Computing node and computing device
By setting up a flow diversion structure and flow hole on the main board of the computing node, longitudinal power supply is achieved, which solves the problems of low flow density and poor energy efficiency in the prior art, improves the power supply density and computing power density, and reduces the heat generation. Enhanced power supply stability and safety.
Patent Information
- Application Number
- CN202411020418.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2022-08-30
- Publication Date
- 2024-11-12
AI Technical Summary
In the existing server power supply technology, the current density is low and the energy efficiency is poor, making it difficult to meet the power supply needs of high-power computing nodes, resulting in large heat generation of the motherboard, affecting the normal operation and safety of electronic devices.
By setting up a flow guide structure and flow hole on the main board of the computing node, longitudinal power supply is achieved using conductive dielectric and flow guide structure, the power supply density is improved, and the circuit supply path is optimized through the high-conducting and current-carrying layer hierarchical structure to reduce the heat generation and motherboard thickness.
It improves the power supply density and computing power density of the computing nodes, reduces the heat generation of the motherboard, enhances the stability and safety of power supply, and supports the power supply demand of high-density electronic devices.
Smart Images

Figure CN118945995A_ABST
Abstract
Description
[0001] This application is a divisional application of the invention patent application with application number 202211049239.9 submitted to the China Patent Office, application date August 30, 2022, and invention name “Computing Node and Computing Device”. Technical Field
[0002] The present invention relates to the field of computing devices, and in particular to a computing node and a computing device. Background Art
[0003] With the rise of big data, cloud computing, and AI (Artificial intelligence), the number of processing cores, number of particles, and capacity of processors, memory sticks, and hard disks have continued to increase, and system power consumption has also increased significantly (the power consumption of a single processor has evolved from 200w to 1000w). As a result, the power supply capacity of data centers and servers has become increasingly demanding; however, due to the limitations of the current traditional horizontal power supply architecture, the current density is low and the energy efficiency is poor, which makes it difficult to meet the needs of the long-term evolution and development of server power. Summary of the invention
[0004] The embodiments of the present application are to improve the power supply density of the mainboard in the computing node.
[0005] The first aspect of the embodiment of the present application proposes a computing node. The computing node includes a mainboard and a flow guide structure, an electronic device is mounted on the first surface of the mainboard; a flow hole is opened along the thickness direction of the mainboard, and the hole wall of the flow hole has a conductive medium; the flow hole is electrically connected to the electronic device; the flow guide structure is connected to the power supply, and is used to transmit the power supply current output from the power supply; the flow guide structure is arranged on one side of the mainboard, and is electrically connected to the hole wall of the flow hole, so as to transmit the power supply current to the flow hole along the thickness direction of the mainboard.
[0006] Here, there are one or more electronic devices on the mainboard, which may include one or more processors. The electronic devices may also include memory. The flow-guiding structure and the hole wall of the through hole may be fixedly connected by a conductive fixing member.
[0007] In the first aspect, in the process of supplying power to the electronic devices on the mainboard, the power supply current is first conducted outside the mainboard through the flow guide structure, and enters the mainboard through the electrical connection between the flow guide structure and the flow hole of the mainboard. Inside the mainboard, the power supply current is transmitted along the flow hole in the thickness direction of the mainboard, thereby reaching the power supply end of the electronic device, and realizing power supply for the electronic device. In the embodiment of the present application, since the flow guide structure and the mainboard are independent structures of each other, it is possible to select a flow guide structure with a suitable flow capacity according to the power required by the mainboard to be powered, without being limited by the flow capacity of the mainboard, thereby improving the power supply density of the computing node. And it is conducive to promoting the development of computing nodes towards high-density computing power and high-density power. In addition, since there is no need for large current to flow horizontally in the mainboard, the heat generated by the mainboard is greatly reduced, and the safety of the computing node operation is improved.
[0008] In a possible implementation of the first aspect, the mainboard includes a current-carrying layer extending along the transverse direction of the mainboard; the current-carrying layer is electrically connected to the hole wall of the flow hole; the flow hole is electrically connected to the electronic device, including: the flow hole is electrically connected to at least one electronic device through the current-carrying layer.
[0009] In this implementation, the current-carrying layer and the hole wall of the through hole are in contact to achieve electrical connection. The current-carrying layer and the electronic device can be electrically connected through the through hole. Therefore, the power supply current is transmitted to the current-carrying layer along the through hole, and on the current-carrying layer, the power supply current is transmitted along the cross-sectional direction of the mainboard, and then transmitted to the power supply end of the electronic device through the through hole along the thickness direction of the mainboard.
[0010] In a possible implementation of the first aspect, the plurality of through-holes include a first through-hole, which penetrates the mainboard along a thickness direction of the mainboard; a hole wall of the first through-hole contacts the current-carrying layer for achieving electrical connection.
[0011] The first through-hole and the flow-guiding structure can be fixedly connected by a conductive fixing member, thereby ensuring the structural stability of the flow-guiding structure and the main board. In addition, since the conductive fixing member is inserted into the first through-hole, the flow rate of the current in the first through-hole is increased, thereby improving the longitudinal flow capacity of the main board.
[0012] In a possible implementation of the first aspect, the electronic device includes a processor module; there are multiple through-holes, including a second through-hole among the multiple through-holes; the mainboard also includes a high-conductivity portion; the second through-hole extends along the thickness direction of the mainboard, and the first orifice of the second through-hole is connected to the flow-guiding structure; the second orifice of the second through-hole is connected to the high-conductivity portion; the high-conductivity portion extends from the second orifice along the thickness direction of the mainboard to the first surface of the mainboard, and contacts the power supply end of the processor module. Here, the first orifice and the flow-guiding structure can be connected by a conductive fixing member.
[0013] The second through-hole is electrically conductive by the conductive fixing member and the hole wall of the second through-hole, thereby ensuring a large longitudinal through-current capacity. Furthermore, by arranging a high-conductivity portion connected between the second through-hole and the processor module, the current is directly and stably transmitted to the surface of the mainboard, thereby being connected to the processor module, without relying on other layer structures in the mainboard for current carrying, thereby improving the reliability and stability of the power supply to the processor.
[0014] In a possible implementation manner of the first aspect, a cross-sectional area of the high conductivity portion is greater than or equal to a cross-sectional area of the second through-flow hole.
[0015] In this possible implementation, the longitudinal current carrying capacity of the mainboard can be improved to meet the processor's demand for high current power supply.
[0016] In a possible implementation manner of the first aspect, the cross-sectional area of the high-conductivity portion is gradually widened from the second opening of the second through-hole toward the first surface of the main board.
[0017] In this possible implementation, by setting the shape of the high-conductivity portion, the internal resistance of the high-conductivity portion can be reduced, the heat generation can be reduced, and the shape and area of the high-conductivity portion exposed on the first surface of the mainboard can be adjusted to adapt to the power supply end of the switching device.
[0018] In a possible implementation manner of the first aspect, the second opening of the second through hole and the high-conductivity portion are connected at a first position of the current-carrying layer; and the first position of the current-carrying layer is also electrically connected to at least one through hole.
[0019] In this possible implementation, since the high-conductivity portion is electrically connected to the second through-hole and the first through-hole at the same time, the current in the second through-hole and the first through-hole can be simultaneously obtained to power the processor, thereby reducing the current-carrying pressure on the second through-hole and ensuring stable power supply to the processor. In addition, it avoids the second through-hole from generating concentrated heat due to a large current-carrying capacity, which may cause local deformation of the mainboard. Furthermore, when the second through-hole is unstable in current carrying due to an unstable connection with the guide structure, the processor can obtain current through the first through-hole, thereby ensuring stable power supply.
[0020] In a possible implementation of the first aspect, the current-carrying layer includes multiple first conductive layers, which are used to conduct the power supply current; the multiple first conductive layers are arranged along the transverse direction of the mainboard, and the multiple first conductive layers are spaced apart; each first conductive layer is electrically connected to at least one through hole.
[0021] In this possible implementation, since the guide structure can be connected to the mainboard at multiple points, the first conductive layer can be set locally and in blocks, thereby forming multiple first conductive layers. While providing power to the electronic devices on the mainboard, it effectively reduces the amount of copper laid between the mainboard layers, shortens the flow path of the current in the mainboard layers, and reduces the flow area, thereby reducing the heat generated by the mainboard itself and improving the safety and stability of the mainboard operation.
[0022] In a possible implementation manner of the first aspect, the current-carrying layer includes a first conductive layer, the area of the first conductive layer is smaller than the cross-sectional area of the main board; and the first conductive layer is electrically connected to the through hole.
[0023] In this possible implementation, since the high-power electronic device can be powered vertically by the second through hole in conjunction with the high-conductivity portion, the power supply current that needs to be transmitted by the current-carrying layer is reduced, so that the copper layer area laid on the current-carrying layer can be reduced. A smaller copper layer area can reduce the heat generated inside the motherboard.
[0024] In a possible implementation of the first aspect, the current-carrying layer includes a first layer and a second layer arranged in sequence along the thickness direction of the mainboard; the first layer includes a first conductive layer, and the second layer includes a second conductive layer; the second conductive layer is electrically connected to the through hole; and the thickness of the second conductive layer is less than or equal to the thickness of the first conductive layer.
[0025] In this possible implementation, the first conductive layer on the first layer or the second conductive layer on the second layer can be connected to the first conductive layer on the first layer or the second conductive layer on the second layer through a via or a blind hole according to the power supply requirements of different electronic devices, so as to obtain an adaptive power supply current. By arranging the current-carrying layer in layers, the adaptability of the current-carrying capacity of the conductive layer to the power supply requirements of the electronic devices on the mainboard can be improved.
