Hybrid cooling device for accelerated hardware

Abstract

This document describes a hybrid cooling device and method that combines phase change cooling and air cooling. The hybrid cooling device includes a closed-loop two-phase system, one or more fans, and a mounting clamp. The two-phase system also includes a cooling plate, integrated channels, a heat sink, and a pressure sensor. The cooling plate may include a phase change fluid for extracting heat from electronics on a printed circuit board clamped between the cooling plate and the mounting clamp. One or more fans are used to generate airflow to cool the cooling plate and the heat sink. The pressure sensor is used to control the operation of the hybrid cooling device, which can be deployed in various system environments and server configurations.

Description

Technical Field

[0001] The embodiments of this disclosure generally relate to cooling systems. More specifically, the embodiments of this disclosure relate to hybrid cooling apparatuses and hybrid cooling methods that simultaneously utilize phase change cooling and air cooling. Background Technology

[0002] High-power-density devices are computing devices that encapsulate high-performance processors (such as GPUs, ASICs, heterogeneous computing-based IC chips, or chiplets). These devices are increasingly popular due to persistently high computing demands. High-power-density devices often generate significant heat and are typically integrated into server chassis. Therefore, a suitable thermal environment for servers, racks, and data center facilities must be provided for high-power-density devices to function properly.

[0003] While liquid cooling is a promising cooling solution for high power density devices, especially when the power budget for a single chip exceeds a threshold (e.g., 400W), the required infrastructure can become a bottleneck. This is because such liquid cooling solutions have requirements for supply inlet temperatures, flow rates, and pressures that exceed the capabilities of typical data centers. Even if suitable data center infrastructure could be developed, the cost would be prohibitively high.

[0004] To complicate matters further, many high-performance hardware components are connected via the PCIe (Peripheral Component Interconnect Standard) expansion bus. Liquid cooling solutions for such hardware components and packages require a completely different architecture compared to connector-based mezzanine cards.

[0005] Previously, PCIe-based electronic device cooling solutions primarily focused on desktop products, rather than hyperscale cloud data centers. Therefore, these cooling solutions may not be easily integrated into cloud data center servers. They may also be scalable, lack versatility, be unreliable, or prohibitively expensive. Furthermore, most solutions are based on air cooling, which may not meet the demands of ever-increasing power densities. Summary of the Invention

[0006] A hybrid cooling device includes: a phase change system including a cooling plate, a heat sink, and an integrated channel connecting the cooling plate and the heat sink; a mounting clip for positioning electronic hardware to be cooled between the mounting clip and the cooling plate, wherein the heat sink is located above the cooling plate, and wherein the cooling plate is vertically positioned to attach to the electronic hardware when the mounting clip is clamped onto the cooling plate; and one or more fans for providing air cooling to the heat sink and the electronic hardware, wherein the electronic hardware includes a printed circuit board and electronic devices packaged thereon, the electronic devices including one or more chips or power electronic devices, and wherein the phase change system, the mounting clip, and one or more fans, together with the electronic devices packaged on the printed circuit board, are capable of being inserted into a peripheral bus as an integrated peripheral device.

[0007] According to some embodiments, the hybrid cooling device further includes a device frame, wherein the radiator, integrated channel, and cooling plate are attached to the device frame.

[0008] According to some embodiments, the hybrid cooling device further includes: an adapter reinforcement located between the cooling plate and the electronic hardware; one or more resilient channels; wherein the adapter reinforcement works in conjunction with one or more resilient channels to maintain appropriate pressure on the electronic hardware.

[0009] According to some embodiments, the hybrid cooling device further includes: a movable shaft located in one of the elastic channels; wherein one end of the mounting clamp is inserted into the elastic channel via the movable shaft, such that the one end of the mounting clamp is movable in the elastic channel.

[0010] According to some embodiments, the elastic channel provides force on the moving axis on both sides in the horizontal direction to properly secure the electronic hardware within the hybrid cooling device.

[0011] According to some embodiments, the integrated channel includes a steam line and a liquid line, the liquid line being used to transfer liquid from the radiator to the cooling plate, and the steam line being used to transfer steam from the cooling plate to the radiator.

[0012] According to some embodiments, each of one or more fans is integrated into a hybrid cooling unit or is a separate fan.

