WO2024021606A1 - 芯片散热方法、单板、电子设备、计算机设备和存储介质 - Google Patents

芯片散热方法、单板、电子设备、计算机设备和存储介质 Download PDF

Info

Publication number
WO2024021606A1
WO2024021606A1 PCT/CN2023/079590 CN2023079590W WO2024021606A1 WO 2024021606 A1 WO2024021606 A1 WO 2024021606A1 CN 2023079590 W CN2023079590 W CN 2023079590W WO 2024021606 A1 WO2024021606 A1 WO 2024021606A1
Authority
WO
WIPO (PCT)
Prior art keywords
chip
heat exchange
thermal resistance
mode
heat dissipation
Prior art date
Application number
PCT/CN2023/079590
Other languages
English (en)
French (fr)
Inventor
陶成
刘帆
周晓东
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2024021606A1 publication Critical patent/WO2024021606A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16—Constructional details or arrangements
    • G06F1/20—Cooling means
    • G06F1/206—Cooling means comprising thermal management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16—Constructional details or arrangements
    • G06F1/20—Cooling means
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F28—HEAT EXCHANGE IN GENERAL
    • F28D—HEAT-EXCHANGE APPARATUS, NOT PROVIDED FOR IN ANOTHER SUBCLASS, IN WHICH THE HEAT-EXCHANGE MEDIA DO NOT COME INTO DIRECT CONTACT
    • F28D15/00—Heat-exchange apparatus with the intermediate heat-transfer medium in closed tubes passing into or through the conduit walls ; Heat-exchange apparatus employing intermediate heat-transfer medium or bodies
    • F28D15/02—Heat-exchange apparatus with the intermediate heat-transfer medium in closed tubes passing into or through the conduit walls ; Heat-exchange apparatus employing intermediate heat-transfer medium or bodies in which the medium condenses and evaporates, e.g. heat pipes
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • Embodiments of the present application relate to, but are not limited to, the field of heat dissipation technology, and in particular, to a chip heat dissipation method, a single board, electronic equipment, computer equipment, and a computer-readable storage medium.
  • Embodiments of the present application provide a chip heat dissipation method, a single board, electronic equipment, computer equipment, and a computer-readable storage medium.
  • inventions of the present application provide a chip heat dissipation method.
  • the chip heat dissipation method includes: determining the allowable thermal resistance value of each chip according to the first characteristic parameters of multiple chips; determining the heat exchange mode of each chip.
  • the heat exchange mode It includes a first heat exchange mode and a second heat exchange mode; when the heat exchange mode is the first heat exchange mode, the heat dissipation mode of the corresponding chip is determined according to the allowable thermal resistance value and the first thermal resistance threshold; when the heat exchange mode is In the case of the second heat exchange mode, the heat dissipation mode of the corresponding chip is determined based on the allowed thermal resistance value and the second thermal resistance threshold, where the first thermal resistance threshold is greater than the second thermal resistance threshold.
  • embodiments of the present application provide a single board on which multiple chips are disposed, and the chip heat dissipation method as described in the first aspect is used to determine the heat dissipation mode of each of the chips.
  • an embodiment of the present application provides an electronic device, including the single board described in the second aspect.
  • embodiments of the present application provide a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, The chip heat dissipation method described in the first aspect above.
  • embodiments of the present application provide a computer-readable storage medium.
  • the computer-readable storage medium stores A computer executable program is stored, and the computer executable program is used to cause the computer to execute the chip cooling method described in the first aspect.
  • Figure 1 is a flow chart of a chip heat dissipation method provided by an embodiment of the present application
  • FIG. 2 is a sub-flow chart of step S102 in Figure 1;
  • FIG. 3 is a sub-flow chart of step S203 in Figure 2;
  • Figure 4a is a schematic structural diagram of a single board provided by an embodiment of the present application.
  • Figure 4b is another schematic structural diagram of a single board provided by an embodiment of the present application.
  • Figure 4c is another schematic structural diagram of a single board provided by an embodiment of the present application.
  • Figure 4d is another structural schematic diagram of a single board provided by an embodiment of the present application.
  • Figure 4e is another schematic structural diagram of a single board provided by an embodiment of the present application.
  • Figure 5 is a schematic structural diagram of a computer device provided by an embodiment of the present application.
  • Embodiments of the present application provide a chip heat dissipation method, a single board, electronic equipment, computer equipment and a computer-readable storage medium.
  • the method includes: determining the allowable thermal resistance value of each chip according to the first characteristic parameters of multiple chips; determining each The heat exchange mode of the chip, the heat exchange mode includes a first heat exchange mode and a second heat exchange mode; when the heat exchange mode is the first heat exchange mode, the corresponding chip is determined according to the allowable thermal resistance value and the first thermal resistance threshold The heat dissipation mode; when the heat exchange mode is the second heat exchange mode, the heat dissipation mode of the corresponding chip is determined according to the allowable thermal resistance value and the second thermal resistance threshold, where the first thermal resistance threshold is greater than the second thermal resistance threshold.
  • FIG. 1 is a flow chart of a chip heat dissipation method provided by an embodiment of the present application.
  • Chip cooling methods include but are not limited to the following steps:
  • Step S101 Determine the allowable thermal resistance value of each chip according to the first characteristic parameters of the multiple chips, where the multiple chips Integrated on the same board;
  • Step S102 determine the heat exchange mode of each chip, where the heat exchange mode includes a first heat exchange mode and a second heat exchange mode;
  • Step S103 when the heat exchange mode is the first heat exchange mode, determine the heat dissipation mode of the corresponding chip according to the allowable thermal resistance value and the first thermal resistance threshold;
  • Step S104 when the heat exchange mode is the second heat exchange mode, determine the heat dissipation mode of the corresponding chip according to the allowed thermal resistance value and the second thermal resistance threshold, where the first thermal resistance threshold is greater than the second thermal resistance threshold.