[0026] In a possible implementation of the first aspect, an orthographic projection of at least one second conductive layer on the first layer at least partially overlaps with the first conductive layer, and the orthographic projection of the second conductive layer partially overlaps with the first conductive layer. The second conductive layer and the first conductive layer may be electrically connected in a direction along the thickness of the mainboard.
[0027] In this possible implementation, the current-carrying layer is improved to improve the longitudinal current-carrying capacity of the mainboard, thereby improving the stability of power supply to high-power devices on the first surface and avoiding local overheating of the single-layer conductive layer in the mainboard due to excessive current-carrying capacity.
[0028] In a possible implementation of the first aspect, the mainboard also includes a signal layer, which is arranged on the same layer as the current-carrying layer; the signal layer includes one or more third conductive layers, which are arranged on one side of the first conductive layer, or between two adjacent first conductive layers.
[0029] In this possible implementation, by improving the power supply structure, the first conductive layer in the current-carrying layer can be partially laid, so that the signal layer and the current-carrying layer can be arranged on the same layer, reducing the thickness of the mainboard. The reduction in the thickness of the mainboard can make it easier to weld electronic devices and reduce the cost of board manufacturing.
[0030] In a possible implementation of the first aspect, the mainboard also includes a grounding layer, which is arranged on the same layer as the current-carrying layer; the grounding layer includes one or more fourth conductive layers, which are arranged on one side of the first conductive layer, or between two adjacent first conductive layers.
[0031] In this possible implementation, by improving the power supply structure, the first conductive layer in the current-carrying layer can be partially laid, so that the grounding layer and the current-carrying layer can be arranged on the same layer, reducing the thickness of the mainboard. The reduction in the thickness of the mainboard can make it easier to weld electronic devices and reduce the cost of board manufacturing.
[0032] In a possible implementation of the first aspect, the mainboard has a second surface away from the first surface; the mainboard includes two current-carrying layers, one of which is disposed close to the first surface of the mainboard, and the other current-carrying layer is disposed close to the second surface of the mainboard; the mainboard also includes at least one functional inner layer, and the functional inner layer is any one or more of a signal layer, a power layer, or a ground layer. For example, the functional inner layer is a signal layer and a power layer disposed on the same layer.
[0033] In this possible implementation, by providing two current-carrying layers, bidirectional power supply can be achieved, so that electronic devices can be provided on both surfaces of the mainboard. In addition, the two current-carrying layers supply power together, further improving the current carrying capacity of the mainboard.
[0034] In a possible implementation manner of the first aspect, the guide structure includes one or more guide bars; the guide bars extend along any one direction of a length direction, a width direction, or a diagonal direction of the mainboard.
[0035] In this possible implementation, by adjusting the extension track of the guide bar relative to the mainboard, the guide bar can provide vertical power supply to more high-power electronic devices, thereby improving the power supply flexibility.
[0036] In a possible implementation manner of the first aspect, the guide structure includes a plurality of guide plates stacked in layers, and an anti-corrosion layer sandwiched between two adjacent layers of guide plates.
[0037] In this possible implementation, when the guide plate on the surface of the guide structure is corroded, the anti-corrosion layer can isolate or slow down the speed of the corrosion spreading downward, thereby reducing the impact of corrosion on the overall flow capacity of the guide bar structure and ensuring stable power supply to the motherboard. In addition, since the corresponding number of guide plates can be selected according to the required flow rate, it is helpful to match computing nodes with different flow requirements and achieve universality for different computing nodes.
[0038] In a possible implementation of the first aspect, in a possible implementation of the first aspect, there are multiple main boards, including a first main board and a second main board; the second main board is located on one side of the first main board, and the first surface of the first main board and the first surface of the second main board face the same direction; the guide structure includes a first guide bar and a first connecting bar, the first guide bar is arranged on one side of the first main board and is fixedly connected to the flow hole of the first main board; one end of the first connecting bar is fixedly connected to the first guide bar, and the other end of the first connecting bar is connected to the flow hole of the second main board.
[0039] In this possible implementation, power supply to multiple mainboards is achieved through the improvement of the diversion structure, thereby improving the power supply density. Since multiple mainboards are set up, the computing power of the computing node is improved.
[0040] In a possible implementation of the first aspect, there are multiple main boards, including a third main board and a fourth main board; the third main board is adjacent to the fourth main board, and the first surface of the third main board and the first surface of the fourth main board are facing opposite sides respectively; the guide structure is clamped between the third main board and the fourth main board, and is fixedly connected to the flow holes of the third main board and the flow holes of the fourth main board respectively.
[0041] In this possible implementation, power supply to multiple mainboards is achieved through the improvement of the diversion structure, thereby improving the power supply density. Since multiple mainboards are set up, the computing power of the computing node is improved.
[0042] In a possible implementation of the first aspect, the mainboard includes multiple computing units, each computing unit includes a processor, a memory bar arranged on one side of the processor, and a heat sink; the computing units are arranged in an array or in sequence on the mainboard.
[0043] In this possible implementation, the power supply capacity is improved through the diversion structure and the mainboard structure setting of the embodiment of the present application, and the number of computing units arranged on the mainboard can be increased, thereby improving the computing power.
[0044] In a possible implementation of the first aspect, the computing node also includes a confluence device; the confluence device has multiple power input parts and power output parts; each power input part is connected to a power supply, and the power output part is connected to the guide structure.
[0045] In this possible implementation, the power supply currents output by multiple power modules are converged by a converging device, thereby improving the output capacity of the power supply current and improving the power supply density.
[0046] A second aspect of an embodiment of the present application provides a computing device, which includes a housing and the computing node described in the first aspect; the computing node is installed in the housing.
[0047] Depending on the type of computing device, the housing can be in different styles. For example, when the computing device is a rack-type or tower-type, the housing can be a box-type structure; when the computing device is a blade server, the housing can be a base style or a box-type. No specific limitation is made here.
[0048] In a possible implementation of the second aspect, the computing device also includes an immersion cabinet and an electric power input port, a liquid outlet and a liquid inlet provided on the immersion cabinet; the immersion cabinet is filled with cooling liquid, and the computing nodes are immersed in the cooling liquid; the power supply is arranged outside the immersion cabinet; the power supply and the guide structure are connected by a power cord; the electric power input port is used for the power cord to pass into the immersion cabinet; the liquid outlet is used for gaseous or liquid cooling liquid to flow out, and the cooled cooling liquid flows back into the immersion cabinet through the liquid inlet.
[0049] In this possible implementation, the immersion liquid-cooled computing device breaks through the bottleneck of traditional on-board power supply based on the efficient heat dissipation capability of liquid cooling and the high power supply density power supply architecture of the embodiment of the present application, thereby allowing multiple processors and associated sockets, radiators and memory to be set on the motherboard, thereby achieving an increase in computing power density while ensuring the power supply capability and power supply stability for high-power devices on the motherboard.
[0050] In a possible implementation of the second aspect, a plurality of parallel guide structures are provided in the immersion cabinet, each guide structure being used to supply power to one or more mainboards; the computing device further comprises a confluence device, which is strip-shaped and extends along the arrangement direction of the plurality of guide structures; each guide structure is fixedly connected to the confluence device.
[0051] In this possible implementation, the flow guiding structure may be in a strip shape. Each flow guiding structure is provided to supply power to one or more mainboards, thereby further improving the power supply density. By providing multiple flow guiding structures arranged in parallel in the immersion cabinet, the utilization rate of the space in the immersion cabinet is improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 This is a top view of a computing node according to an embodiment of the present application.
[0053] Figure 2 It is a side view of a motherboard computing node according to an embodiment of the present application.
[0054] Figure 3 It is a schematic diagram showing the arrangement of a guide structure relative to a mainboard according to an embodiment of the present application.
[0055] Figure 4 It is a schematic diagram showing another arrangement of a guide structure relative to a mainboard according to an embodiment of the present application.
[0056] Figure 5 It is a schematic diagram showing another arrangement of a guide structure relative to a mainboard according to an embodiment of the present application.
[0057] Figure 6 It is a schematic diagram showing another arrangement of a guide structure relative to a mainboard according to an embodiment of the present application.
[0058] Figure 7 This is a partial structural breakdown diagram of a diversion structure according to an embodiment of the present application.
[0059] Figure 8 It is a side sectional view of a mainboard according to an embodiment of the present application.
[0060] Fig. 9 It is a side sectional view of another mainboard according to an embodiment of the present application.
[0061] Fig.10 This is a side sectional view of another mainboard according to an embodiment of the present application.
[0062] Fig.11 It is a side view of a computing node in an architecture where two mainboards are stacked in the same direction according to an embodiment of the present application.
[0063] Fig.12 It is a side view of another computing node under the architecture of two mainboards stacked in the same direction shown in an embodiment of the present application.
[0064] Fig.13 It is a side view of another computing node under the architecture of two mainboards stacked in the same direction shown in an embodiment of the present application.
[0065] Fig.14 It is a side view of another computing node under the architecture of motherboards stacked in the same direction shown in an embodiment of the present application.
[0066] Fig.15This is a side view of the architecture in which two main boards are arranged back to back according to an embodiment of the present application.
[0067] Fig.16 It is a side view of another computing node under the architecture of two mainboards arranged back to back according to the embodiment of the present application.
[0068] Fig.17 It is a side view of a computing node in an architecture where multiple motherboards are arranged in the same direction according to an embodiment of the present application.
[0069] Fig.18 It is a schematic diagram of the connection structure between a single-phase immersion liquid-cooled computing device and an external cooling device according to an embodiment of the present application.
