator 22 and the side wall of the computing board 21 are offset, that is, not flush, such as the first radiator 22 can be extended out of the computing board 21, specifically, the heat sink 221 can extend out of the substrate 211; for example, on the side close to the air inlet 11, the substrate 211 can also extend out of the first radiator 22. In another embodiment, on the side where the air inlet 11 is located, the first radiator 22 and the side wall of the computing board 21 are aligned, such as Figure 2 , Figure 3As shown, on the side close to the air inlet 11, the substrate 211 and the first heat sink 22 can be aligned, that is, the extension structure 223 is flush with the side wall of the bare board area 2111 close to the air inlet to reduce the air resistance of the airflow on the air inlet side of the supercomputing unit 20.

[0046] Regardless of which of the above methods is adopted by the first heat sink 22 and the computing board 21 on the air inlet side, the seal 23 can be directly arranged between the first heat sink 22 and the computing board 21 to achieve sealing at the air inlet end, or can be arranged on the extension structure 23 and the side wall of the bare board area 2111 to achieve sealing. In the embodiment where the first heat sink 22 and the computing board 21 are not flush on the side where the air inlet 11 is located, the seal 23 can be arranged on the side wall of the first heat sink 22 and the substrate 211 close to the air inlet 11, and sealed with the bottom surface of the other. If the first heat sink 22 extends out of the computing board 21, the seal can be attached to the side wall of the substrate 211 close to the air inlet, and attached to the bottom surface of the first heat sink 22, specifically sealed with the first surface of the heat sink 221; or the seal can be directly arranged between the extension structure 223 and the bare board area 2111.

[0047] In one embodiment, the extension structure 23 is flush with the bare board area 2111 on the side close to the air inlet 11, that is, the first heat sink 22 and the computing board 21 are aligned with the side walls on the side close to the air inlet 11; the seal 23 is arranged on the side walls of the first heat sink 22 and the substrate 211, that is, it is located at the side wall of the substrate 211 close to the air inlet 11 and the side wall of the first heat sink 22 close to the air inlet 11.

[0048] In a preferred embodiment, the sealing member 23 is located between the first heat sink 22 and the substrate 211. Specifically, Figure 2 , Figure 3 As shown, the seal 23 is located between the bare board area 2111 and the extension structure 223. In this way, all the computing chips 212 are located on the side of the seal 23 away from the air inlet 11. At the same time, the seal 23 is pressed against the first heat sink 22 through the computing board 21, thereby improving the sealing between the substrate 211 and the first heat sink 22. When the seal 23 is a strip plate structure, the two opposite surfaces of the plate structure are respectively attached to the extension structure 223 and the bare board area 2111, that is, the extension structure 223, the seal 23 and the bare board area 2111 form a sandwich structure. In particular, when the seal 23 is a strip plate structure, the two opposite surfaces of the seal 23 are respectively attached to the extension structure 223 and the bare board area 2111.

[0049] The extension structure 223 is tilted toward the windward surface 2231 of the air inlet 11, that is, the end of the extension structure 223 away from the air inlet 11 is tilted toward the direction away from the computing board 21 relative to the end close to the air inlet 11. Figure 2 , Figure 3By setting the extension structure 223 as an inclined structure, the airflow can be guided, and more importantly, dust and other foreign matter in the airflow can be prevented from being deposited on the end wall of the first heat sink 22 to affect the flow of the entire airflow, thereby further improving the performance of the entire supercomputing unit 20.

[0050] In order to better prevent dust and other foreign matter from being deposited on the computing chip 212 and other components on the substrate 211, in the relative direction between the air inlet 11 and the air outlet 12, i.e., the aforementioned first direction X, the projection of the chip area 2112 is located within the projection range of the seal 23, i.e., both ends of the seal 23 along the second direction Y exceed the chip area 2112. In another embodiment, the projection contour of the chip area 2112 coincides with the projection contour of the seal 23.

[0051] Among them, only one first heat sink 22 can be set on the same side of the same substrate 211, and the first heat sink 22 covers each computing power chip 212, that is, the first heat sink 22 is simultaneously attached to all computing power chips 212. In a preferred embodiment, multiple first heat sinks 22 are set on the same side of the same substrate 211 along the arrangement direction, that is, the above-mentioned second direction Y, and the multiple first heat sinks 22 cover multiple computing power chips 212 in the arrangement direction. For example, two, three or more first heat sinks 22 are set, wherein, Figure 2 The structure in which two first radiators 22 are provided is shown. Among these first radiators 22, each first radiator 22 covers part of the computing chip 212, so that the contact area between each first radiator 22 and the computing chip 212 is required to be small, thereby reducing the difficulty of processing the first radiator 22, and enabling the first radiator 22 to better fit the computing chip 212, thereby improving the heat dissipation efficiency of the computing chip 212. Among them, the above arrangement direction is parallel to the substrate 211 and perpendicular to the relative direction of the air inlet 11 and the air outlet 12. Preferably, two first radiators 22 are arranged on the side of the same substrate 211 where the computing chip 212 is arranged. The two first radiators 22 can not only reduce the difficulty of processing the first radiator 22, ensure the fitting area between the first radiator 22 and the computing chip 212, but also reduce the assembly time of the first radiator 22 and the computing board 21.

