Computing power server
By fixing the graphics card vertically on the motherboard and adopting an L-shaped chassis stacking design, the problems of low heat dissipation efficiency and space utilization in traditional server layouts are solved, achieving efficient space utilization and heat dissipation effects, and reducing computer room and production costs.
Patent Information
- Application Number
- CN202521807714.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Utility models(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-25
- Publication Date
- 2025-10-10
- Estimated Expiration
- 2035-08-25
AI Technical Summary
The traditional server layout has low heat dissipation efficiency and space utilization, which makes it difficult to meet the compactness and efficiency requirements of high-density data centers.
The computing server adopts an L-shaped structure design, with the graphics card fixed vertically on the upper surface of the motherboard, and the motherboard set horizontally at the bottom of the chassis. The chassis is designed to be L-shaped, and two chassis are stacked to form an integrated unit, and the heat dissipation effect is improved through the side air intake and rear air outlet cooling method.
It significantly reduces the overall width of the server module, increases the number of servers accommodated, simplifies wiring and maintenance processes, reduces computer room and production costs, and improves heat dissipation efficiency and operational efficiency.
Smart Images

Figure CN223427085U_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The utility model relates to server equipment technical field, concretely is a kind of computing power server. BACKGROUND
[0002] The server structure usually adopted modular design on the market, and the components such as graphic card, mainboard, radiator are placed in parallel in the same plane or stacked up and down, forming a square shape. In the operation and management of computer room, these servers are usually placed on the rack, and then connected by network cable to communicate with each other, and multiple racks are placed to form a server cluster in the computer room.
[0003] Based on the existing 4U server, its height is fixed, so it is not suitable to accommodate too many servers in the height space of computer room. On the other hand, the design of 4U height is conducive to the smooth development of operation and debugging. ("4U" represents the height unit of the case, 1U is equal to 1.75 inches (44.45 mm), that is, the height of 4U server case is 4U x 1.75 inches = 7 inches (177.8 mm). However, in the limited space of computer room, it is extremely necessary to arrange more computing power servers which are more convenient for programmers to operate for commercial application market, especially at the current time of AI computing power rising, and more edge computing nodes are particularly necessary for computing speed.
[0004] The heat dissipation efficiency and space utilization of the traditional layout server are low, which is difficult to meet the needs of compactness and high efficiency of high-density data center. Therefore, it is urgent to provide a server with improved function, which can store more servers in limited space without changing the traditional 4U height and convenient operation. UTILITY MODEL CONTENT
[0005] The utility model aims at providing a kind of computing power server to solve the problem of low heat dissipation efficiency and space utilization of traditional parallel layout in the background art, which is difficult to meet the needs of compactness and high efficiency of high-density data center.
[0006] To achieve the above purpose, the utility model provides the following technical scheme: a kind of computing power server, including mainboard, graphic card and case, the mainboard is horizontally arranged at the bottom of case, the graphic card is vertically fixed on the upper surface of mainboard, the plane of graphic card and the plane of mainboard are perpendicular, the case is L-shaped structure, the bottom of L-shaped structure contains mainboard, the vertical part of L-shaped structure contains graphic card, two cases are stacked to form an integrated unit.
[0007] Preferably, a slot is provided on the upper part of the outer wall of the mainboard and the connection part of the graphic card.
[0008] Preferably, the bottom and vertical portion of the L-shaped structure of the chassis are connected by integral molding, and the width L1 of the bottom of the L-shaped structure of the chassis is greater than the height L2 of the vertical portion of the L-shaped structure of the chassis.
[0009] Preferably, the graphics card is provided with a first heat sink perpendicular to the plane of the motherboard and parallel to the motherboard, and a gap is maintained between the bottom of the first heat sink and the upper surface of the motherboard.
[0010] Preferably, a circuit board is fixed to the upper surface of the mainboard, the circuit board is arranged parallel to the graphics card, and the circuit board is arranged perpendicular to the mainboard.
[0011] Preferably, a cabinet is further included, the height of the vertical part of the L-shaped structure of the chassis is not higher than the height of a single row of stacking space in the cabinet, and the bottom depth L3 of the L-shaped structure of the chassis is adapted to the depth of the cabinet.
[0012] Preferably, a second radiator is provided on the rear inner wall of the chassis, and heat dissipation holes are opened on the left and right inner walls of the chassis.
[0013] Compared with the prior art, the beneficial effects of the present invention are:
[0014] 1. By changing the traditional parallel stacking arrangement of graphics cards to a vertical arrangement on the motherboard, an L-shaped chassis structure is formed. The stacking method significantly reduces the overall width of the server module. Compared with the traditional square module layout, this structure can accommodate more server modules and reduce computer room costs.
[0015] 2. The stacking design of the L-shaped modules and the external communication line interface simplify the wiring, debugging and maintenance processes, improve the programmer's operating efficiency, reduce the cabinet height requirement and simplify the part design, reduce the cabinet and production costs, and the stacking structure facilitates modular production and assembly. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 It is a schematic diagram of the three-dimensional structure of the utility model;
[0017] Figure 2 This is a schematic diagram of the internal structure of the chassis in the present utility model;
[0018] Figure 3 This is a schematic diagram of the structure in which the motherboard and the graphics card are separated in the present invention;
[0019] Figure 4 This is a schematic diagram of the three-dimensional structure of the chassis in the utility model;
[0020] Figure 5 It is a schematic diagram of the three-dimensional structure of the unit in this utility model.