[0070] Fig.19 It is a schematic diagram of a two-phase immersion liquid-cooled computing device according to an embodiment of the present application.
[0071] The reference numerals are described as follows: 100, computing node;
[0072] 10. Mainboard; 11. Electronic device; 111. Processor module; 112. Computing unit; 113. Heat sink; 114. Memory; 115. Processor; 121. First surface; 122. Second surface; 10a. First mainboard; 10b. Second mainboard; 10c. Third mainboard; 10d. Fourth mainboard; 13. Current-carrying layer; 131. First layer; 1311. First conductive layer; 132. Second layer; 1321. Second conductive layer; 133. Signal layer; 1331. Third conductive layer; 134. Ground layer; 1341. Fourth conductive layer; 135. Insulating layer; 14. Functional inner layer; 15. Through hole; 151. First through hole; 152. Second through hole; 16. High conductive part; 17. Conductive fixing member; 181. First via hole; 182. First buried hole; 136. First position;
[0073] 20. guide structure; 21. guide strip; 211. first guide strip; 221. first connecting strip; 222. second connecting strip; 212. second guide strip; 213. third guide strip; 223. third connecting strip; 224. fourth connecting strip; 214. fourth guide strip; 23. guide plate; 24. anti-corrosion layer;
[0074] 30. Power module;
[0075] 40. Converging device; 41. Electric energy input unit;
[0076] 50. PCIE card; 60. Fan; 70. Hard disk backplane; 80. Hard disk; 90. Housing;
[0077] 200. Computing equipment; 210. Immersed cabinet; 2101. Power input port; 2102. Liquid outlet; 2103. Liquid inlet; 220. Cooling device. DETAILED DESCRIPTION
[0078] In order to enable those skilled in the art to better understand the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly described below in conjunction with the drawings in the embodiments of the present application.
[0079] It should be understood that the "multiple" mentioned in the embodiments of the present application refers to two or more. In the description of the embodiments of the present application, unless otherwise specified, " / " means or, for example, A / B can represent A or B; "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships, for example, A and / or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, in order to facilitate the clear description of the technical solution of the embodiments of the present application, the words "first" and "second" are used to distinguish the same or similar items with the same function and action base. Those skilled in the art can understand that the words "first" and "second" do not limit the quantity and execution order, and the words "first" and "second" do not limit them to be different.
[0080] In order to further illustrate the principle and structure of the present invention, embodiments of the present invention are now described in detail with reference to the accompanying drawings.
[0081] As the demand for computing power of computing nodes increases, more and more types of processors need to be deployed on the computing node motherboard, such as CPU (Central Processing Unit), TPU (Tensor Processing Unit), GPU (Graphics Processing Unit), DPU (Deep learning Processing Unit), NPU (Neural network Processing Unit), etc.; the power of the processor continues to evolve towards high power. From the physical form, the development of high-density computing power of computing nodes has also led to the continuous increase in the arrangement density of electronic devices on the motherboard, such as the emergence of half-width computing nodes. Since the width of a half-width computing node is half the width of a conventional standard-sized computing node, the arrangement of electronic devices on its motherboard is very tight. The development of high-density computing power of computing nodes needs to rely on stable and reliable power supply to electronic devices on the motherboard.
[0082] However, existing server power supply technologies are mainly based on a two-dimensional horizontal method; for example, the server PSU (Powersupply unit) receives the power supply current through the connector on the motherboard, and the copper layer inside the motherboard PCB is set as a power supply plane to transmit the power supply current introduced from the connector, thereby supplying power to the various components on the PCB board. However, based on this type of motherboard power supply structure, when a higher current flow density is required, it can only be achieved by increasing the thickness of the copper layer or the number of copper layers in the motherboard PCB. However, too thick or too many copper layers will cause difficulties in the PCB board punching layout, and will also generate more heat due to the continuous conduction of large currents, causing the internal temperature of the motherboard to be too high, seriously affecting the normal operation of various electronic devices; even in some serious cases, it may cause fires, explosions and other accidents.
[0083] In response to the above problems, an embodiment of the present application proposes a computing node, which improves the power supply structure of the mainboard in the computing node to improve the flow capacity of the power supply current, thereby meeting the power supply requirements of high-power devices and high-density electronic device layouts on the mainboard, and supporting the development trend of continued growth in computing power of the computing node in the future.
[0084] First embodiment
[0085] See also Figure 1 , Figure 1 It is a top view of a computing node 100 shown according to an embodiment of the present application. In this embodiment, the computing node 100 includes a housing 90, a mainboard 10 disposed in the housing 90, a power module 30, a confluence device 40, a flow guide structure 20, etc. The mainboard 10 includes a variety of electronic devices 11, for example, the mainboard 10 includes a processor 115, a memory 114, etc. Among them, the processor 115 is a high-power device. Taking a rack server as an example, the housing 90 can also generally include a PCIE (Peripheral Component Interconnect Express, a high-speed serial computer expansion bus standard) card 50, a fan 60, a hard disk backplane 70, a hard disk 80 and other devices, a light board and other components.
[0086] For blade servers, the composition of the computing node 100 is different from that of rack servers. The blade server includes a base, on which multiple single board computers can be inserted. When a single board computer is regarded as a computing node 100, the computing node 100 in this embodiment may not have the above-mentioned housing 90, and may not have the above-mentioned power module 30, converging device 40, PCIE card 50, fan 60, hard disk 80 backplane 70, hard disk 80, light board and other components.
[0087] See also Figure 2 , Figure 2It is a side view of a computing node of the mainboard 10 according to an embodiment of the present application. Figure 2 The dotted lines in the figure are used to illustrate the current heat transmission path. The power supply may be a power module 30 disposed in the housing 90 of the computing node 100, or an external power supply disposed outside the housing 90 of the computing node 100. In one example, the power supply may include a plurality of power modules 30, and the plurality of power modules 30 may be stacked in sequence or arranged horizontally in the same direction. The electric energy output by the plurality of power modules 30 may be converged by a convergence device 40, and the electric energy after convergence by the convergence device 40 is transmitted to the flow guide structure 20.
[0088] Specifically, the power module 30 has an electric energy output terminal, and the converging device 40 may have a plurality of electric energy input parts 41 and electric energy output parts; each electric energy input part 41 is connected to a power module 30, and the electric energy output part is connected to the guide structure 20, thereby realizing the converging of the output currents of the plurality of power modules 30. The specific structure of the converging device 40 is not limited here.
[0089] In one example, the busbar 40 includes a busbar made of a conductive metal, such as copper or aluminum. The busbar 40 may be a strip-shaped structure, and its extension direction is substantially consistent with the arrangement direction of the multiple power modules 30. When the positions of the multiple power modules 30 are relatively dispersed, the busbar 40 is provided with one or more bends to match the arrangement positions of the multiple power modules 30, thereby shortening the connection path with each power module 30.
[0090] The confluence device 40 may also include an overcurrent protector. When the current on the guide structure 20 or the confluence device 40 exceeds a threshold value, the overcurrent protector is activated to cut off the electrical connection between the confluence device 40 and the guide structure 20, thereby protecting the electronic devices 11 on the mainboard 10 from overcurrent shock.
[0091] In this embodiment, the convergence device 40 is used to converge multiple power modules 30, thereby improving the current output capacity, which is beneficial for powering the motherboard 10 with high power density / high-density computing power, and is beneficial for powering multiple motherboards 10 at the same time.
[0092] Here with Figure 2 The orientation shown in is for reference, the upper and lower surfaces of the mainboard 10 are respectively the first surface 121 and the second surface 122; the processor 115 and other electronic devices are arranged on the first surface 121; of course, they can also be arranged on the first surface 121 and the second surface 122 at the same time. The flow guide structure 20 is connected between the confluence device 40 and the mainboard 10, and is used to transmit the power output by the confluence device 40 to the mainboard 10, thereby powering the electronic devices 11 on the mainboard 10.
[0093] It should be noted that the guide structures 20 are usually arranged in pairs, one guide structure 20 is connected to the positive pole of the power output part of the converging device 40 for outputting current; the other guide structure 20 (not shown) is connected to the negative pole of the power output part of the converging device 40 for returning current. The structures and styles of the two guide structures 20 can be the same. No distinction is made in the following description of the embodiments.
[0094] The flow guiding structure 20 is arranged on one side of the mainboard 10 and is arranged substantially along the extension direction of the surface of the mainboard 10. Here, "one side" means that the flow guiding structure 20 can be located on the side where the first surface 121 of the mainboard 10 is located, or on the side where the second surface 122 of the mainboard 10 is located. For example, the flow guiding structure 20 is located on one side of the second surface 122 of the mainboard 10, and can be attached to the second surface 122 of the mainboard 10, or there can be a gap between the flow guiding structure 20 and the second surface 122 of the mainboard 10. In the process of supplying power to the electronic device 11 on the mainboard 10, the power supply current is first conducted outside the mainboard 10 through the flow guiding structure 20, and enters the mainboard 10 through the electrical connection between the flow guiding structure 20 and the second surface 122 of the mainboard 10. For example, the power supply current is first conducted outside the mainboard 10 through the flow guiding structure 20, and enters the mainboard 10 substantially in a direction perpendicular to the surface of the mainboard 10 through the electrical connection between the flow guiding structure 20 and the second surface 122 of the mainboard 10. In the mainboard 10, the power supply current can be transmitted upward substantially along the thickness direction of the mainboard 10, so as to reach the power supply end of the electronic device 11.