[0052] In order to further improve the heat dissipation effect of the entire computing board 21, a second heat sink 24 may be provided on the side of the substrate 211 away from the computing chip 212. Figure 6As shown, the second heat sink 24 is attached to and covers the substrate 211. The second heat sink 24 can adopt the structure of the first heat sink 22, such as also including a heat sink and fins. Considering that the substrate 211 is basically flat on the side away from the computing chip 212, in a preferred embodiment, a second heat sink 24 is provided on the side of the substrate 211 away from the computing chip 212, and the second heat sink 24 is attached to the substrate 211, that is, no seal 23 may be provided between the two.

[0053] Among them, the supercomputing unit 20 and the shell 10 can be installed through the first radiator 22. In an embodiment in which a second radiator 24 is provided, the supercomputing unit 20 can be installed on the shell 10 through the second radiator 24. For example, a slot can be provided on the inner wall of the shell 10, and both ends of the second radiator 24 in the first direction are arranged to extend beyond the computing board 21 and the first radiator 22, and the extending part is connected to the slot on the shell 10 by plugging. In this way, the structure and setting method of the first radiator 22 will not be affected by the need to be installed with the shell 10, that is, through the second radiator 24, the heat dissipation effect of the computing board 21 can be increased, and the installation of the entire supercomputing unit 20 can be facilitated.

[0054] The utility model also provides a data center, including the supercomputing device described in any of the above embodiments, which may include only one supercomputing device or may include two or more supercomputing devices.

[0055] Those skilled in the art will appreciate that, without conflict, the above-mentioned preferred solutions can be freely combined and superimposed.

[0056] It should be understood that the above-mentioned embodiments are merely illustrative and not restrictive. Without departing from the basic principles of the present invention, various obvious or equivalent modifications or substitutions that can be made by technicians in this field to the above details will be included in the scope of the claims of the present invention.

Claims

1. A supercomputing device, characterized in that: including a shell and a supercomputing unit; The housing has an air inlet and an air outlet arranged opposite to each other; The supercomputing unit is installed in the housing and located between the air inlet and the air outlet; The supercomputing unit includes a computing board, a first heat sink and a seal. The computing board includes a substrate and a plurality of computing chips. A chip area and a bare board area are provided on the first surface of the substrate along a first direction. The bare board area is closer to the air inlet than the chip area, and the plurality of computing chips are provided in the chip area. The first heat sink is installed on the first surface of the substrate, and the first heat sink has an extension structure opposite to the bare board area. The seal is provided on the edge area of ​​the extension structure and the bare board area close to the air inlet. The first direction is the relative direction of the air inlet and the air outlet.

2. The supercomputer device according to claim 1, characterized in that: The seal is located between the bare board area and the extension structure.

3. The supercomputer device according to claim 2, characterized in that: The sealing member is a strip plate structure, and two opposite surfaces thereof are respectively in contact with the first heat sink and the substrate.

4. The supercomputer device according to claim 1, characterized in that: The extension structure is flush with the bare panel area on a side close to the air inlet; the sealing member is attached to the extension structure and the side wall of the bare panel area close to the air inlet.

5. The supercomputer device according to claim 1, characterized in that: The sealing member is a strip structure, comprising a first section, a second section and a third section, wherein the first section is located at an edge region between the extension structure and the bare panel area close to the air inlet, and extends along an edge of the edge region; the second section and the third section are respectively bent and connected to two ends of the first section, and are respectively located at two edge regions of the first heat sink and the bare panel area that are opposite in a second direction; wherein the second direction is perpendicular to the first direction.

6. The supercomputer device according to claim 1, characterized in that: Projected in the first direction, the projection of the chip area is located within the projection range of the sealing member.

7. The supercomputer device according to claim 1, characterized in that: The sealing member is a silicone member or a sealing foam.

8. The supercomputer device according to claim 1, characterized in that: The sealing member is a heat-conducting silicone member.

9. The supercomputer device according to claim 1, characterized in that: The extension structure is arranged to be inclined toward the windward surface of the air inlet, and an end of the windward surface away from the air inlet is inclined toward a direction away from the computing board relative to an end closer to the air inlet.

10. The supercomputer device according to any one of claims 1 to 9, characterized in that: A plurality of the first heat sinks are arranged along the second direction on the same side of the same substrate, and the plurality of the first heat sinks cover a plurality of the computing power chips in the second direction; wherein the second direction is perpendicular to the first direction; and the sealing member is arranged on the edge of each of the first heat sinks close to the air inlet; A second heat sink is disposed on a second surface of the substrate facing away from the computing chip. The second heat sink is attached to and covers the substrate. The supercomputing unit is connected to the shell through the second heat sink.

11. A data center, characterized in that: A supercomputing device comprising any one of claims 1-10.

Images