[0013] According to some embodiments, airflow generated by one or more fans passes through a first dedicated channel through a cooling plate and through a second dedicated channel through a heat sink.

[0014] According to some embodiments, the hybrid cooling device further includes: a temperature sensor; a pressure sensor; wherein the temperature sensor and the pressure sensor are used to control the operation of the hybrid cooling device.

[0015] A server chassis includes a hybrid cooling device according to the above embodiments; and a chassis fan for providing airflow to cool the server chassis and the hybrid cooling device.

[0016] An electronic rack includes multiple server chassis according to the above embodiments. Attached Figure Description

[0017] Embodiments of the present invention are illustrated in the accompanying drawings by way of example rather than limitation, wherein the same reference numerals denote the same parts.

[0018] Figure 1A-1B Hardware for a hybrid cooling device according to one embodiment is shown.

[0019] Figure 2A-2B A hardware design with a fan in a hybrid cooling device according to one embodiment is shown.

[0020] Figures 3A-3C A hybrid cooling device, according to one embodiment, is shown that is integrated with electronic devices on a printed circuit board.

[0021] Figures 4A-4C Thermal management inside a hybrid cooling device according to one embodiment is shown.

[0022] Figures 5A-5B This illustrates the use of a hybrid cooling system at the system level in a server, according to one embodiment.

[0023] Figure 6 Figure 2 illustrates a hybrid cooling unit deployed in a server chassis according to one embodiment.

[0024] Figures 7A-7B The operation control of a hybrid cooling device according to one embodiment is shown.

[0025] Figure 8 This is a flowchart illustrating the control flow 800 for a hybrid cooling device according to one embodiment.

[0026] Figure 9 A method 900 for cooling a heterogeneous computing architecture according to one embodiment is shown.

[0027] Figure 10 This is an electronic rack block diagram according to one embodiment. Detailed Implementation

[0028] Various embodiments and aspects of the invention will be described with reference to the details discussed below, and the accompanying drawings will illustrate various embodiments. The following description and drawings are illustrative and should not be construed as limiting. Many details are described in detail to provide a thorough understanding of various embodiments of the invention. In some cases, well-known or conventional details will not be described to provide a brief discussion of embodiments of the invention.

[0029] The reference to "an embodiment" or "embodiment" in this specification means that a specific feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the invention. The phrase "in one embodiment" appearing throughout the specification does not necessarily refer to the same embodiment.

[0030] According to various embodiments, a hybrid cooling apparatus and method using a combination of phase change cooling and air cooling is described herein. The hybrid cooling apparatus includes a closed-loop two-phase system, one or more fans, and a mounting clamp. The two-phase system further includes a cooling plate, integrated channels, and a heat sink serving as a condenser. The cooling plate may include a phase change fluid for extracting heat from electronic components on a printed circuit board (PCB) clamped between the cooling plate and the mounting clamp. One or more fans may be used to generate airflow to cool the electronic components on the PCB and the heat sink. Pressure and temperature sensors may be used to control the operation of the hybrid cooling apparatus, which can be integrated into different system environments and server configurations.

[0031] In one embodiment, the hybrid cooling device further includes a device frame to which the heat sink, integrated channels, and cooling plate are attached. Additionally, the hybrid cooling device may include an adapter reinforcement between the cooling plate and electronics on the PCB, as well as one or more resilient channels. The adapter reinforcement works in conjunction with one or more resilient channels to maintain appropriate pressure on the electronics on the PCB.

[0032] In one embodiment, the hybrid cooling device further includes a movable shaft in one of the resilient channels, one end of the mounting clamp being inserted into the resilient channel via the movable shaft, thereby making that end of the mounting clamp movable on the resilient channel. The resilient channel can provide force on the movable shaft on both sides in a horizontal direction to properly secure the PCB at a specific location within the hybrid cooling device.

[0033] In one embodiment, the electronics on the PCB may include one or more chips or power electronic devices, wherein the PCB on which the electronics are mounted is connected to the server main PCB via a peripheral component high-speed interconnect standard (PCIe) bus.

[0034] In one embodiment, the integrated channel includes a steam line and a liquid line, the liquid line for conveying liquid from the radiator to the cooling plate, and the steam line for conveying steam from the cooling plate to the radiator. In one embodiment, the steam line and the liquid line may be designed with different physical dimensions for better performance.