[0095] Here, the transmission path of the power supply current in the mainboard 10 may include a transmission path transmitted along the thickness direction of the mainboard 10, so as to realize power supply to the electronic device 11 to be powered. The transmission path of the power supply current in the mainboard 10 may also include a transmission path along the transverse direction of the mainboard 10. For example, after the power supply current enters the mainboard 10 along the thickness direction of the mainboard 10, it is transmitted laterally along the mainboard 10 to realize power supply to the electronic device 11.
[0096] In this embodiment, the guide structure 20 may include a guide strip 21 and / or a guide plate.
[0097] See also Figure 3 , Figure 3 1 is a schematic diagram showing an arrangement of a guide structure 20 relative to a main board 10 according to an embodiment of the present application. In one implementation of the guide bar 21, the guide structure 20 includes a guide bar 21, and the guide bar 21 can extend along the length, width, or diagonal direction of the main board 10; of course, it can also extend in other directions.
[0098] See also Figure 4 and Figure 5 , Figure 4 and Figure 5It is a schematic diagram of another arrangement of a guide structure 20 relative to the mainboard 10 according to an embodiment of the present application. In another implementation of the guide bar 21, there are multiple guide bars 21, and the multiple guide bars 21 can be arranged in parallel, and all extend along the length, width or diagonal direction of the mainboard 10. The multiple guide bars 21 can also extend in different directions respectively. For example, at least one guide bar 21 extends along the length direction of the mainboard 10, and at least one guide bar 21 extends along the width direction of the mainboard 10. Among them, the flow capacity of the multiple guide bars 21 can be different; the cross-sectional area of the guide bar 21 can be adjusted so that each guide bar 21 has a different flow capacity, and accordingly supplies power to electronic devices 11 of different power levels on the mainboard 10.
[0099] See also Figure 6 , Figure 6 1 is a schematic diagram of another arrangement of the guide structure 20 relative to the mainboard 10 according to an embodiment of the present application. In these two implementations, the guide bar 21 can be a linear structure, or it can have one or more bends according to the arrangement position of the electronic device 11 on the mainboard 10, so as to adjust the extension track of the guide bar 21 relative to the mainboard 10, so that the guide bar 21 can vertically power more high-power electronic devices.
[0100] In this embodiment, the number and extension direction of the guide bars 21 can be flexibly set according to the layout positions of the processor 115 and other high-power devices on the motherboard 10, so that the guide bars 21 can supply power to the processor 115 and other high-power devices, shorten the lateral flow path of the power supply current in the motherboard 10, thereby reducing the current density in the motherboard 10 and reducing the heat generated by the motherboard 10.
[0101] In one implementation of the guide plate, the guide structure 20 is plate-shaped and substantially parallel to the main board 10. The area of the guide structure 20 may be smaller than that of the main board 10, or may be substantially equal to that of the main board 10. The guide structure 20 provided in a plate shape can greatly improve the flow capacity, and by connecting with the main board 10 at multiple points, vertical power supply to more electronic devices 11 is achieved, thereby further shortening the lateral flow path of the power supply current in the main board 10, and significantly reducing the heat generated by the main board 10.
[0102] It can be understood that since the guide structure 20 is plate-shaped, the guide structure 20 can be vertically connected to a larger number of high-power devices on the mainboard 10, so as to directly transmit current to these high-power devices, effectively shortening the lateral flow path of the power supply current in the mainboard 10, thereby reducing the current density in the mainboard 10 and reducing the heat generated by the mainboard 10.
[0103] In the present embodiment, the power supply current output from the converging device 40 is conducted to the mainboard 10 through the flow guiding structure 20. Since the flow guiding structure 20 and the mainboard 10 are independent structures, the flow guiding structure 20 with a suitable flow capacity can be selected according to the power of the mainboard 10 to be powered, without being limited by the flow capacity of the mainboard 10, thereby improving the power supply density of the computing node 100, which is conducive to promoting the development of the computing node 100 towards high-density computing power and high-density power.
[0104] Each motherboard 10 may have one or more processors 115. Corresponding to each processor 115, a memory 114 slot and a heat sink 113 packaged with the processor 115 are generally configured. The motherboard 10 may also include a chipset, a PCIE slot, and various ports. The type of the processor 115 on the motherboard 10 is not limited here, and the multiple processors 115 may be of the same or different types.
[0105] Here, a processor 115, a memory 114 slot of the processor 115 and a heat sink 113 can form a computing unit; therefore, the motherboard 10 can have one or more computing units. Multiple computing units can be arranged in an array, in a single row with intervals, or in an irregular arrangement, etc. The processor 115 corresponding to each computing unit can be vertically powered by the guide structure 20 in this embodiment, thereby improving the stability of power supply.
[0106] Since the electronic devices 11 in the computing node 100 are closely arranged and have a high power density, the computing node 100 generates a large amount of heat inside; and since the flow-guiding structure 20 may supply power to multiple motherboards 10 at the same time, the flow rate is large, resulting in a correspondingly large amount of heat. If space permits, a gap may be provided between the flow-guiding structure 20 and the powered motherboard 10. In an air-cooled computing node, a fan 60 is included, and the fan 60 takes away the heat in the computing node 100 by blowing air. Therefore, by setting a gap between the flow-guiding structure 20 and the powered motherboard 10, air circulation can be provided, thereby better dissipating the heat of the flow-guiding structure 20 and the motherboard 10. For a liquid-cooled computing node 100, the setting of the gap allows liquid cooling medium to circulate, thereby taking away the heat from the flow-guiding structure 20 and the motherboard 10.
[0107] When the guide structure 20 and the powered mainboard 10 are closely fitted or the gap is very small, a heat insulating layer may be provided between the guide structure 20 and the powered mainboard 10 to reduce the heat conducted from the guide structure 20 to the powered mainboard 10 .
[0108] As mentioned above, in this embodiment, the guide structure 20 is independently provided from the mainboard 10. When directly exposed to the environment, for the air-cooled computing node 100, the guide structure 20 will be at risk of being corroded by moisture, oxygen and other corrosive media in the air, thereby affecting the flow capacity.
[0109] See also Figure 7 , Figure 7 1 is a partial structural exploded diagram of a guide structure 20 according to an embodiment of the present application. In order to solve this problem, in this embodiment, the guide structure 20 may include a plurality of guide plates 23 stacked in layers. The plurality of guide plates 23 may be fixedly connected and electrically connected by welding at the ends, or may be fixedly connected and electrically connected by using metal conductive columns that penetrate the guide plates 23 at the same time. Each guide plate 23 may be isolated by an anti-corrosion layer 24.
[0110] The anti-corrosion layer 24 can be a separate layer, sandwiched between two adjacent guide plates 23, or a coating applied on the surface of each guide plate 23. The material of the anti-corrosion layer 24 can specifically be a polymer material such as acrylic acid, polyurethane, epoxy system, etc.; a membrane structure formed by metal tin, zinc, etc. can also be used. When the guide plate 23 located on the surface of the guide structure 20 is corroded, due to the isolation of the anti-corrosion layer 24, the speed of corrosion spreading downward can be isolated or slowed down, thereby reducing the impact of corrosion on the overall flow capacity of the guide bar 21 structure, and ensuring stable power supply to the mainboard 10. In addition, since the corresponding number of guide plates 23 can be selected according to the required flow rate, it is beneficial to match computing nodes 100 with different flow requirements, and realize the universality for different computing nodes 100.
[0111] See also Figure 8 , Figure 8 It is a side sectional view of a mainboard according to an embodiment of the present application. Figure 8 The dotted line in the middle is a schematic diagram of the current transmission direction. In this embodiment, the mainboard 10 includes a through hole 15 extending along its own thickness direction and a current-carrying layer 13, a signal layer 133, a functional inner layer 14, and a grounding layer 134 arranged along the transverse direction of the mainboard 10. The through hole 15 can be matched with a conductive fixing member 17 to achieve the connection between the guide structure 20 and the mainboard 10; the current-carrying layer 13 is used to provide a flow path for the power supply current to be conducted along the transverse direction of the mainboard 10. The current-carrying layer 13 is electrically connected to the hole wall of the through hole 15, so that the power supply current can be obtained from the through hole 15, and the electronic device 11 arranged along the extension path of the current-carrying layer 13 can obtain the power supply current from the current-carrying layer 13 through the first via 181.
[0112] Specifically, the through hole 15 is used to achieve the connection between the guide structure 20 and the main board 10. The hole wall of the through hole 15 is coated with a conductive medium, which can be copper or aluminum. The extension direction of the through hole 15 is roughly along the thickness direction of the main board 10. The guide structure 20 is fixedly connected to the through hole 15 through a conductive fixing, and an electrical connection is achieved while the fixed connection is achieved. The conductive fixing can be a bolt or a surface-mounted nut. Through holes are correspondingly provided on the guide structure 20, and the conductive fixings are sequentially penetrated into the through holes on the through hole 15 and the guide structure 20, thereby achieving a fixed connection and an electrical connection between the two.
[0113] See also Fig. 9 , Fig. 9 It is a side sectional view of another mainboard according to an embodiment of the present application. Among them, the dotted line with an arrow is a schematic diagram of the current transmission direction. A plurality of through-holes 15 can be provided to achieve multi-point vertical power supply between the flow-guiding structure 20 and the mainboard 10. The types of through-holes 15 may also not be exactly the same. In one implementation, the plurality of through-holes 15 include a first through-hole 151, and the first through-hole 151 penetrates the mainboard 10 along the thickness direction of the mainboard 10; the hole wall of the first through-hole 151 contacts the current-carrying layer 13 for achieving electrical connection.