[0021] In the figure: 1. Motherboard; 2. Graphics card; 3. Chassis; 4. Cabinet; 5. First radiator; 6. Slot; 7. Circuit board; 8. Unit; 9. Second radiator; 10. Heat dissipation holes. DETAILED DESCRIPTION
[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0023] The "4U" listed in this embodiment represents the height unit of the chassis 3. 1U is equal to 1.75 inches (44.45 mm), that is, the height of the 4U server chassis 3 is 4U×1.75 inches=7 inches (177.8 mm). According to the common industry name, 4U2 represents two computing nodes, 4U4 represents four computing nodes, and one computing node usually represents one server. Figure 1 As shown, the exemplary listed rack includes four layers of servers.
[0024] See also Figure 2 This embodiment provides a technical solution: a computing server comprising a motherboard 1, a graphics card 2, and a chassis 3. The motherboard 1 is horizontally arranged at the bottom of the chassis 3, and the graphics card 2 is vertically fixed to the upper surface of the motherboard 1. The plane of the graphics card 2 is perpendicular to the plane of the motherboard 1. The chassis 3 has an L-shaped structure. The bottom of the L-shaped structure accommodates the motherboard 1, and the vertical portion of the L-shaped structure accommodates the graphics card 2. When two chassis 3 are stacked, they form an integrated unit 8. Specifically, when two chassis 3 are stacked, they form an integrated unit 8. Other stacking methods, such as steps, are also possible, and are not specifically limited here. Thus, the improvement idea for a single node is to update the horizontally arranged graphics card 2 to a vertical arrangement. After stacking in the height direction to form a unit, it still meets the 4U requirement. Since the graphics card does not occupy space separately in the horizontal direction, the single node saves space in the horizontal direction without affecting the layout requirements of other electronic components on the motherboard.
[0025] Further explanation: In existing designs, graphics card 2, which is almost universally adopted worldwide using NVIDIA's standard packaged module, such as the GeForce RTX 5060, is approximately 350mm long, 100mm wide, and 50mm thick. This makes it impossible to stack directly horizontally on motherboard 1. For example, the packaged graphics card with fan is over 50mm thick, and the thickness of the motherboard with components is only less than 100mm under extreme conditions. Considering that the motherboard requires more space to integrate other electronic components when stacking, the physical conditions for double-layer stacking are not met in any way. In existing technology, to accommodate a 4U rack height, the two need to be arranged side by side at the bottom of chassis 3.
[0026] like Figure 3 and Figure 5 As shown, two chassis 3 are stacked into a unit 8, and the outer wall of the unit 8 is slidably connected to the cabinet 4. By setting up a stacking structure, the unit 8 and the cabinet 4 can be slid and pulled, which is convenient for actual inspection and maintenance. The graphics card 2 is vertically set on the motherboard 1 to form an L-shaped chassis 3 structure, and a stacking method is adopted to significantly reduce the overall width of the server module. Compared with the traditional square module layout, this structure can accommodate more server modules at the same cabinet 4 height, reducing the deployment cost of the computer room.
[0027] In the assembly of the chassis, the two L-shaped structures are slidably stacked on each other, such as Figure 4 and Figure 5 As shown, one to two slides (or steps, etc.) are installed on the outer wall of the L-shaped chassis. The slides are oriented parallel to the direction of extension, that is, parallel to the front-to-back movement. This allows for the extension and adjustment of a single L-shaped chassis on the rack. For example, if the top chassis is pulled out separately, the L-shaped chassis have inner slides, so when the two chassis are stacked, they engage with each other to form an integrated unit. This allows them to be transported as a single unit and then pulled apart again when needed.
[0028] See also Figure 1 、 Figure 2 、 Figure 3 、 Figure 4 and Figure 5A slot 6 is provided at the upper portion of the outer wall of the motherboard 1 where it connects to the graphics card 2. The bottom and vertical portions of the L-shaped structure of the chassis 3 are integrally connected. The width L1 of the bottom portion of the L-shaped structure of the chassis 3 is greater than the height L2 of the vertical portion of the L-shaped structure of the chassis 3. A first heat sink 5 is provided on the graphics card 2, perpendicular to the plane of the motherboard 1, with a gap between the bottom portion of the first heat sink 5 and the upper surface of the motherboard 1. A circuit board 7 is fixed to the upper surface of the motherboard 1, parallel to the graphics card 2 and perpendicular to the motherboard 1. The chassis 3 also includes a cabinet 4. The vertical portion of the L-shaped structure of the chassis 3 is no higher than the height of a single-row stacking space within the cabinet 4. The depth L3 of the bottom portion of the L-shaped structure of the chassis 3 matches the depth of the cabinet 4. A second heat sink 9 is provided on the rear inner wall of the chassis 3. Multiple cooling holes 10 are provided on the left and right inner walls of the chassis 3, with only one shown in the figure. In traditional cooling methods, due to the parallel arrangement of the graphics cards 2, the cooling fan can only reach the upper and lower surfaces of the server, resulting in poor cooling when multiple servers are stacked one above the other. In this embodiment, rear and side heat dissipation is adopted to improve the heat dissipation effect of the server.