[0035] In one embodiment, each of the one or more fans may be a fan integrated into the hybrid cooling system or a separate fan. The airflow generated by the one or more fans passes through the PCB via a first dedicated channel and through the heatsink via a second dedicated channel.

[0036] In one embodiment, the hybrid cooling device may include a temperature sensor and a pressure sensor for controlling the operation of the hybrid cooling device. In another embodiment, the hybrid cooling device may include only a pressure sensor, which is pre-integrated into the steam line of the hybrid cooling device.

[0037] In one embodiment, the hybrid cooling method can be applied to different chassis, such as blade servers. Furthermore, multiple electronic components on one or more PCBs can be encapsulated within the hybrid cooling unit. Various clamping methods can be used to hold the PCB.

[0038] Hybrid cooling systems can be deployed in any server or chassis environment, compatible with diverse heterogeneous hardware configurations to handle complex and varied heterogeneous computing workloads. Therefore, hybrid cooling systems are scalable and interoperable with different server system designs and configurations, including various heterogeneous hardware extensions. Furthermore, this solution is highly efficient because the fluid can be self-driven through phase change technology.

[0039] Several embodiments provide solutions for hyperscale data center applications and corresponding servers in cloud environments, as well as edge computing systems in edge clusters or edge devices. The cooling solutions described in these embodiments can be used to cool high-power-density electronic devices. With a complete packaging approach to hybrid cooling device design, the cooling solutions can be configured into different hybrid designs, such as a combination of phase change and air cooling, or phase change liquid cooling alone.

[0040] Figure 1A-1B A hybrid cooling device according to one embodiment is shown. Figure 1A This is a front view of the hybrid cooling system. Figure 1B This is a side view of the hybrid cooling system.

[0041] As shown in the figure, the hybrid cooling device includes a radiator 101, an integrated channel 103, a cooling plate 105, a mounting clamp 107, and a device frame 109. The radiator 101, the integrated channel 103, and the cooling plate 105 can form a single unit. This unit constitutes the main components of the hybrid cooling device.

[0042] Although they are single units, the overall design of the three components 101, 103 and 105 can be different depending on the actual implementation and specific requirements of different users.

[0043] The device frame 109 may be a hardware frame to which the heat sink 101, integrated channel 103, and cooling plate 105 are attached. The integrated channel 103 may include liquid and vapor lines for connecting the heat sink and cooling plate. The mounting clip 107, which can be used to hold electronic components on a printed circuit board (PCB) at appropriate pressure, will be described in detail below.

[0044] Figure 2A-2B A hybrid cooling device according to one embodiment is further illustrated. Figure 2A This is a front view of the hybrid cooling system. Figure 2B This is a side view of the hybrid cooling system.

[0045] As shown in the figure, the hybrid cooling device may include a fan 201. The fan 201, together with the single unit described above, provides a hybrid cooling environment for the printed circuit board (PCB) 203 mounted thereon, which has high power density electronic devices.

[0046] In one embodiment, PCB 203 is an acceleration PCB that includes multiple hardware components to accelerate data communication, storage retrieval, encryption / decryption, mathematical operations, graphics, and web browsing. PCB 203 can be attached to cooling plate 105. Both heat sink 101 and PCB 203 can be air-cooled by fan 201. Figure 2A The scheme shown can be understood as integrating the fan into a single unit. This means that the fan design is optimized in terms of location, fan selection, and airflow control.

[0047] In one embodiment, the hybrid cooling system's layout allows the fan 201 to direct or indirect airflow towards the heatsink 101 and the electronic components on the PCB 203. This allows the fan 201 to provide both direct and indirect air cooling. The fan 201 can be an integrated unit of the hybrid cooling system or a separate module attached to it.

[0048] Figures 3A-3C A hybrid cooling device integrated with a PCB is shown according to one embodiment. Figure 3A The overall structure of the hybrid cooling device is explained. Figure 3B and Figure 3C Additional implementation details are provided.

[0049] exist Figure 3A In this embodiment, different types of chips or power electronic devices can be mounted on PCB 203. An adapter reinforcement 305 can be used between the cooling plate 105 and the chips or power electronic devices on PCB 203 to ensure proper assembly of the hybrid cooling device. In one embodiment, the adapter reinforcement 305 can be elastic, and it can be made of an elastic material to accommodate electronic devices mounted at different heights on PCB 203.