[0114] In some other implementations, the through hole 15 may also be a blind hole. When the through hole 15 is a blind hole, the closed end of the through hole 15 is electrically connected to the current-carrying layer 13 in the main board 10; the open end of the through hole 15 is for the conductive fixing member 17 to penetrate to connect the flow guiding structure 20.
[0115] By setting the flow-guiding structure 20 and the flow-through hole 15, it is convenient to power the high-power devices on the mainboard 10 without considering the setting restrictions of the transmission path of the large current and the current-carrying capacity of the mainboard 10. Specifically, the flow-through hole 15 can be set close to the high-power devices on the mainboard 10, for example, the flow-through hole 15 can be located directly below the processor 115 or set corresponding to the side of the processor 115. During the power supply process, the power supply current is first transmitted outside the mainboard 10 through the flow-guiding structure 20, and after reaching the vicinity of the processor 115 to be powered, it enters the flow-through hole 15 through the conductive fixing member 17, and then along the extension direction of the flow-through hole 15, it is directly or indirectly transmitted to the power supply end of the processor 115, so as to realize vertical power supply for the processor 115. For example, when the flow-through hole 15 is a blind hole, the power supply current can also flow out of the flow-through hole 15 and then be transmitted to the power supply end of the processor 115 through the current-carrying layer 13 inside the mainboard 10.
[0116] In one implementation, in the circuit design, the processor 115 is usually powered on and off by a switching device, so the through hole 15 can be directly or indirectly electrically connected to the switching device to power the processor 115 under the control of the switching device. In the following example, the processor 115 and the switching device are combined and referred to as the processor module 111.
[0117] Based on the above embodiments, it can be known that the embodiment of the present application improves the current carrying capacity by providing the guide structure 20. Furthermore, the embodiment further improves the longitudinal current carrying capacity of the mainboard 10 based on the improvement of the internal structure of the mainboard 10, so that the large current transmitted on the guide structure 20 can be effectively and stably and reliably transmitted to the processor 115 on the mainboard 10 to power the processor 115.
[0118] Please continue reading Fig. 9 In one implementation, the longitudinal current-carrying capacity of the mainboard 10 is improved by improving the through-hole 15. Specifically, one or more through-holes 15 are provided on the mainboard 10, and the one or more through-holes 15 include at least one second through-hole 152. The first opening of the second through-hole 152 is for the conductive fixing member 17 to be inserted to achieve a fixed connection and an electrical connection with the flow-guiding structure 20. Here, the conductive fixing member 17 can be a surface-mounted nut. The second opening of the second through-hole 152 is connected to a high-conductivity portion 16. The high-conductivity portion 16 is a solid conductive structure, and extends to the top surface of the mainboard 10 along the thickness direction of the mainboard 10, and is electrically connected by directly contacting the power supply terminal of the switching device in the processor module 111, thereby achieving vertical power supply to the processor 115.
[0119] Here, the high-conductivity portion 16 is made of a conductive material, such as copper or aluminum; and the cross-sectional area of the high-conductivity portion 16 is greater than or equal to the cross-sectional area of the second through-hole 152, thereby improving the flow capacity and meeting the processor 115's demand for high current power supply.
[0120] Furthermore, the high-conductivity portion 16 can be arranged to gradually widen in the direction of the first surface 121 of the mainboard 10 to achieve a gradual increase in the cross-sectional area, so as to reduce the internal resistance of the high-conductivity portion 16 and reduce the heat generation, and adjust the shape and area of the high-conductivity portion 16 exposed on the first surface 121 of the mainboard 10 to adapt to the power supply end of the switching device.
[0121] In this implementation, the second through-hole 152 is electrically conductive by the conductive fixing member 17 and the hole wall of the second through-hole 152, thereby ensuring a large longitudinal current capacity. Furthermore, by providing a high-conductivity portion 16 connected between the second through-hole 152 and the processor module 111, the current is directly and stably transmitted to the surface of the motherboard 10, thereby being connected to the processor module 111, without relying on other layer structures in the motherboard 10 for current carrying, thereby improving the reliability and stability of power supply to the processor 115.
[0122] Further, the second opening of the second through-hole 152 and the high-conductivity portion 16 can be arranged to be connected at the first position 136 of the current-carrying layer 13, and when the first position 136 is also directly or indirectly electrically connected to other through-holes 15, such as the first through-hole 151, since the high-conductivity portion 16 is electrically connected to the second through-hole 152 and the first through-hole 151 at the same time, the current in the second through-hole 152 and the first through-hole 151 can be obtained at the same time to power the processor 115, thereby reducing the current-carrying pressure on the second through-hole 152 and ensuring stable power supply to the processor 115. In addition, it is avoided that the second through-hole 152 generates concentrated heat due to a large current carrying capacity, causing local deformation of the mainboard 10. Furthermore, when the second through-hole 152 is unstable in connection with the guide structure 20, resulting in unstable current carrying, the processor 115 can obtain current through the first through-hole 151, thereby ensuring stable power supply.
[0123] It should be understood that the motherboard 10 has at least two CPUs 115, and may also have other types of processors 115. In this solution, there may be multiple matching structures of the second through-hole 152 and the high-conductivity portion 16, so as to supply power to each processor 115 in a one-to-one correspondence manner.
[0124] As mentioned above, in this embodiment, along the transverse direction of the mainboard 10, the mainboard 10 includes a current-carrying layer 13, a signal layer 133, a functional inner layer 14 and a grounding layer 134; each layer can be isolated by an insulating layer. Among them, the current-carrying layer 13 and the hole wall of the through hole 15 are in contact to achieve electrical connection, so the current transmitted from the current-conducting structure 20 can be transmitted to the current-carrying layer 13 through the conductive fixing member 17 and the hole wall of the through hole 15. Since the current-carrying layer 13 extends along the transverse direction of the mainboard 10, the current can be transmitted along the transverse direction of the mainboard 10. In addition, the current-carrying layer 13 can be electrically connected to the power supply end of the electronic device 11 on the mainboard 10 by setting a via, a blind hole or a buried hole, so as to achieve connection with the power supply end of the electronic device 11. In practical applications, the current-carrying layer 13 can be used to power medium and low-power electronic devices 11.
[0125] In order to meet the power supply requirements of the electronic device 11, the current-carrying layer 13 includes a first conductive layer 1311, and the first conductive layer 1311 can be made of metal copper or aluminum. In this embodiment, copper is used as an example, and the first conductive layer 1311 is a thick copper layer; specifically, it can be a 2-ounce to 4-ounce copper layer, for example, a 2-ounce copper layer.
[0126] Please continue reading Fig. 9 In one implementation, the current-carrying layer 13 includes a first conductive layer 1311, and the area of the first conductive layer 1311 is smaller than the cross-sectional area of the mainboard 10. Since the high-power electronic device can be powered vertically by the second through hole 152 in conjunction with the high-conductivity portion 16, the power supply current that needs to be transmitted by the current-carrying layer 13 is reduced, so that the area of the copper layer laid by the current-carrying layer 13 can be reduced. In this implementation, the shape of the current-carrying layer 13 can be square or irregular, and its contour can be designed to be straight or curved according to the circuit board layout.
[0127] In another implementation, since the current in the flow-guiding structure 20 can be conveniently led vertically to the current-carrying layer 13 through the flow-through hole 15, it is not necessary to lay the first conductive layer 1311 in the entire current-carrying layer 13, but it can be laid locally. For example, a plurality of first conductive layers 1311 are included in the current-carrying layer 13, and the plurality of first conductive layers 1311 can be electrically connected through local parts or electrically isolated. In this implementation, since the flow-guiding structure 20 can be connected to the mainboard 10 at multiple points, the first conductive layer 1311 can be set locally and in blocks, thereby forming a plurality of first conductive layers 1311, which effectively reduces the amount of copper laid between the board layers of the mainboard 10 while realizing power supply to the electronic device 11 on the mainboard 10, shortens the flow path of the current in the board layers of the mainboard 10, and reduces the flow area, thereby reducing the heat generation of the mainboard 10 itself and improving the safety and stability of the operation of the mainboard 10.
[0128] In one implementation, the current-carrying layer 13 may have only one layer. In another implementation, the current-carrying layer includes a first layer 131, a second layer 132, and an insulating layer 135 sandwiched between the first layer 131 and the second layer 132, which are sequentially arranged along the thickness direction of the main board 10. The first layer 131 and the second layer 132 both extend along the transverse direction of the main board 10, and are both electrically connected to the hole wall of the through hole 15 to obtain the current transmitted by the self-guiding structure 20. The first layer 131 has at least one first conductive layer 1311, and the second layer 132 has at least one second conductive layer 1321; the thickness of the second conductive layer 1321 is less than the thickness of the first conductive layer 1311. Here, the first conductive layer 1311 and the second conductive layer 1321 can obtain current from different through holes 15.
[0129] Here, a copper layer with a thickness within a certain range may be referred to as a first conductive layer 1311, and thus the thicknesses of the plurality of first conductive layers 1311 may not be equal. A copper layer with a thickness within another range may be referred to as a second conductive layer 1321, and the thicknesses of the plurality of second conductive layers 1321 may not be equal. The thickness range of the first conductive layer 1311 and the thickness range of the second conductive layer 1321 may overlap.
[0130] Since the current carrying capacity of the first conductive layer 1311 and the second conductive layer 1321 are different, and the power supply requirements of various electronic devices 11 on the mainboard 10 are also different, it is possible to connect to the first conductive layer 1311 on the first layer 131 or to the second conductive layer 1321 on the second layer 132 through vias or blind holes according to the power supply requirements of different electronic devices 11, so as to obtain an adaptive power supply current. By arranging the current-carrying layer 13 in layers, the adaptability of the current carrying capacity of the conductive layer to the power supply requirements of the electronic devices 11 on the mainboard 10 can be improved.