  • the first heat exchange mode refers to a single-phase heat exchange mode, that is, during the heat dissipation process, the coolant remains in a liquid state and will not undergo a phase change and convert into a gaseous state due to the heat dissipated by the chip.
  • the second heat exchange mode refers to the two-phase heat exchange mode, that is, during the heat dissipation process, the coolant will undergo a phase change due to the heat dissipated by the chip, converting from a liquid state to a gaseous state, and then proceed in the heat dissipation pipe. Condensation reliquefies into liquid heat exchange mode.
  • the first characteristic parameters may include the temperature requirements of the chip, the liquid inlet temperature of the single board, the power consumption value of the chip, the interface thermal resistance between the chip and the single board, and the shell thermal resistance of the chip.
  • the temperature requirements of the chip, the board liquid temperature and the power consumption value of the chip the total thermal resistance limit between the chip and the outside world can be calculated.
  • the interface thermal resistance sum is subtracted from the total thermal resistance limit value.
  • the allowable cold plate thermal resistance can be calculated to meet the chip's heat dissipation requirements.
  • the first thermal resistance threshold is the thermal resistance limit value of the single-phase conventional cold plate, that is, the minimum value of the thermal resistance of the single-phase conventional cold plate.
  • the second thermal resistance threshold refers to the minimum thermal resistance of the two-phase conventional cold plate.
  • the heat dissipation mode of the corresponding chip is determined based on the allowable thermal resistance value and the first thermal resistance threshold, including: when the allowable thermal resistance value When the allowed thermal resistance value is greater than the first thermal resistance threshold, it is determined that the corresponding chip adopts the first heat dissipation mode; when the allowed thermal resistance value is less than the first thermal resistance threshold, it is determined that the corresponding chip adopts the second heat dissipation mode.
  • the first heat exchange mode is a single-phase cold plate liquid cooling heat dissipation mode
  • the second heat exchange mode is a single-phase microchannel cold plate liquid cooling heat dissipation mode.
  • the single-phase microchannel cold plate liquid cooling heat dissipation mode is compared with the single-phase cold plate Liquid cooling mode has better heat dissipation capabilities, but at the same time, it also requires higher design costs.
  • the allowable thermal resistance value of the chip is greater than the first thermal resistance threshold, it means that the heat dissipation capability of the single-phase cold plate liquid cooling mode can also meet the heat dissipation needs of the corresponding chip.
  • the chip uses single-phase cooling.
  • the plate liquid cooling heat dissipation mode is sufficient.
  • the allowable thermal resistance value of the chip is greater than the first thermal resistance threshold, it means that the heat dissipation capacity of the single-phase cold plate liquid cooling mode cannot meet the heat dissipation needs of the chip. In this case, use The single-phase cold plate liquid cooling mode may cause damage to the chip. Therefore, the single-phase microchannel cold plate liquid cooling mode must be used. Based on this, a composite liquid cooling solution is used to dissipate heat for each chip on the board, thereby reducing the chip heat dissipation design cost while meeting the heat dissipation needs of each chip on the board.
  • the heat dissipation mode of the corresponding chip is determined based on the allowable thermal resistance value and the second thermal resistance threshold, including: when the allowable thermal resistance value When the allowed thermal resistance value is greater than the second thermal resistance threshold, it is determined that the corresponding chip adopts the second heat dissipation mode; when the allowed thermal resistance value is less than the second thermal resistance threshold, it is determined that the corresponding chip adopts the second heat dissipation mode.
  • the second heat exchange mode is a two-phase cold plate liquid cooling heat dissipation mode.
  • the second heat exchange mode is a two-phase microchannel cold plate liquid cooling heat dissipation mode.
  • the two-phase microchannel cold plate liquid cooling heat dissipation mode is compared with the two-phase cold plate.
  • Liquid cooling mode has better heat dissipation power, but at the same time, it also requires higher design costs.
  • the allowable thermal resistance value of the chip when the allowable thermal resistance value of the chip is greater than the second thermal resistance threshold, it means that the heat dissipation capability of the two-phase cold plate liquid cooling mode can also meet the heat dissipation needs of the corresponding chip.
  • the chip adopts the two-phase cold plate liquid cooling mode.
  • the plate liquid cooling heat dissipation mode is sufficient.
  • the allowable thermal resistance value of the chip is greater than the second thermal resistance threshold, it means that the heat dissipation capacity of the two-phase cold plate liquid cooling mode cannot meet the heat dissipation needs of the chip. In this case, use The two-phase cold plate liquid cooling mode may cause damage to the chip. Therefore, the two-phase microchannel cold plate liquid cooling mode must be used. Based on this, a composite liquid cooling solution is used to dissipate heat for multiple chips on a single board, thereby reducing chip heat dissipation design costs while meeting the heat dissipation needs of each chip on the single board.
  • each chip on the single board adopts different heat dissipation methods. mode, which can keep the temperatures of multiple chips on a single board within the operating temperature range, effectively solving the problem of serious local chip heating on the single board, resulting in uneven temperatures of each chip, and effectively improving the stability between multiple chips. temperature uniformity.
  • step S102 includes but is not limited to the following steps:
  • Step S201 obtain the second characteristic parameter
  • Step S202 determine the heat exchange power consumption threshold according to the second characteristic parameter
  • Step S203 determine the heat exchange modes of multiple chips according to the heat exchange power consumption threshold.
  • the heat exchange mode includes a single-phase heat exchange mode and a two-phase heat exchange mode.