[0050] The hybrid cooling system also includes a connection bus 301 for connecting different electronic devices on PCB 203. Connection bus 301 may be a PCIe (Peripheral Component Interconnect Standard) bus, which is an interface standard for connecting high-speed components.

[0051] exist Figure 3A In this system, the cold section of the hybrid cooling device can cover all electronic components on PCB 203, so that they can all be cooled by the cooling plate 105.

[0052] Figure 3B The assembly clamp 107 is shown in detail. The assembly clamp 107 comprises four parts: a resilient channel 307, a movable shaft 309, and two assembly shafts 311 and 313. In this figure, the assembly clamp 107 is not locked. As shown, one end of the assembly shaft 311 is inserted into the resilient channel 307 via the movable shaft 309, allowing this end to move within the resilient channel 307. The resilient channel 307 can provide force to the movable shaft 309 on both sides in the horizontal direction to ensure proper fixation of the hybrid cooling device relative to the PCB 203 and to ensure proper thermal contact between the electronics and the reinforcement.

[0053] Figure 3C The diagram shows a view of the hybrid cooling device when locked. As shown, in addition to the elastic channel 307, the hybrid cooling device may also include another elastic channel 315 on the assembly shaft 315. The two assembly shafts 311 and 313 are connected together to form the assembly clamp 107.

[0054] The mounting clamp 107 can be locked and unlocked by rotating about the moving axis 309. When the mounting clamp 107 is locked, the PCB 203, the chip 303 (also referred to as an electronic device) on the PCB 203, and the adapter reinforcement 305 can be clamped between the cooling plate 105 and the mounting axis 107. Furthermore, when the mounting clamp 107 is locked, the two resilient channels 307 and 315 ensure that appropriate pressure is applied to the chip 303 and the PCB 203 to avoid damage and prevent them from malfunctioning. The resilient channels 307 and 315 also ensure proper thermal contact between the cooling plate 105 and the chip 303.

[0055] Figures 4A-4C A hybrid cooling device and thermal management within a hybrid environment are illustrated according to one embodiment. Figure 4A This is a front view of the hybrid cooling system. Figure 4B-4C This is a front view of the hybrid cooling system.

[0056] like Figure 4A As shown, the cooling plate 105 and the radiator 101 are connected via a liquid line 403 and a steam line 404, each line being a conduit through which steam, air, or other fluids can pass. The liquid line 403 and the steam line 404 constitute... Figure 2A The described integrated channel 103.

[0057] Figure 4A In this process, a phase change 405 can occur within the cooling plate 105 due to heat extraction from the electronic devices / chips on the PCB 203. Fluid from the heat sink 101 reaches the cooling plate 105 via the liquid line 403, where it absorbs the heat extracted from the chips on the PCB 203 and transforms into vapor 404. The vapor carrying latent heat does not change its temperature due to the phase change. The phase change causes an increase in pressure within the cooling plate 105, which in turn causes the vapor to rise through the vapor line 404 to the heat sink 101.

[0058] The radiator 101 can be used as a condensation unit to condense the steam lifted from the cooling plate 105 into a liquid by extracting its latent heat. The liquid can then return to the cooling plate under the influence of gravity.

[0059] Figure 4B It shows a fan (e.g.) Figure 2A and Figure 2B The fan 201 shown generates airflows 407 and 409. The fan can generate airflows 407 and 409 by blowing or drawing air. The fan can be located on either side of the hybrid cooling unit.

[0060] In one embodiment, airflow 407 may pass through heat sink 101 to assist heat sink 101 in condensing vapor into liquid, and airflow 409 may pass through chips or electronic devices on PCB 203 to provide air cooling to the chips or electronic devices on PCB 203. Figure 4B In the mixed cooling unit, dedicated channels are used to manage and optimize airflow 407 and 409.

[0061] Alternatively, Figure 4C Another design for thermal management is also shown, in which airflow 411 passes in parallel through heat sink 101 and PCB 203 and the electronics on PCB 203, since no dedicated channel for airflow 411 is used.