[0131] The insulating layer 135 may be made of a corrosive resin material. The insulating layer 135 can block the electrical interference between the first layer 131 and the second layer 132 and support the first layer 131 or the second layer 132 .
[0132] It should be understood that the current-carrying layer 13 may include three or more layers, and a conductive layer for power supply is provided on each layer, so that the electronic devices 11 on the mainboard 10 can be classified according to their demand for power supply current, thereby realizing graded power supply to the electronic devices 11.
[0133] See also Fig.10 , Fig.10 1 is a cross-sectional view of another mainboard 10 according to an embodiment of the present application. Here, the longitudinal current carrying capacity of the mainboard 10 can also be improved by improving the current carrying layer 13. Specifically, the positive projection of at least one second conductive layer 1321 in the second layer 132 on the first layer 131 can be set to at least partially overlap with the first conductive layer 1311. It is assumed that the projection of the A area of the second conductive layer 1321 overlaps with the B area of the first conductive layer 1311. Therefore, in the process of transmitting the power supply current to the first surface 121 of the mainboard 10, the positive projection of the second conductive layer 1321 on the first layer 131 can be used. Fig.10 The first buried via 182 in the mainboard 10 electrically connects the A region and the B region along the thickness direction of the mainboard 10, thereby improving the longitudinal current carrying capacity of the mainboard 10, improving the stability of the power supply to the high-power devices on the first surface 121, and also avoiding the situation in which a single-layer conductive layer in the mainboard 10 is locally overheated due to excessive current carrying capacity. It should be understood that the first buried via 182 can be replaced by a through hole, a blind hole or other solid conductor structure.
[0134] The signal layer 133 is used to conduct signals between the electronic devices 11. As mentioned above, since the current in the current-conducting structure 20 can be conveniently led vertically to the current-carrying layer 13 through the through-holes 15, it is not necessary to lay the first conductive layer 1311 in the entire current-carrying layer 13, but it can be laid locally. Therefore, in one implementation, the signal layer 133 can be set in the area of the current-carrying layer 13 where the first conductive layer 1311 is not laid; for example, the signal layer 133 is set in the same layer as the first layer 131, thereby reducing the thickness of the mainboard 10 and improving the space utilization rate in the mainboard 10.
[0135] When the current-carrying layer 13 includes a first layer 131 and a second layer 132, the signal layer 133 can be arranged in the same layer as the first layer 131 or the second layer 132, or can be arranged in the same layer as the first layer 131 and the second layer 132 at the same time. It should be understood that at this time, the signal layer 133 is also divided into two layers accordingly.
[0136] Here, the signal layer 133 is set in the same layer as the first layer 131. A third conductive layer 1331 for conducting signals is set in the current-carrying layer 13. The third conductive layer 1331 can also be made of metal copper or aluminum. The thickness of the third conductive layer 1331 can be less than the thickness of the first conductive layer 1311. Specifically, it can be 0.5 ounces to 2 ounces of copper, for example, a copper layer with a thickness of 1 ounce. The third conductive layer 1331 can be electrically connected to the electronic device 11 on the mainboard 10 by setting a via, a blind hole or a buried hole, so as to transmit the signal between the electronic devices 11.
[0137] Here, there may be one or more third conductive layers 1331. As mentioned above, the first conductive layer 1311 on the first layer 131 may include one or more first conductive layers 1311. When there is only one integral first conductive layer 1311 on the first layer 131, the third conductive layer 1331 may be arranged on one side of the first conductive layer 1311. When there are multiple first conductive layers 1311 on the first layer 131, since there may be multiple first conductive layers 1311, there may also be one or more gaps between the multiple first conductive layers 1311, and the gap areas between the multiple first conductive layers 1311 may be different, so one or more third conductive layers 1331 may be laid in each of the above gaps. Multiple third conductive layers 1331 may be electrically connected or electrically isolated, which is specifically determined according to the layout design of the motherboard 10.
[0138] The signal layer 133 and the second layer 132 are arranged in the same layer in a similar manner, which will not be described in detail here.
[0139] In the implementation method of setting the signal layer 133 and the current-carrying layer 13 in the same layer, during the board making process, the entire copper layer or aluminum layer can be laid in the board layer, and the first conductive layer 1311 and the third conductive layer 1331 with different thicknesses can be produced by etching, thereby improving the board making efficiency.
[0140] In this embodiment, the first conductive layer 1311 in the current-carrying layer 13 can be partially laid by improving the power supply structure, so that the signal layer 133 and the current-carrying layer 13 can be arranged on the same layer, thereby reducing the thickness of the main board 10. The reduction in the thickness of the main board 10 can make it easier to weld the electronic device 11 and reduce the board manufacturing cost.
[0141] The grounding layer 134 is used to return the current returned by the electronic device 11 to the ground. Fig. 9 In one implementation, the grounding layer 134 can be independently arranged from the current-carrying layer 13. For example, the grounding layer 134 is arranged close to the second surface 122 of the mainboard 10, and the current-carrying layer 13 is arranged close to the first surface 121 of the mainboard 10. The distribution and thickness of the copper layer on the grounding layer 134 can be set to achieve impedance matching between the grounding layer 134 and the current-carrying layer 13.
[0142] In another implementation, the grounding layer 134 can be arranged in the area on the current-carrying layer 13 where the conductive layer is not laid, and the grounding layer 134 and the current-carrying layer 13 are arranged on the same layer. Specifically, the grounding layer 134 includes a fourth conductive layer 1341, and the fourth conductive layer 1341 can also be a copper or aluminum layer. The grounding layer 134 can be arranged on the same layer as one of the first layer 131 or the second layer 132, and can also be arranged on the same layer as both the first layer 131 and the second layer 132.
[0143] Here, there may be one or more fourth conductive layers 1341. As mentioned above, the first conductive layer 1311 on the first layer 131 may include one or more first conductive layers 1311. When there is only one integral first conductive layer 1311 on the first layer 131, the fourth conductive layer 1341 may be arranged on one side of the first conductive layer 1311. When there are multiple first conductive layers 1311 on the first layer 131, since there may be multiple first conductive layers 1311, there may also be one or more gaps between the multiple first conductive layers 1311, and the gap areas between the multiple first conductive layers 1311 may be different, so one or more fourth conductive layers 1341 may be laid in each of the above gaps, and the multiple fourth conductive layers 1341 may be electrically connected or electrically isolated, which is specifically determined according to the layout design of the motherboard 10.
[0144] The ground layer 134 and the second layer 132 are arranged in the same layer in a similar manner, which will not be described in detail here.
[0145] As mentioned above, the mainboard 10 may further include one or more functional inner layers 14, where the functional inner layer 14 may be a signal layer, a power layer, or one or more of a ground layer. Whether the functional inner layer 14 is provided and the number of layers provided may be determined based on the signal transmission amount and power supply requirements of the electronic device 11 on the mainboard 10. In some other implementations, the provision of the functional inner layer 14 is not necessary.
[0146] Based on the above description of each layer of the mainboard 10, it can be understood that in this embodiment, the number of current-carrying layers 13 can be one or more. In one implementation, two current-carrying layers 13 can be provided, namely, a first current-carrying layer and a second current-carrying layer. The first current-carrying layer can be provided close to the first surface 121 of the mainboard 10, and the second current-carrying layer can be provided close to the second surface 122 of the mainboard 10. One or more functional inner layers 14 can be provided between the first current-carrying layer and the second current-carrying layer. In this arrangement, the first current-carrying layer and / or the second current-carrying layer can be provided in the same layer as the signal layer 133 and the ground layer 134.
[0147] In another implementation, the second current-carrying layer may be replaced by the ground layer 134. In this case, the first current-carrying layer may be disposed only on the same layer as the signal layer 133.
[0148] Second embodiment
[0149] The computing node 100 may have only one mainboard 10 or may have multiple mainboards 10 at the same time. In this embodiment, the computing node 100 includes multiple mainboards 10, and the multiple mainboards 10 are stacked in the same direction. "Multiple" here means two or more. The stacking direction can be stacked upward or stacked in the horizontal direction. "Same direction" means that the electronic devices 11 on the multiple mainboards 10 are all arranged on the first surface 121.
[0150] See also Fig.11 , Fig.11 It is a side view of a computing node in an architecture of two motherboards stacked in the same direction according to an embodiment of the present application. Specifically, the plurality of motherboards 10 include a first motherboard 10a and a second motherboard 10b, and the second motherboard 10b is located on one side of the first motherboard 10a; the upper side is taken as an example. Since the electronic device 11 is arranged on the first motherboard 10a, the gap height between the first motherboard 10a and the second motherboard 10b is relatively large.
[0151] When multiple mainboards 10 are stacked in the same direction, the structure of the guide structure 20 can be set according to the relative position relationship of the multiple mainboards 10, so as to provide power to the multiple mainboards 10 at the same time and improve the power supply density.
[0152] Here, Fig.11The orientation shown in is for reference. In one implementation, the guide structure 20 includes a first guide bar 211 and a first connecting bar 221. Among them, the first guide bar 211 can be arranged on the upper side of the first mainboard 10a, and the current on the first guide bar 211 is transmitted downward to the first mainboard 10a. One end of the first connecting bar 221 is fixedly connected to the first guide bar 211, and the other end is fixedly connected to the second mainboard 10b to transmit current to the second mainboard 10b and support the second mainboard 10b. The first connecting bar 221 can play a role in conducting current on the one hand, and can be used to support the second mainboard 10b on the other hand, thereby improving the structural stability within the computing node 100.