[0029] In terms of specific heat dissipation methods, one implementation utilizes side-intake ventilation. Cool air flows through the sides of the chassis through the heat dissipation holes, then exits the chassis through the cooling fans, creating a continuous cooling cycle. Because the front, back, and sides are hollowed out and have external insulation, heat can be directly dissipated from the front and back, resulting in superior cooling compared to traditional top and bottom cooling methods.
[0030] As for the specific heat dissipation method, in one embodiment, when rear air intake is adopted, air is stirred in through the rear cooling fan, and then the hot air is discharged from the sides and front, and a cycle is formed to form air cooling.
[0031] By setting up the stacking design of L-shaped modules and the external communication line interface, the wiring, debugging and maintenance processes are simplified, the programmer's operating efficiency is improved, and the cabinet and production costs are reduced by reducing the cabinet 4 height requirement and simplifying the part design. The stacking structure facilitates modular production and assembly.
[0032] Working principle: The motherboard 1 is fixed horizontally to the bottom of the chassis 3, which serves as the basic platform of the server module. The graphics card 2 is fixed to the upper surface of the motherboard 1 through a vertical slot 6. Its plane is perpendicular to the plane of the motherboard 1. The slot 6 adopts a standard interface to ensure that the graphics card 2 is firmly connected and electrically connected to the motherboard 1. The chassis 3 is designed as an L-shaped structure. The bottom accommodates the motherboard 1 and the vertical part accommodates the graphics card 2. The bottom width of the L-shaped chassis 3 is greater than the width of the vertical part to adapt to the size of the motherboard 1, and the height of the vertical part adapts to the height requirements of the graphics card 2 and the first radiator 5. The first radiator 5 is installed on the side wall of the graphics card 2, parallel to the plane of the graphics card 2, and the bottom is protected from the motherboard 1. Maintain a gap to avoid contact with other components. The circuit board 7 is fixed on the motherboard 1 and is arranged parallel to the graphics card 2. When multiple L-shaped chassis 3 are arranged in the server cabinet 4, they are stacked. The vertical part of one chassis 3 is close to the inner side of the bottom of the adjacent chassis 3, forming a compact stacking structure. The vertical parts and the bottom of the adjacent chassis 3 partially overlap in the height direction to reduce the overall height. The bottom depth of the chassis 3 is adapted to the depth of the cabinet 4, and the height of the vertical part is controlled within the single-row space requirement of the cabinet 4. Multiple chassis 3 are connected through communication lines and are uniformly managed by the cabinet 4 management system to facilitate wiring and debugging. The above is the working process of the entire device.
[0033] Although the embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A computing power server, characterized in that: The invention comprises a motherboard (1), a graphics card (2) and a chassis (3), wherein the motherboard (1) is horizontally arranged at the bottom of the chassis (3), the graphics card (2) is vertically fixed to the upper surface of the motherboard (1), and the plane of the graphics card (2) is perpendicular to the plane of the motherboard (1). The chassis (3) is an L-shaped structure, wherein the bottom of the L-shaped structure accommodates the motherboard (1), and the vertical part of the L-shaped structure accommodates the graphics card (2). When two chassis (3) are stacked, an integrated unit (8) is formed.
2. The computing power server according to claim 1, wherein: A slot (6) is provided at the connection between the upper portion of the outer wall of the mainboard (1) and the graphics card (2).
3. The computing power server according to claim 1, wherein: The bottom and vertical portion of the L-shaped structure of the chassis (3) are connected by integral molding, and the width L1 of the bottom of the L-shaped structure of the chassis (3) is greater than the height L2 of the vertical portion of the L-shaped structure of the chassis (3).
4. The computing power server according to claim 1, wherein: The graphics card (2) is provided with a first heat sink (5) perpendicular to the plane of the motherboard (1) and parallel to the motherboard, and a gap is maintained between the bottom of the first heat sink (5) and the upper surface of the motherboard (1).
5. The computing power server according to claim 1, wherein: A circuit board (7) is fixed to the upper surface of the mainboard (1); the circuit board (7) and the graphics card (2) are arranged in parallel; and the circuit board (7) and the mainboard (1) are arranged in perpendicular relation.
6. The computing power server according to claim 1, wherein: It also includes a cabinet (4), wherein the height of the vertical portion of the L-shaped structure of the chassis (3) is not higher than the height of a single-row stacking space in the cabinet (4), and the bottom depth L3 of the L-shaped structure of the chassis (3) is adapted to the depth of the cabinet (4).
7. The computing power server according to claim 1, wherein: A second radiator (9) is provided on the rear inner wall of the chassis (3), and heat dissipation holes (10) are provided on the left and right inner walls of the chassis (3).