[0062] Figure 4B and Figure 4C This illustrates different airflow management within a hybrid cooling system with different fan implementations. This can be understood as creating different hybrid cooling environments, and how the entire hybrid cooling system is used and configured. Figure 4B As shown, a portion of the incoming airflow is used to cool the heatsink, causing the vapor to condense back into liquid, while another portion is used to directly cool other air-cooled electronic components on the PCB. Heated air is concentrated into a dedicated channel by a fan drive. However, in Figure 4C In this process, the two parts of the airflow form independent paths.

[0063] Figures 5A-5B The use of a hybrid cooling device according to one embodiment at the system level is illustrated. As shown, the hybrid cooling device 501 can be integrated into a server chassis 507, wherein the hybrid cooling device 501 can adapt to the environment of the server chassis 507 and make full use of the existing server chassis environment and structure.

[0064] exist Figure 5A In this configuration, the hybrid cooling device 501 includes a phase change cooling section occurring in the cooling plate 507 and the dedicated fan 505, used to cool the accelerated PCB 503 and the electronic components mounted thereon. The dedicated fan 505 may be a crossflow fan and may be used to assist in generating the airflow shown in Figure 4.

[0065] As further shown in the figure, the server chassis 507 may include a server PCB 505 and a chassis fan 509 mounted on the right side of the hybrid cooling unit 501. The chassis fan 509, as part of the existing server chassis structure, can serve as a primary airflow booster. Therefore, the hybrid cooling unit 501 can make full use of the existing server chassis structure.

[0066] exist Figure 5BIn this configuration, an additional fan 511 is integrated into the hybrid cooling unit 501 to enhance airflow. The additional fan 511 can be used as redundancy, as the server chassis 507 may not be dedicated to accelerating the PCB 503. The additional fan 511 can further improve system performance.

[0067] Figure 6 Figure 2 illustrates a hybrid cooling unit deployed in a server chassis according to one embodiment.

[0068] In this embodiment, with Figures 5A-5B Unlike the embodiment shown, the hybrid cooling unit 601 is completely separate from the server chassis 507 in terms of airflow management, which means that no server fan is required.

[0069] In the various embodiments described above, Figures 5A-5B and Figure 6 The hybrid cooling unit can be reconfigured by adding additional features to take advantage of the environment in server chassis 507.

[0070] Figures 7A-7B This illustrates how a hybrid cooling device can be controlled according to one embodiment.

[0071] As shown in the figure, the hybrid cooling device may include two sensors. A pressure sensor 701 may be attached to steam line 403 to measure the pressure of steam passing through steam line 404. A temperature sensor 703 may be disposed in the cooling plate to measure the temperature of the cooling plate. These two sensors 701 and 703 are decoupled from any electronics on PCB 503. Decoupling can significantly improve the adaptability and reliability of the cooling solution. In one embodiment, the temperature sensor may be a sensor within a chip package, such as a sensor that measures the case temperature. In this case, only pressure is required on the hybrid cooling device to control its operation.

[0072] In one embodiment, the two sensors 701 and 703 are used to control only one or more fans of the hybrid cooling unit, and this unit control applies only to the unit's hardware, not to PCBs 503 and 505 and the electronics on those PCBs. This design increases the deployability, adjustability, and interoperability of the hybrid cooling unit. The design aims to simplify system integration and commissioning procedures, i.e., "plug and play."

[0073] Figure 8 This is a flowchart illustrating the control flow 800 for a hybrid cooling device according to one embodiment.

[0074] like Figure 8As shown, temperature and pressure sensors are used to control the operation of the hybrid cooling system, which includes a main fan and a secondary fan. The flow control process 400 can be executed by processing logic, which may include software, hardware, or a combination of both.

[0075] In operation 801, the processing logic activates the temperature sensor to measure the temperature inside the cooling plate in the hybrid cooling device and activates the pressure sensor to measure the pressure of the steam passing through the steam line.

[0076] In operation 803, the processing logic determines whether the measured temperature is below a predetermined threshold (i.e., T). 案例设计 ).

[0077] In operation 805, if the measured temperature is not below a predetermined threshold, the processing logic can send a command to cause the main fan in the hybrid cooling unit to run at its maximum speed.

[0078] In operation 806, the processing logic determines whether the measured temperature has dropped below a predetermined threshold due to the main fan blowing at its maximum speed.