[0153] In this implementation, the first guide bar 211 and the electronic device 11 on the mainboard 10 are located on the same side of the mainboard 10. Since the processor 115, memory 114 and other devices on the mainboard 10 are relatively high, in order to avoid these relatively high electronic devices 11, the first guide bar 211 can be arranged in any one or more of the following three solutions:
[0154] (1) The first guide bar 211 can be bent to change its trajectory, thereby avoiding these taller electronic devices 11 .
[0155] (2) A plurality of first guide bars 211 may be provided. Each section is individually connected to the power module 30 or the converging device 40. For example, Fig.11 In the example, for a computing unit 112 formed by a processor 115, a memory 114, and a heat sink 113, two first guide bars 211 are arranged on both sides of the computing unit 112, and the two first guide bars 211 can be connected to the same or different power modules 30 or current converging devices 40, respectively.
[0156] (3) A plurality of first current guide bars 211 may be provided. Each current guide bar 21 is fixedly connected to the main board 10, and adjacent current guide bars 21 may be electrically connected through the current-carrying layer 13 in the main board 10 to ensure continuous transmission of the power supply current.
[0157] (4) Please refer to Fig.12 By raising the height of the first guide bar 211, the height of the first guide bar 211 is higher than the electronic devices on the main board 10 that need to be avoided. Then, the second connecting bar 222 is provided to connect between the first guide bar 211 and the main board 10. In this implementation, the first connecting bar 221 may not be provided, and the first guide bar 211 is directly connected to the second main board 10b.
[0158] See also Fig.13 In another implementation, the first guide bar 211 may also be disposed on the lower side of the first main board 10a, and in this case, the first connecting bar 221 may be fixedly connected to the second main board 10b by penetrating the first main board 10a.
[0159] See also Fig.14 The first connecting strip 221 may also be connected to the second main board 10b indirectly, for example, by providing a second guide strip 212 to connect to the second main board 10b, wherein the second guide strip 212 extends substantially along the surface of the second main board 10b.
[0160] It should be understood that the number of the first guide bar 211, the second guide bar 212, and the first connecting bar 221 can be set to one or more. In addition, by setting the first connecting bar 221 to penetrate the second main board 10b or adding connecting bars, more layers of main boards 10 can be stacked on the second main board 10b, thereby meeting the power supply requirements and improving the computing power density.
[0161] In other implementations, the second mainboard 10b may also be other types of circuit boards in the computing node 100, such as a riser card.
[0162] In this embodiment, please refer to the first embodiment for the structural description of components such as the power module 30, the confluence device 40, the flow guide structure 20 and the mainboard 10, and no further details will be given here.
[0163] Third embodiment
[0164] This embodiment can be applied to computing devices such as clusters or supercomputers. The computing node 100 includes multiple mainboards 10, and two mainboards 10 in the multiple mainboards 10 are arranged in a "back-to-back" structure. Each mainboard includes multiple computing units 112. The computing node 100 may include one or more "back-to-back" mainboard 10 arrangements.
[0165] Specifically, the two main boards 10 in a "back-to-back" structure include a third main board 10c and a fourth main board 10d. The third main board 10c and the fourth main board 10d may have different areas and shapes. The first surface 121 of the third main board 10c and the first surface 121 of the fourth main board 10d face two different directions respectively.
[0166] See also Fig.15 , Fig.15 1 is a side view of the structure of two main boards arranged back to back according to an embodiment of the present application. Fig.15 The orientation shown in is taken as an example. The guide structure 20 is sandwiched between the third main board 10c and the fourth main board 10d, and the guide structure 20 is fixedly connected to the third main board 10c and the fourth main board 10d respectively. The guide structure 20 supplies power to the third main board 10c by conducting current upwards, and supplies power to the fourth main board 10d by conducting current downwards.
[0167] In an implementation, when the distance between the third main board 10 c and the fourth main board 10 d is small, the guide structure 20 may only include the third guide bar 213 .
[0168] See also Fig.16 , Fig.16 It is a side view of another computing node under the architecture of the back-to-back arrangement of two main boards shown in the embodiment of the present application. In another implementation, when the spacing between the third main board 10c and the fourth main board 10d is large. The guide structure 20 may also include a third connecting bar 223 and a fourth connecting bar 224. The third connecting bar 223 is fixedly connected between the third guide bar 213 and the third main board 10c, and is used to transmit the current on the third guide bar 213 to the third main board 10c. The fourth connecting bar 224 is fixedly connected between the third guide bar 213 and the fourth main board 10d, and is used to transmit the current on the third guide bar 213 to the fourth main board 10d. The number of the third guide bar 213, the third connecting bar 223 and the fourth connecting bar 223 can be one or more.
[0169] It should be understood that the third connecting bar 223 and the fourth connecting bar 224 are not necessary and can be reasonably arranged according to the distance between the third main board 10c and the fourth main board 10d.
[0170] In this embodiment, by arranging the mainboards 10 back to back and providing the guide structure 20 , the power supply density is doubled and the PUE (Power Usage Effectiveness) is reduced.
[0171] In this embodiment, please refer to the first embodiment for the structural description of components such as the power module 30, the confluence device 40, the flow guide structure 20 and the mainboard 10, and no further details will be given here.
[0172] Fourth embodiment
[0173] See also Fig.17 , Fig.17 It is a side view of a computing node 100 according to an architecture in which multiple motherboards are arranged in the same direction as shown in an embodiment of the present application. In this embodiment, the computing node 100 includes multiple motherboards 10, and the multiple motherboards 10 are arranged side by side in the same direction, and the first surfaces 121 of the multiple motherboards 10 are all facing the same direction, and there may be a gap between two adjacent motherboards 10, or they may be arranged in contact. Specifically, the multiple motherboards 10 are arranged in sequence in the horizontal direction, and the first surfaces 121 of the motherboards 10 are all facing upward.
[0174] Here, Fig.17Taking the orientation in FIG. 2 as an example, the guide structure 20 can be arranged below the plurality of mainboards 10 and extend along the arrangement direction of the plurality of mainboards 10, so as to realize simultaneous power supply to the plurality of mainboards 10. Specifically, the guide structure 20 can include a fourth guide bar 214, which extends substantially along the arrangement direction of the plurality of mainboards 10 and can be directly or indirectly connected to each mainboard 10 through a connecting bar, and power is supplied to each mainboard 10 by transmitting current upward.
[0175] The first embodiment, the second embodiment, the third embodiment, and the fourth embodiment described above record some possible arrangements of the plurality of motherboards 10, and the corresponding arrangement of the flow guide structure 20. It should be understood that as the computing power of the computing node 100 continues to increase, a computing node 100 will include a plurality of motherboards 10. Limited by the space conditions in the computing node 100, the arrangement of the plurality of motherboards 10 can be diverse. Since the flow capacity and the extended shape of the flow guide structure 20 can be customized, the flow guide structure 20 can be provided to realize simultaneous power supply for more than two motherboards 10, thereby improving the power supply density and meeting the power supply requirements of the high-power devices on the motherboard 10, thereby improving the working stability of the computing node 100.
[0176] Fifth embodiment
[0177] In this embodiment, an immersion liquid cooling computing device 200 is proposed. Immersion liquid cooling can be a direct contact type of liquid cooling, which is to immerse the heat-generating electronic device 11 in the cooling liquid to achieve cooling of the electronic device 11. In immersion liquid cooling, since the heat-generating element is in contact with the cooling liquid, the heat dissipation efficiency is higher than that of traditional heat dissipation methods, such as air cooling and water cooling; and the noise is lower than that of cold plate or spray liquid cooling.
[0178] See also Fig.18 , Fig.18 1 is a schematic diagram of a connection structure of a single-phase immersion liquid-cooled computing device and an external cooling device according to an embodiment of the present application. In this embodiment, the computing device 200 includes an immersion cabinet 210 and a computing node 100 disposed in the immersion cabinet 210. In some other embodiments, the computing device 200 may also include a cooling device 220 disposed outside the immersion cabinet 210. For the embodiments of the computing node 100, please refer to the first, second, third, and fourth embodiments described above.
[0179] As mentioned in the above embodiment of the computing node 100, the computing node 100 includes a mainboard 10, a flow guiding structure 20, and a flow converging device 40. For high-density, cluster, supercomputing and other types of computing nodes 100, the mainboard 10 may include multiple computing units 112, and the computing unit 112 includes a processor 115, a memory 114, and may also include a heat sink 113 packaged together with the processor 115; the multiple computing units 112 are arranged in an array on the mainboard 10.
[0180] The arrangement of the motherboard 10 in the immersion cabinet 210 can be arranged in a "back-to-back" manner as described in the above-mentioned embodiment of the computing node 100. Two motherboards 10 form a group of computing nodes 100 by the "back-to-back" arrangement, and each group of computing nodes 100 shares a guide structure 20. Multiple groups of computing nodes 100 can be arranged in the immersion cabinet 210, and the multiple groups of computing nodes 100 are arranged roughly in parallel. The confluence device 40 can be arranged in a strip shape, extending roughly along the arrangement direction of the multiple groups of computing nodes 100. The guide structure 20 in each group of computing nodes 100 is fixedly connected to the confluence device 40.
[0181] In some other implementations, the mainboards 10 in the immersion cabinet 210 may be arranged in a manner such as “stacking in the same direction” in the above-mentioned embodiment, and multiple mainboards 10 may be arranged in the same direction.