  • the second characteristic parameters include: the latent heat of the physical properties of the working fluid, the total power consumption of the single board, the dryness requirement, the subcooling degree of the incoming liquid of the single board, and the specific heat capacity of the physical properties of the working fluid.
  • the working fluid refers to the medium substance used for the mutual conversion of thermal energy and mechanical energy.
  • the working fluid refers to the cooling liquid
  • the latent heat refers to the heat absorbed or released by the working fluid without changing the temperature.
  • Subcooling refers to the difference between the temperature at which the coolant condenses into a liquid state under a certain pressure and the saturation temperature under that pressure.
  • step S202 may include: determining the flow rate according to the latent heat of the working fluid physical properties, the total power consumption of the single board, and the dryness requirements; determining the heat exchange according to the flow rate, the subcooling degree of the single plate inlet liquid, and the specific heat capacity of the working fluid physical properties. Power consumption threshold.
  • the total heat generated by the single board can be estimated based on the total power consumption of the single board.
  • the heat dissipation required to maintain the single board in a normal working state can be estimated, combined with the latent heat of the working fluid physical properties and dryness. degree requirements, the total flow rate of coolant required can be calculated. Based on this and based on the flow rate of the coolant, the subcooling degree of the inlet liquid of the single plate and the specific heat capacity of the working fluid physical properties, the heat threshold that can be taken away while the coolant is maintained in a liquid state without phase change can be calculated. Based on this heat threshold The heat exchange power consumption threshold of the corresponding chip can be determined. Based on the heat exchange power consumption threshold, it can be compared with the total power consumption of multiple chips to determine the heat exchange mode that multiple chips should adopt. In some embodiments, it is determined that the first N chips use single-phase The heat exchange mode adopts the two-phase heat exchange mode starting from the N+1th chip.
  • step S203 includes but is not limited to the following steps:
  • Step S301 determine the chip ordering according to the connection relationship of the heat dissipation devices corresponding to each chip;
  • Step S302 Calculate the first sum of power consumption of the first N chips and the second sum of power consumption of the first N+1 chips according to the chip sorting and the power consumption value of the chips, where N is a positive integer greater than 0;
  • Step S303 compare the first total power consumption and the second total power consumption with the heat exchange power consumption threshold respectively;
  • Step S304 When the first sum of power consumption is less than the heat exchange power consumption threshold and the second sum of power consumption is greater than the heat exchange power consumption threshold, it is determined that the heat exchange mode of the first N chips is the first heat exchange mode, and it is determined that the chip sorting is in The heat exchange mode of the chips after the Nth chip is the second heat exchange mode.
  • the heat dissipation device can be a cold plate or a micro-channel cold plate. In some embodiments, this is further determined according to step S103 or step S104.
  • the heat dissipation devices corresponding to each chip are designed to be connected in series and parallel through metal pipes, hoses, etc.
  • the metal pipes and The metal pipes are connected by welding.
  • barbed interfaces can also be used at the liquid inlet and liquid outlet of the cold plate, and rubber pipes such as Teflon are used for interference fit connection.
  • clamps can also be added between the barbed interface and the intersecting pipelines to enhance the connection.
  • the coolant will flow through the heat dissipation devices corresponding to each chip in a certain order. That is, the flow order of the coolant can be determined based on the connection relationship between the heat dissipation devices. According to the flow order, the number of components on the single board can be determined. The ordering of chips on the board.
  • the chip sorting can represent the order in which the coolant flows through the heat dissipation device corresponding to each chip.
  • the coolant flows through the heat dissipation device, it will take away part of the heat dissipated by the chip through the heat dissipation device, thereby causing the coolant to
  • the total heat dissipated by the first N chips into the coolant can be determined by calculating the sum of the power consumption of the first N chips. After the total heat dissipated by the current N chips into the coolant reaches a certain threshold, it is enough to offset the liquid intrusion of the single board.
  • the degree of subcooling is increased, and the coolant is raised to a certain temperature so that when it takes away the heat emitted by the N+1 chip, it meets the phase change requirements and undergoes a phase change from liquid to vaporization to gas. Based on this, through calculation, The first sum of power consumption of N chips and the second sum of power consumption of the first N+1 chips, and the first sum of power consumption and the second sum of power consumption are respectively compared with the heat exchange power consumption threshold, thereby determining the coolant After flowing through the heat dissipation device corresponding to the Nth chip and taking away the heat, it will still maintain the liquid state.
  • the first N chips can adopt the single-phase heat exchange mode, while starting from the N+1th chip, the two-phase heat exchange mode can be adopted.
  • the chip uses a single-phase cold plate liquid cooling mode. Phase cold plate liquid cooling mode is enough.
  • the allowable thermal resistance value of the chip is less than the thermal resistance threshold of single-phase cold plate liquid cooling mode, it means that the heat dissipation capacity of single-phase cold plate liquid cooling mode is not enough to meet the corresponding requirements. Due to the heat dissipation requirements of the chip, the chip must adopt a single-phase microchannel cold plate liquid cooling mode.
  • the first N chips adopt the single-phase heat exchange mode by comparing the total power consumption of the first N chips with the heat exchange power consumption threshold, thereby avoiding the problem that the front-end chip adopts two-phase heat exchange mode.
  • the heat dissipated by the front-end chip is not enough to cause the phase change of the coolant due to the subcooling of the single board inlet liquid.
  • the allowable thermal resistance of the chip is compared with the thermal resistance threshold of the single-phase cold plate liquid cooling mode to determine whether the chip adopts the single-phase cold plate liquid cooling mode with lower design cost or the single-phase microchannel cold plate liquid cooling mode with better heat dissipation capacity. .
  • the thermal resistance threshold of the single-phase cold plate liquid cooling mode determines whether the chip adopts the single-phase cold plate liquid cooling mode with lower design cost or the single-phase microchannel cold plate liquid cooling mode with better heat dissipation capacity.