[0079] In operation 807, the measured temperature has dropped below the threshold control. The processing logic continues to monitor the temperature and uses the measured pressure to control the operation of the hybrid cooling unit.

[0080] In operation 808, if the measured temperature does not drop below the threshold control, the processing logic will cause the auxiliary fan to run at its maximum speed.

[0081] In operation 809, the processing logic determines whether the measured pressure has increased.

[0082] In operation 811, the processing logic determines that the measured pressure has not increased and accordingly reduces the speed of the main fan.

[0083] In operation 813, the processing logic determines that the measured pressure has increased, and if the main fan is not running at its maximum speed, it increases the speed of the main fan accordingly.

[0084] In operation 815, the processing logic determines whether the measured temperature exceeds a predetermined threshold. If it does, the processing logic monitors the measured temperature to determine whether it drops below the predetermined threshold; otherwise, the processing logic checks whether the measured pressure has increased.

[0085] Figure 9 A method 900 for cooling a heterogeneous computing architecture according to one embodiment is shown.

[0086] like Figure 9As shown, in block 901, the phase change system includes a cooling plate, a heat sink, and an integrated channel for connecting the cooling plate and the heat sink. In block 903, a mounting clip is provided to position the electronic hardware to be cooled between the mounting clip and the cooling plate. In block 903, one or more fans are provided. The fans can be integrated with the cooling plate and heat sink, or they can be separate fans. In block 907, the phase change system is used to cool the electronic hardware, and the airflow generated by one or more fans is used to cool the heat sink and the electronic hardware.

[0087] Figure 10 This is a block diagram of an electronic rack according to one embodiment. The electronic rack 1000 can represent any electronic rack in a data center. (See reference...) Figure 10 According to one embodiment, the electronic rack 1000 includes, but is not limited to, a CDU 1001, a rack management unit (RMU) 1002, and one or more server chassis 1003A-1003E (collectively referred to as server chassis 1003). The server chassis 1003 can be inserted into a server slot array (e.g., a standard rack) from either the front end 1004 or the rear end 1005 of the electronic rack 1000. Note that although five server chassis 1003A-1003E are shown here, more or fewer server chassis may be housed within the electronic rack 1000. Also note that the specific locations of the CDU 1001, RMU 1002, and / or server chassis 1003 are for illustrative purposes only; other arrangements or configurations of the CDU 1001, RMU 1002, and / or server chassis 1003 are also possible. In one embodiment, the electronics rack 1000 can be either open to the environment or partially contained within a rack container, provided that the cooling fan can generate airflow from the front to the rear.

[0088] In addition, for at least some server chassis 1003, optional fan modules (not shown) are associated with the server chassis. Each fan module includes one or more cooling fans. The fan modules may be mounted on the rear end or electronic rack of the server chassis 1003 to generate airflow from the front end 1004, through the air area of ​​the server chassis 1003, and present at the rear end 1005 of the electronic rack 1000.

[0089] In one embodiment, CDU 1001 primarily includes a heat exchanger 1011, a liquid pump 1012, a pump controller (not shown), and other components such as a liquid reservoir, power supply, and monitoring sensors. The heat exchanger 1011 may be a liquid-to-liquid heat exchanger. The heat exchanger 1011 includes a first loop with a first pair of fluid connectors at its inlet and outlet ports for connection to external fluid supply / return lines 131-132 to form a main loop. The connectors connecting to the external fluid supply / return lines 131-132 may be provided or mounted on the rear end 1005 of the electronics rack 1000. These fluid supply / return lines 131-132 are also referred to as space liquid supply / return lines and may be connected to an external cooling system (e.g., a data center space cooling system).

[0090] In addition, heat exchanger 1011 includes a secondary loop with two ports having a second pair of fluid connectors for connection to a liquid manifold 1025 (also referred to as a rack manifold) to form a secondary loop. Liquid manifold 1025 may include a supply manifold (also referred to as a rack liquid supply line or rack supply manifold) and a return manifold (also referred to as a rack liquid return line or rack return manifold). The supply manifold supplies cooling fluid to server chassis 1003, while the return manifold allows warmer fluid to flow back to CDU 1001. Note that CDU 1001 can be any commercially available or custom-made CDU. Therefore, details of CDU 1001 will not be described here.