[0182] The power supply can be arranged outside the immersion cabinet 210, and a power input port 2101 is provided on the immersion cabinet 210; the power module 30 is connected to the flow guiding structure 20 through a power line; the power input port 2101 allows the power line to pass into the immersion cabinet 210. The power module 30 is connected to the confluence device 40. The power line and the power input port 2101 need to be sealed to prevent the coolant from overflowing.
[0183] After the coolant absorbs the heat of the electronic components 11 on the mainboard 10 in the immersion cabinet 210 , it flows out through the liquid outlet 2102 , and after being cooled by the external cooling device 220 , it flows back into the immersion cabinet 210 through the liquid inlet 2103 in the immersion cabinet 210 .
[0184] Please continue reading Fig.18 For single-phase immersion liquid cooling, the cooling device 220 disposed outside the immersion cabinet 210 may include a circulation pump, a heat exchanger, and a circulation pipeline. The coolant in the immersion cabinet 210 may be an electronic fluoride liquid, but the electronic fluoride liquid remains in a liquid state. The mainboard 10 is directly immersed in the electronic fluoride liquid, and heat is transferred from the electronic device 11 on the mainboard 10 to the coolant.
[0185] The cooling device 220 may include a circulating pump and a heat exchanger. The liquid outlet 2103 on the cabinet of the immersion cabinet 210, the circulating pump, the heat exchanger, the liquid inlet 2102, and the pipeline form a circulation loop. The electronic fluoride liquid with a higher temperature flows out from the liquid outlet 2103, flows in the circulation loop under the power of the circulating pump, and exchanges heat with the cooling water in the heat exchanger to form an electronic fluoride liquid with a lower temperature. The electronic fluoride liquid with a lower temperature then flows back to the immersion cabinet 210 through the liquid inlet 2102. The circulating cooling water in this heat exchanger can be provided by a cooling water tower and / or a dry cooler.
[0186] The immersion liquid-cooled computing device 200 in this embodiment breaks through the bottleneck of traditional on-board power supply based on the efficient heat dissipation capability of liquid cooling and the high power supply density power supply architecture of the embodiment of the present application, thereby allowing at least four processors 115 and associated sockets, radiators 113 and memory 114 to be set on the motherboard 10, thereby achieving an increase in computing power density while ensuring the power supply capability and power supply stability for high-power devices on the motherboard 10.
[0187] See also Fig.19 , Fig.19 It is a schematic diagram of a two-phase immersion liquid cooling computing device according to an embodiment of the present application. For two-phase immersion liquid cooling, the liquid inlet 2103 and the liquid outlet 2102 can be the same. The coolant in the immersion cabinet 210 can be an electronic fluoride liquid. The cooling device 220 includes a condensing coil and a dry cooler arranged outside the immersion cabinet 210. The electronic device 11 on the mainboard 10 is directly immersed in the electronic fluoride liquid, and the heat is transferred from the electrical device 11 to the electronic fluoride liquid, causing the electronic fluoride liquid to boil and generate steam. The heat transfer efficiency is greatly improved through the boiling and condensation process of the electronic fluoride liquid. The steam flows out from the liquid outlet 2102 of the immersion cabinet 210 to the condensing coil, and heat is exchanged with the circulating cooling water provided by the dry cooler in the condensing coil. After the steam is condensed and converted into a liquid state, it still flows back to the immersion cabinet 210 through the liquid outlet 2102.
[0188] Although the present invention has been described with reference to several typical embodiments, it should be understood that the terms used are illustrative and exemplary, rather than restrictive. Since the present invention can be embodied in a variety of forms without departing from the spirit or essence of the invention, it should be understood that the above embodiments are not limited to any of the foregoing details, but should be interpreted broadly within the spirit and scope defined by the appended claims, so all changes and modifications falling within the scope of the claims or their equivalents should be covered by the appended claims.
Claims
1. A computing node, It is characterized in that include: A mainboard, wherein electronic components are mounted on a first surface of the mainboard; A through hole is opened along the thickness direction of the main board, and the hole wall of the through hole has a conductive medium; The through hole is electrically connected to the electronic device; A flow guiding structure, the flow guiding structure is connected to a power supply and is used to transmit the power supply current output from the power supply; the flow guiding structure is arranged on one side of the main board and is electrically connected to the hole wall of the through hole, so as to transmit the power supply current to the through hole along the thickness direction of the main board; There are multiple main boards, and the multiple main boards include a first main board and a second main board; The second main board is located on one side of the first main board, and the first surface of the first main board and the first surface of the second main board face the same direction; The guide structure includes a first guide bar and a first connecting bar, wherein the first guide bar is arranged on one side of the first main board and is fixedly connected to the through hole of the first main board; One end of the first connecting strip is fixedly connected to the first guide strip, and the other end of the first connecting strip is connected to the flow hole of the second main board; or There are multiple main boards, and the multiple main boards include a third main board and a fourth main board; The third main board is disposed adjacent to the fourth main board, and the first surface of the third main board and the first surface of the fourth main board face opposite sides respectively; The flow guide structure is sandwiched between the third main board and the fourth main board, and is fixedly connected to the flow holes of the third main board and the flow holes of the fourth main board respectively.
2. The computing node according to claim 1, It is characterized in that The main board includes a current-carrying layer extending in a transverse direction of the main board; the current-carrying layer is electrically connected to the hole wall of the through hole; The through hole is electrically connected to the electronic device, comprising: the through hole is electrically connected to at least one of the electronic devices through the current-carrying layer.
3. The computing node according to claim 2, It is characterized in that There are a plurality of through-holes, wherein the plurality of through-holes include a first through-hole, and the first through-hole penetrates the main board along the thickness direction of the main board; The hole wall of the first flow-through hole contacts the current-carrying layer for achieving electrical connection.
4. The computing node according to any one of claims 1 to 3, It is characterized in that The electronic device includes a processor module; there are a plurality of through-holes, and the plurality of through-holes include a second through-hole; the mainboard also includes a high-conductivity portion; The second through-hole extends along the thickness direction of the main board, and the first opening of the second through-hole is connected to the flow guiding structure; the second opening of the second through-hole is connected to the high-conductivity part; The high-conductivity portion extends from the second opening along the thickness direction of the mainboard to the first surface of the mainboard and contacts the power supply terminal of the processor module.
5. The computing node according to claim 4, It is characterized in that The cross-sectional area of the high-conductivity portion is greater than or equal to the cross-sectional area of the second through-flow hole.
6. The computing node according to claim 4, It is characterized in that The cross-sectional area of the high-conductivity portion is gradually widened from the second opening of the second through-flow hole toward the first surface of the main board.
7. The computing node according to claim 4, It is characterized in that The second opening of the second through-hole and the high-conductivity portion are connected at a first position of the current-carrying layer; The first position of the current-carrying layer is also electrically connected to at least one through hole.
8. The computing node according to any one of claims 2 to 7, It is characterized in that The current-carrying layer includes a plurality of first conductive layers, wherein the first conductive layers are used to conduct the power supply current; A plurality of the first conductive layers are arranged along the transverse direction of the main board, and the plurality of the first conductive layers are arranged at intervals; each of the first conductive layers is electrically connected to at least one of the through holes.
9. The computing node according to claim 2, It is characterized in that The current-carrying layer includes a first conductive layer, the area of the first conductive layer is smaller than the cross-sectional area of the main board; The first conductive layer is electrically connected to the through hole.
10. The computing node according to claim 8 or 9, It is characterized in that The current-carrying layer comprises a first layer and a second layer, and the first layer and the second layer are arranged in sequence along the thickness direction of the main board; The first layer includes the first conductive layer, and the second layer includes the second conductive layer; the first conductive layer and the second conductive layer are both electrically connected to the through hole; The thickness of the second conductive layer is less than or equal to the thickness of the first conductive layer.
11. The computing node according to claim 10, It is characterized in that There are one or more second conductive layers, and the orthographic projection of at least one of the second conductive layers on the first layer at least partially overlaps with the first conductive layer; and the second conductive layer with which the orthographic projection partially overlaps is electrically connected to the first conductive layer.
12. The computing node according to any one of claims 8 to 11, It is characterized in that The mainboard also includes a signal layer, and the current-carrying layer and the signal layer are arranged on the same layer; The signal layer includes one or more third conductive layers, and the third conductive layer is arranged on one side of the first conductive layer or between two adjacent first conductive layers.
13. The computing node according to claim 1, It is characterized in that The mainboard includes a plurality of computing units, each of which includes a processor, a memory bar arranged on one side of the processor, and a heat sink; The computing units are arranged in an array or in sequence on the mainboard.
14. A computing device, It is characterized in that It comprises a shell, and a computing node as described in any one of claims 1 to 13; the computing node is installed in the shell.
15. The computing device of claim 14, It is characterized in that The computing device further comprises an immersion cabinet and an electric energy input port, a liquid outlet port and a liquid inlet port provided on the immersion cabinet; the immersion cabinet is filled with cooling liquid, and the computing node is immersed in the cooling liquid; The power supply is arranged outside the immersion cabinet; the power supply is connected to the flow-guiding structure via a power line; the power input port is used for the power line to pass into the immersion cabinet; The liquid outlet is used for gaseous or liquid cooling liquid to flow out, and the cooled cooling liquid flows back into the immersion cabinet through the liquid inlet.
Citation Information
Patent Citations
Computer mainboard
CN102759974A
Power supply device and single board
CN111065199A
Cabinet, immersion liquid cooling system and immersion liquid cooling method
CN113301764A
Circuit board and manufacturing method thereof
CN114916127A
Circuit board capable of loading large current
CN201550356U