  • each chip on the single board is designed in series, the coolant needs to flow through the heat dissipation devices corresponding to multiple chips and take away the heat. After the heat is taken away, the coolant will also heat up. Therefore, If in coolant flow The chip at the moving end adopts single-phase heat exchange mode. Because the coolant temperature is too high and the temperature difference between the chip and the coolant is small, the coolant cannot take away the heat emitted by the chip or can only take away a small amount of the heat emitted by the chip. The heat dissipation efficiency is low, and the heat dissipation capacity cannot meet the heat dissipation needs of the chip.
  • the heat exchange power consumption threshold is compared with the total power consumption of the first N+1 chips.
  • the current total power consumption of the N+1 chips is greater than the heat exchange power consumption.
  • the threshold value it means that starting from the N+1th chip, the coolant subcooling degree is low, and the heat dissipated by the chip is enough to cause the coolant to undergo a phase change. Therefore, starting from the N+1th chip, subsequent chips can use two Phase heat exchange mode.
  • the coolant In the two-phase heat exchange mode, after the coolant takes away the heat emitted by the chip, it will undergo a phase change from liquid to vaporization to gaseous state, and then condense and re-liquefy in the pipeline, thereby realizing the cooling of the coolant. Cooling down effectively avoids the problem that when the coolant flows to the end of the pipeline, the heat dissipated by the first N chips causes the coolant temperature to be high, resulting in poor heat dissipation for the end chips. Based on this, the allowable thermal resistance value of the chip is compared with the thermal resistance threshold of the two-phase cold plate liquid cooling mode.
  • the allowable thermal resistance value of the chip When the allowable thermal resistance value of the chip is greater than the thermal resistance threshold, it means that the two-phase cold plate liquid cooling is used for heat dissipation. mode can meet the heat dissipation needs of the chip. When the allowable thermal resistance value of the chip is less than the thermal resistance threshold, it means that the two-phase cold plate liquid cooling mode cannot meet the heat dissipation needs of the chip.
  • the chip needs to use a two-phase micro-cooling mode. Channel cold plate liquid cooling mode.
  • embodiments of the present application also provide a single board. Multiple chips are provided on the single board, and the chip heat dissipation method proposed in the above embodiment is used to determine the heat dissipation mode of each chip. The following is combined with Figures. 4a to 4e for illustration.
  • the heat dissipation devices corresponding to chips 1 to 5 are all connected in series, where the total power consumption of chip 1 and chip 2 is less than the heat exchange power consumption threshold, and the power consumption of chips 1 to 3 is The total heat consumption is greater than the heat exchange power consumption threshold.
  • the allowable thermal resistance of chip 1 is greater than the first thermal resistance threshold.
  • the allowable thermal resistance of chip 2 is less than the first thermal resistance threshold.
  • the allowable thermal resistance of chip 3 and chip 4 are both greater than the second thermal resistance. threshold, the allowable thermal resistance of chip 5 is less than the second thermal resistance threshold.
  • chip 1 adopts the single-phase cold plate liquid cooling mode
  • chip 2 adopts the single-phase microchannel cold plate liquid cooling mode
  • both chip 3 and chip 4 The two-phase cold plate liquid cooling mode is used.
  • Chip 5 adopts the two-phase microchannel cold plate liquid cooling mode. The coolant enters the pipeline from the left end connector and flows through the corresponding heat dissipation devices from chip 1 to chip 5 in sequence.
  • the heat dissipation devices corresponding to chips 1 to 5 are all connected in series, where the total power consumption of chip 1 and chip 2 is less than the heat exchange power consumption threshold, and the power consumption of chips 1 to 3 is The total heat consumption is greater than the heat exchange power consumption threshold.
  • the allowable thermal resistance of chip 1 is greater than the first thermal resistance threshold.
  • the allowable thermal resistance of chip 2 is less than the first thermal resistance threshold.
  • the allowable thermal resistances of chips 3 to 5 are all less than the second thermal resistance threshold. Resistance threshold, based on this, chip 1 adopts single-phase cold plate liquid cooling mode, chip 2 adopts single-phase microchannel cold plate liquid cooling mode, chip 3, chip 4 and chip 5 all adopt two-phase cold plate liquid cooling mode.
  • the coolant enters the pipeline from the left connector and flows through the heat dissipation devices corresponding to chip 1 to chip 5 in sequence.
  • the heat dissipation devices corresponding to chip 1 and chip 2 adopt a parallel structure based on the single-board structure installation requirements, and are connected in series with the heat dissipation devices corresponding to chip 3, chip 4 and chip 5.
  • the total power consumption of chip 1 and chip 2 is less than the heat exchange power consumption threshold, and the total power consumption of chip 1 to chip 3 is greater than the heat exchange power consumption threshold.
  • the allowable thermal resistance of chip 1 is greater than the first thermal resistance threshold
  • the allowable thermal resistance of chip 2 is less than the first thermal resistance threshold
  • the allowable thermal resistance of chip 3 and chip 4 is greater than the second thermal resistance threshold
  • the allowable thermal resistance of chip 5 is less than the second thermal resistance threshold.
  • chip 1 adopts single-phase cold plate liquid cooling mode
  • chip 2 adopts single-phase microchannel cold plate liquid cooling mode
  • chip 3 and chip 4 both adopt two-phase cold plate liquid cooling mode
  • Chip 5 adopts a two-phase microchannel cold plate liquid cooling mode.
  • the coolant enters the pipe from the left end connector, flows through the corresponding heat dissipation devices of chip 1 and chip 2 at the same time, and then flows through the corresponding heat dissipation devices of chip 3 to chip 5 in turn.
  • the flow resistance is lower.