[0091] Each server chassis 1003 may include one or more IT components (e.g., a central processing unit (i.e., a CPU), a general-purpose / graphics processing unit (i.e., a GPU), memory and / or storage devices, etc.). Each IT component can perform data processing tasks, wherein the IT component may include software installed on storage devices, loaded into memory, and executed by one or more processors to perform data processing tasks. The server chassis 1003 may include a host server (referred to as a host node) connected to one or more compute servers (also referred to as compute nodes, such as CPU servers and GPU servers). The host server (having one or more CPUs) typically interfaces with clients via a network (e.g., the Internet) to receive requests for specific services (such as storage services, such as cloud-based storage services like backup and / or recovery) to execute applications to perform certain operations (e.g., image processing, deep data learning algorithms, or modeling, etc., as part of a "Software as a Service" (SaaS) platform). In response to a request, the host server assigns the task to one or more compute nodes or compute servers (which have one or more GPUs) managed by the host server. The compute servers perform the actual tasks, which may generate heat during operation.

[0092] The electronic rack 1000 also includes an optional RMU 1002, which is configured to provide and manage the power supplied to the server 1003 and CDU 1001. The RMU 1002 can be coupled to a power supply unit (not shown) to manage the power consumption of the power supply unit. The power supply unit may include the necessary circuitry (e.g., AC-to-DC or DC-to-AC power converters, batteries, transformers, or regulators) to provide power to the rest of the electronic rack 1000.

[0093] In one embodiment, RMU 1002 includes an optimization module 1021 and a rack management controller (RMC) 1022. RMC 1022 may include a monitor for monitoring the operational status of various components within the electronic rack 1000, such as compute nodes 1003, CDU 1001, and fan modules. Specifically, the monitor receives operational data from various sensors representing the operating environment of the electronic rack 1000. For example, the monitor may receive operational data representing the temperature of the processor, coolant, and airflow, which can be captured and collected by various temperature sensors. The monitor may also receive data representing the fan power and pump power generated by the fan modules and liquid pump 1012, which may be proportional to their respective speeds. This operational data is referred to as real-time operational data. Note that the monitor may be implemented as a separate module within RMU 1002.

[0094] Based on operational data, the optimization module 1021 uses a predetermined optimization function or model to determine a set of optimal fan speeds for the fan module and an optimal pump speed for the liquid pump 1012. This minimizes the total power consumption of the liquid pump 1012 and the fan module while ensuring that the operational data associated with the cooling fans of the liquid pump 1012 and the fan module are within their respective design specifications. Once the optimal pump speed and optimal fan speed are determined, the RMC 1022 can configure the cooling fans of the liquid pump 1012 and the fan module accordingly.

[0095] For example, based on the optimal pump speed, RMC 1022 communicates with the pump controller of CDU 1001 to control the speed of liquid pump 1012, thereby controlling the liquid flow rate of coolant supplied to liquid manifold 1025, so that coolant is distributed to at least some server chassis 1003. Similarly, based on the optimal fan speed, RMC 1022 communicates with each fan module to control the speed of each cooling fan of the fan module, thereby controlling the airflow rate of the fan module. Note that each fan module can be individually controlled with its specific optimal fan speed, and different fan modules and / or different cooling fans within the same fan module can have different optimal fan speeds.

[0096] Please note, Figure 10 The rack configuration shown is for illustrative purposes only; other configurations or arrangements may also be applicable. For example, CDU 1001 may be an optional unit. The cooling plate of server chassis 1003 may be connected to a rack manifold, which may be directly connected to space manifolds 131-132 without using a CDU. Although not shown, a power supply unit may be arranged within the electronics rack 1000. The power supply unit may be implemented as a standard chassis identical or similar to the server chassis, wherein the power supply chassis can be inserted into any standard rack to replace any server chassis 1003. Furthermore, the power supply chassis may further include a battery backup unit (BBU) to provide battery power to server chassis 1003 when the main power supply is unavailable. The BBU may include one or more battery packs, each battery pack including one or more battery cells, and charging and discharging circuitry required for charging and discharging the battery cells.

[0097] In one embodiment, as shown in the figure, the cooling device disposed in each server chassis may represent any cooling device described in this application.

[0098] The foregoing specification has described embodiments of this disclosure with reference to specific exemplary models. It is obvious that various modifications can be made to the embodiments without departing from the broader spirit and scope set forth in the claims. Therefore, this specification and the accompanying drawings should be considered illustrative rather than restrictive.

[0099] As previously described, embodiments of this disclosure may be (or include) a non-transitory machine-readable medium (e.g., microelectronic memory) storing instructions that program one or more data processing components (generally referred to herein as "processors") to perform airflow management operations, such as controlling the fan speed of one or more fans in a battery module (and / or BBU rack). In other embodiments, some of these operations may be performed by specific hardware components containing hard-wired logic. Those operations may also be performed by any combination of programmable data processing components and fixed hard-wired circuit components of any battery module described herein.

[0100] While certain aspects have been described and illustrated in the accompanying drawings, it should be understood that these aspects are merely illustrative and not limiting of the broad disclosure, nor is the disclosure limited to the specific constructions and arrangements shown and described, as various other modifications can be made by those skilled in the art. Therefore, this description should be considered illustrative rather than restrictive.

[0101] In some aspects, this disclosure may include expressions such as "at least one of [element A] and [element B]". Such expressions refer to one or more elements. For example, "at least one of A and B" could mean "A", "B", or "A and B". Specifically, "at least one of A and B" could mean "at least one A and at least one B", or "at least one of A or B". In some aspects, this disclosure may include expressions such as "[element A], [element B], and / or [element C]". Such expressions can refer to any one of these elements, or any combination of these elements. For example, "A, B, and / or C" could mean "A", "B", "C", "A and B", "A and C", "B and C", or "A, B, C".

Claims

1. A hybrid cooling device, comprising: A phase change system includes a cooling plate, a radiator, and an integrated channel connecting the cooling plate and the radiator; An assembly clamp is used to position electronic hardware to be cooled between the assembly clamp and a cooling plate, wherein the heat sink is located above the cooling plate, and wherein the cooling plate is vertically positioned to attach to the electronic hardware when the assembly clamp is clamped onto the cooling plate. as well as One or more fans are used to provide air cooling to heat sinks and electronic hardware, wherein the electronic hardware includes printed circuit boards and electronic devices packaged thereon, the electronic devices including one or more chips or power electronic devices, and wherein the phase change system, mounting clips, and one or more fans, together with the electronic devices packaged on the printed circuit board, are capable of being inserted into a peripheral bus as integrated peripheral devices, the printed circuit board being connected to the server main PCB via the peripheral component high-speed interconnect standard (PCIe) bus, the printed circuit board being an acceleration PCB; Further includes: An adapter reinforcement, located between the cooling plate and the electronic hardware, is elastic and made of an elastic material to accommodate electronic devices mounted at different heights on a printed circuit board. One or more flexible channels; Temperature and pressure sensors, which are decoupled from any electronic components on the printed circuit board, are used to control the operation of the hybrid cooling device. Among them, the adapter reinforcement works in conjunction with one or more flexible channels to maintain appropriate pressure on the electronic hardware; Further includes: The movable axis is located in one of the elastic channels; One end of the assembly clamp is inserted into the elastic channel via a movable shaft, making the one end of the assembly clamp movable on the elastic channel. The elastic channel provides force to the moving axis on both sides in the horizontal direction to properly fix the electronic hardware in the hybrid cooling device.

2. The hybrid cooling device of claim 1, further comprising: The device frame, to which the radiator, integrated channel and cooling plate are attached.

3. The hybrid cooling device according to any one of claims 1 to 2, wherein, The integrated channel includes a steam line and a liquid line. The liquid line is used to transfer liquid from the radiator to the cooling plate, while the steam line is used to transfer steam from the cooling plate to the radiator.

4. The hybrid cooling device according to any one of claims 1 to 2, wherein, Each of one or more fans can be integrated into a hybrid cooling unit or be a separate fan.

5. The hybrid cooling device according to any one of claims 1 to 2, wherein, Airflow generated by one or more fans passes through a first dedicated channel across the cooling plate and through a second dedicated channel across the radiator.

6. A server chassis, comprising: The hybrid cooling device according to any one of claims 1 to 5; as well as Chassis fans are used to provide airflow to cool the server chassis and hybrid cooling system.

7. An electronic rack, comprising: Multiple server chassis as described in claim 6.

Images