  • the heat dissipation devices corresponding to chip 3 and chip 4 adopt a parallel structure and are connected in series with the heat dissipation devices corresponding to chip 1, chip 2, and chip 5, where the heat dissipation devices of chip 1 and chip 2 are connected in series.
  • the total power consumption is less than the heat exchange power consumption threshold.
  • the total power consumption of chip 1, chip 2 and chip 3 is greater than the heat exchange power consumption threshold.
  • the total power consumption of chip 1, chip 2 and chip 4 is greater than the heat exchange power consumption threshold.
  • Chip 1 The allowable thermal resistance of chip 2 is greater than the first thermal resistance threshold, the allowable thermal resistance of chip 2 is less than the first thermal resistance threshold, the allowable thermal resistance of chip 3 and chip 4 is greater than the second thermal resistance threshold, and the allowable thermal resistance of chip 5 is less than the second thermal resistance threshold.
  • chip 1 adopts single-phase cold plate liquid cooling mode
  • chip 2 adopts single-phase microchannel cold plate liquid cooling mode
  • chip 3 and chip 4 both adopt two-phase cold plate liquid cooling mode
  • chip 5 Using the two-phase micro-channel cold plate liquid cooling mode, the coolant enters the pipeline from the left end connector, flows through the heat dissipation devices corresponding to chip 1 and chip 2 in sequence, and then flows through the heat dissipation devices corresponding to chip 3 and chip 4 at the same time, and then Flow through the heat dissipation device corresponding to chip 5.
  • the flow resistance is lower.
  • the heat dissipation devices corresponding to chips 1 to 5 are all connected in series, the total power consumption of chips 1 to 5 is less than the heat exchange power consumption threshold, and the allowable heat dissipation of chips 2 and 5 is The resistance is less than the first thermal resistance threshold, and the allowable thermal resistance of chip 1, chip 3, and chip 4 is greater than the first thermal resistance threshold.
  • chip 2 and chip 5 adopt the single-phase microchannel cold plate liquid cooling mode.
  • Chip 1, Chip 3 and chip 4 adopt the single-phase cold plate liquid cooling mode.
  • the coolant enters the pipeline from the left end connector and flows through the corresponding heat dissipation devices from chip 1 to chip 5 in sequence.
  • the heat exchange mode of the chip is determined based on the connection relationship of the heat dissipation devices corresponding to each chip, the power consumption value and the power consumption threshold of each chip, and then based on the allowable thermal resistance of the chip and the corresponding heat dissipation
  • the thermal resistance threshold in the mode determines the heat dissipation mode of the chip, thereby adopting different heat dissipation modes for multiple chips on the single board.
  • Multiple chips on the single board are cooled through a composite liquid cooling solution, thus effectively taking into account Thermal efficiency and design cost of multiple chips on a single board.
  • Figures 4a to 4e and relevant parts of the description are only for illustrative purposes and do not limit the embodiments of the present application.
  • those skilled in the art will determine the connection relationship between the heat dissipation devices, the heat exchange power consumption threshold and the chip.
  • the relationship between the total power consumption and any modification to the relationship between the allowable thermal resistance of the chip and the corresponding thermal resistance threshold are within the protection scope of this application.
  • An embodiment of the present application also provides an electronic device, including the single board provided in the above embodiment.
  • an embodiment of the present application also provides a computer device 500.
  • the computer device 500 includes: one or more processors 520 and a memory 510.
  • one processor 520 and a memory 510 are taken as an example.
  • the processor 520 and the memory 510 may be connected through the bus 530 or other means. In FIG. 5 , the connection through the bus is taken as an example.
  • the memory 410 can be used to store non-transitory software programs and non-transitory computer executable programs, such as the chip heat dissipation method in the above embodiments of the present application.
  • the processor 420 implements the chip heat dissipation method in the above embodiment of the present application by running non-transient software programs and programs stored in the memory 410 . For example, perform steps S101 to S104 in Figure 1, steps S201 to S203 in Figure 2, and step S301 in Figure 3. Go to step S304.
  • the heat exchange mode includes a first heat exchange mode and a second heat exchange mode; in the heat exchange mode, the first heat exchange mode is In the case of thermal mode, the heat dissipation mode of the corresponding chip is determined based on the allowable thermal resistance value and the first thermal resistance threshold; when the heat exchange mode is the second heat exchange mode, the heat dissipation mode of the corresponding chip is determined based on the allowable thermal resistance value and the second thermal resistance threshold.
  • the first thermal resistance threshold is greater than the second thermal resistance threshold.
  • the memory 510 may include a storage program area and a storage data area, wherein the storage program area may store an operating system and an application program required for at least one function; the storage data area may store information required to perform the chip cooling method in the embodiment of the present application. Data etc.
  • memory 510 may include high-speed random access memory and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage device.
  • the memory may include memory located remotely from the processor, and the remote memory may be connected to the computer device through a network. Examples of the above-mentioned networks include but are not limited to the Internet, intranets, local area networks, mobile communication networks and combinations thereof.
  • the non-transient software programs and programs required to implement the chip heat dissipation method in the above embodiment of the present application are stored in the memory 510. When executed by one or more processors, the chip heat dissipation method in the above embodiment of the present application is executed.
  • embodiments of the present application also provide a computer-readable storage medium.
  • the computer-readable storage medium stores a computer-executable program.
  • the computer-executable program is executed by one or more control processors. For example, execute the above figure. Steps S101 to S104 in 1, steps S201 to S203 in Figure 2, and steps S301 to S304 in Figure 3.
  • the heat exchange mode includes a first heat exchange mode and a second heat exchange mode; in the heat exchange mode, the first heat exchange mode is In the case of thermal mode, the heat dissipation mode of the corresponding chip is determined based on the allowable thermal resistance value and the first thermal resistance threshold; when the heat exchange mode is the second heat exchange mode, the heat dissipation mode of the corresponding chip is determined based on the allowable thermal resistance value and the second thermal resistance threshold.
  • the first thermal resistance threshold is greater than the second thermal resistance threshold.
  • Embodiments of the present application provide a chip heat dissipation method, a single board, electronic equipment, computer equipment and a computer-readable storage medium.
  • the method includes: determining the allowable thermal resistance value of each chip according to the first characteristic parameters of multiple chips; determining each The heat exchange mode of the chip, the heat exchange mode includes a first heat exchange mode and a second heat exchange mode; when the heat exchange mode is the first heat exchange mode, the corresponding chip is determined according to the allowable thermal resistance value and the first thermal resistance threshold The heat dissipation mode; when the heat exchange mode is the second heat exchange mode, the heat dissipation mode of the corresponding chip is determined according to the allowable thermal resistance value and the second thermal resistance threshold, where the first thermal resistance threshold is greater than the second thermal resistance threshold.
  • the cold heat dissipation solution dissipates heat for multiple chips on a single board, reducing the cost of chip heat dissipation design on the premise of meeting the heat dissipation needs of each chip on the single board, and effectively balancing design cost and heat dissipation efficiency.
  • Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, Digital Versatile Disk (DVD) or other optical disk storage, magnetic cassettes, tapes, disk storage or other magnetic storage devices, or may Any other medium used to store the desired information and that can be accessed by a computer.
  • communication media typically embodies a computer-readable program, data structure, program module or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Cooling Or The Like Of Electrical Apparatus (AREA)

Abstract

本申请实施例提供了一种芯片散热方法、单板、电子设备、计算机设备和计算机可读存储介质，方法包括：根据多个芯片的第一特征参数确定各个芯片的允许热阻值（S101）；确定各个芯片的换热模式，换热模式包括第一换热模式和第二换热模式（S102）；在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式（S103）；在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式（S104）。

Description

芯片散热方法、单板、电子设备、计算机设备和存储介质
相关申请的交叉引用
本申请基于申请号为202210896127.0、申请日为2022年07月27日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本申请实施例涉及但不限于散热技术领域，特别是涉及一种芯片散热方法、单板、电子设备、计算机设备和计算机可读存储介质。
背景技术
近些年来，随着互联网技术不断发展，通讯和I T等领域对电子设备的集成度越来越高，高集成化的大功耗芯片越来越多的被应用到单板***中，而由于单板上集成有多个功耗不一的芯片，各芯片发热程度不一，由此对散热的需求不一，在相关技术中，单相液冷单板虽然技术成熟度高，当多芯片串联下存在温度不均的问题，流动末端的芯片易出现超温现象，两相液冷单板散热能力更强，同时能保证多芯片串联的温度均匀性，但设计成本更高，且由于冷却液过冷度的存在，冷却液先流经的芯片所散发的热量不足以使冷却液发生相变时，会存在单相换热风险，同时两相液冷单板散热无法满足超高功率密度芯片的散热需求，而芯片级的液冷微通道散热技术设计成本过高，各种散热模式各有利弊，在单板设计中如何为各芯片选择合理的散热模式，以兼顾散热设计成本以及散热效率成为亟需解决的问题。
发明内容
以下是对本文详细描述的主题的概述。本概述并非是为了限制权利要求的保护范围。
本申请实施例提供了一种芯片散热方法、单板、电子设备、计算机设备和计算机可读存储介质。
第一方面，本申请实施例提供了一种芯片散热方法，芯片散热方法包括：根据多个芯片的第一特征参数确定各个芯片的允许热阻值；确定各个芯片的换热模式，换热模式包括第一换热模式和第二换热模式；在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式；在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，其中，第一热阻阈值大于第二热阻阈值。
第二方面，本申请实施例提供了一种单板，单板上设置有多个芯片，并应用如第一方面所述的芯片散热方法确定各个所述芯片的散热模式。
第三方面，本申请实施例提供了一种电子设备，包括如第二方面所述的单板。
第四方面，本申请实施例提供了一种计算机设备，包括：存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，其中，所述处理器执行所述计算机程序时实现如上述第一方面所述的芯片散热方法。
第五方面，本申请实施例提供了一种计算机可读存储介质，所述计算机可读存储介质存 储有计算机可执行程序，计算机可执行程序用于使计算机执行如上第一方面所述的芯片散热方法。
本申请的其它特征和优点将在随后的说明书中阐述，并且，部分地从说明书中变得显而易见，或者通过实施本申请而了解。本申请的目的和其他优点可通过在说明书、权利要求书以及附图中所特别指出的结构来实现和获得。
附图说明
图1是本申请一个实施例提供的一种芯片散热方法的流程图；
图2是对图1中步骤S102的子流程图；
图3是对图2中步骤S203的子流程图；
图4a本申请实施例提供的单板的一种结构示意图；
图4b本申请实施例提供的单板的另一种结构示意图；
图4c本申请实施例提供的单板的又一种结构示意图；
图4d本申请实施例提供的单板的还一种结构示意图；
图4e本申请实施例提供的单板的再一种结构示意图；
图5是本申请一个实施例提供的一种计算机设备结构示意图。
附图用来提供对本申请技术方案的进一步理解，并且构成说明书的一部分，与本申请的实施例一起用于解释本申请的技术方案，并不构成对本申请技术方案的限制。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的实施例仅用以解释本申请，并不用于限定本申请。
应了解，在本申请实施例的描述中，多个(或多项)的含义是两个以上，大于、小于、超过等理解为不包括本数，以上、以下、以内等理解为包括本数。如果有描述到“第一”、“第二”等只是用于区分技术特征为目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量或者隐含指明所指示的技术特征的先后关系。
本申请实施例提供了一种芯片散热方法、单板、电子设备、计算机设备和计算机可读存储介质，方法包括：根据多个芯片的第一特征参数确定各个芯片的允许热阻值；确定各个芯片的换热模式，换热模式包括第一换热模式和第二换热模式；在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式；在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，其中，第一热阻阈值大于第二热阻阈值。通过确定芯片的换热模式，再比较芯片的允许热阻值与该种换热模式下的预设热阻阈值的大小关系，基于该比较结果，判断芯片应当采用哪种散热模式。由此，为各个芯片选择合理的散热模式，通过复合液冷散热方案对单板上的各个芯片进行散热，以实现在满足单板上各个芯片的散热需求的情况下降低芯片的散热设计成本。
参照图1，图1是本申请实施例提供的一种芯片散热方法的流程图。芯片散热方法包括但不限于如下步骤：
步骤S101，根据多个芯片的第一特征参数确定各个芯片的允许热阻值，其中，多个芯片 集成于同一单板上；
步骤S102，确定各个芯片的换热模式，其中，换热模式包括第一换热模式和第二换热模式；
步骤S103，在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式；
步骤S104，在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，其中，第一热阻阈值大于第二热阻阈值
可以理解的是，第一换热模式是指单相换热模式，即在散热过程中，冷却液一直维持液态，不会因为芯片所散发的热量而发生相变转换为气态的换热模式。
可以理解的是，第二换热模式是指两相换热模式，即在散热过程中，冷却液会因为芯片所散发的热量而发生相变，从液态转换为气态，在散热管道中再进行冷凝重新液化成为液态的换热模式。
可以理解的是，第一特征参数可以包括芯片的温度要求、单板进液温度、芯片的功耗值、芯片与单板之间的界面热阻、芯片的结壳热阻。根据芯片的温度要求、单板进液温度以及芯片的功耗值可以计算出芯片与外界之间的热阻总和极限值，在此基础上，从热阻总和极限值中减去界面热阻和结壳热阻的部分，即可算出满足芯片散热需求的情况下，所允许的冷板热阻值。
可以理解的是，第一热阻阈值是单相常规冷板的热阻极限值，即单相常规冷板的热阻的最小值。同理第二热阻阈值是指两相常规冷板的热阻的最小值，需要指出的是，由于采用单相常规冷板进行液冷散热时，要求芯片所散发并传导到冷却液的热量不能使冷却液发生相变，即在这种散热模式下，散热能力较差，由此对冷板热阻要求较低，单相常规冷板的热阻阈值会大于两相常规冷板的热阻阈值，即第一热阻阈值大于第二热阻阈值。
可以理解的是，在一些实施例中，在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式，包括：当允许热阻值大于第一热阻阈值时，确定对应芯片采用第一散热模式；当允许热阻值小于第一热阻阈值时，确定对应芯片采用第二散热模式。其中，第一换热模式单相冷板液冷散热模式，第二换热模式为单相微通道冷板液冷散热模式，单相微通道冷板液冷散热模式相较于单相冷板液冷散热模式具有更好的散热能力，但与此同时，也需要更高的设计成本。在一些实施例中，当芯片的允许热阻值大于第一热阻阈值时，即说明采用单相冷板液冷散热模式的散热能力也能满足对应芯片的散热需求，该芯片采用单相冷板液冷散热模式即可，而当芯片的允许热阻值大于第一热阻阈值时，即说明单相冷板液冷散热模式的散热能力无法满足芯片的散热需求，在此情况下，采用单相冷板液冷散热模式可能导致该芯片损坏，由此，必须采用单相微通道冷板液冷散热模式。基于此，通过复合液冷散热方案对单板上各个芯片进行散热，实现在满足单板上各个芯片散热需求的情况下降低芯片散热设计成本。
可以理解的是，在一些实施例中，在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，包括：当允许热阻值大于第二热阻阈值时，确定对应芯片采用第二散热模式；当允许热阻值小于第二热阻阈值时，确定对应芯片采用第二散热模式。其中，第二换热模式两相冷板液冷散热模式，第二换热模式为两相微通道冷板液冷散热模式，两相微通道冷板液冷散热模式相较于两相冷板液冷散热模式具有更好的散热能 力，但与此同时，也需要更高的设计成本。在一些实施例中，当芯片的允许热阻值大于第二热阻阈值时，即说明采用两相冷板液冷散热模式的散热能力也能满足对应芯片的散热需求，该芯片采用两相冷板液冷散热模式即可，而当芯片的允许热阻值大于第二热阻阈值时，即说明两相冷板液冷散热模式的散热能力无法满足芯片的散热需求，在此情况下，采用两相冷板液冷散热模式可能导致该芯片损坏，由此，必须采用两相微通道冷板液冷散热模式。基于此，通过复合液冷散热方案对单板上的多个芯片进行散热，实现在满足单板上各个芯片散热需求的情况下降低芯片散热设计成本。
可以理解的是，在本申请实施例中，通过将芯片的允许热阻值与芯片所处换热模式下对应的热阻极限值进行比较，从而确定单板上的各个芯片分别采用不同的散热模式，由此可使单板上的多个芯片的温度均维持工作温度范围内，有效解决了单板上局部芯片发热严重，导致各芯片温度不均的问题，有效提高了多个芯片之间的温度均匀性。
如图2所示，在一些实施例中，步骤S102包括但不限于如下步骤:
步骤S201，获取第二特征参数；
步骤S202，根据第二特征参数确定换热功耗阈值；
步骤S203，根据换热功耗阈值确定多个芯片的换热模式。
可以理解的是，换热模式包括单相换热模式和两相换热模式。
可以理解的是，第二特征参数包括：工质物性潜热、单板的总功耗、干度要求、单板进液过冷度、工质物性比热容。其中，工质是指用于热能和机械能相互转化的媒介物质，在本申请中工质指冷却液，潜热是指工质在温度不发生变化的情况下吸收或放出的热量，单板进液过冷度是指冷却液在一定压力下冷凝为液态的温度与该压力下饱和温度之间的差值，由于过冷度的存在，会使得冷却液在带走芯片所散发的热量后，当芯片所散发的热量不足以抵消过冷度时，冷却液会维持液态不发生相变。在一些实施例中，步骤S202可以包括：根据工质物性潜热、单板的总功耗、干度要求确定流量；根据流量，单板进液过冷度、工质物性比热容确定所述换热功耗阈值。在一些实施例中，根据单板总功耗可以估算出单板所产生的总热量，基于此可以估算出使单板维持在正常工作状态所需的散热量，再结合工质物性潜热以及干度要求，可以计算出所需的冷却液的总流量。基于此再根据冷却液的流量，单板进液过冷度以及工质物性比热容可以计算出使冷却液维持在液态而不发生相变的情况下的可以带走的热量阈值，基于该热量阈值可以确定对应芯片换热功耗阈值。基于该换热功耗阈值，可以将其与多个芯片的功耗总和进行比较，即可确定多个芯片应当采取的换热模式，在一些实施例中，即确定前N个芯片采用单相换热模式，从第N+1个芯片开始采用两相换热模式。
可以理解的是，冷却液在流经前N个芯片对应的散热器件即流经管路前半段时，由于过冷度的存在，冷却液会维持在一定温度范围内，而当冷却液流经从第N+1芯片开始的末端芯片对应的散热器件时，由于采用两相换热模式，通过相变的方式对冷却液进行降温，由此保证冷却液在流经管路后半段时仍能维持在一定温度范围内，由此在保证较高散热效率的前提下，降低了设计成本。
如图3所示，在一些实施例中，步骤S203包括但不限于如下步骤：
步骤S301，根据各个芯片对应的散热器件的连接关系确定芯片排序；
步骤S302，根据芯片排序以及芯片的功耗值计算前N个芯片的第一功耗总和与前N+1个芯片的第二功耗总和，其中，N为大于0的正整数；
步骤S303，将第一功耗总和与第二功耗总和分别与换热功耗阈值进行比较；
步骤S304，当第一功耗总和小于换热功耗阈值且第二功耗总和大于换热功耗阈值时，确定前N个芯片的换热模式为第一换热模式，确定芯片排序中处于第N个芯片之后的芯片的换热模式为第二换热模式。
可以理解的是，散热器件可以是冷板，也可以是微通道冷板，在一些实施例中，根据步骤S103或步骤S104进一步确定。
可以理解的是，单板上集成有多个芯片，各个芯片对应的散热器件之间彼此之间会通过金属管路、软管等方式进行串并联设计，在一些实施例中，金属管路与金属管路之间通过焊接的方式进行连接，在另一些实施例中，也可以在冷板的进液口和出液口采用倒刺接口，通过橡胶管路比如特氟龙进行过盈配合连接，基于此，还可以在倒刺接口和相交管路之间增加卡箍增强连接。基于此，冷却液会依照一定的顺序依次流经各个芯片对应的散热器件，即根据散热器件之间的连接关系可以确定冷却液的流经顺序，根据该流经顺序可以确定单板上的多个芯片在单板上的排序。
可以理解的是，芯片排序可以表征冷却液流经各个芯片对应的散热器件的先后顺序，而冷却液在流过散热器件时，会带走一部分芯片通过散热器件所散发的热量，从而使得冷却液升温，通过计算前N个芯片的功耗总和可以确定前N个芯片散发到冷却液中的总热量，当前N个芯片散发到冷却液中的总热量到达一定阈值后，足以抵消单板进液的过冷度，并使冷却液上升至一定温度，使其在带走第N+1个芯片所散发的热量时，满足相变要求发生相变从液态汽化为气态，基于此，通过计算前N个芯片的第一功耗总和以及前N+1的芯片的第二功耗总和，并将第一功耗总和以及第二功耗总和分别与换热功耗阈值进行比较，从而确定冷却液在流经第N个芯片对应的散热器件并带走热量后仍会维持液态，而在流经第N+1个芯片对应的散热器件并带走热量后会发生相变从液态汽化为气态，从而确定前N个芯片可以采用单相换热模式，而从第N+1芯片开始则采用两相换热模式。
可以理解的是，在本申请实施例中，通过计算换热功耗阈值，将换热功耗阈值与前N个芯片的功耗总和进行比较，从而确定前N个芯片所散发的总热量不足以使冷却液发生相变，由此，确定前N个芯片采用单相换热模式，基于此，再将各个芯片的允许热阻值与单相冷板液冷散热模式的热阻阈值进行比较，当芯片的允许热阻值大于单相冷板液冷散热模式的热阻阈值时，即说明单相冷板液冷散热模式的散热能力可以满足对应芯片的散热需求，由此该芯片采用单相冷板液冷散热模式即可，而当芯片的允许热阻值小于单相冷板液冷散热模式的热阻阈值时，即说明单相冷板液冷散热模式的散热能力不足以满足对应芯片的散热需求，由此该芯片须采用单相微通道冷板液冷散热模式。基于本申请实施例所提出的方法，一方面通过将前N个芯片的功耗总和与换热功耗阈值进行比较确定前N个芯片采用单相换热模式，由此规避若前端芯片采用两相换热模式时，由于单板进液过冷度而导致前端芯片所散发的热量不足以使冷却液发生相变所导致的单相换热风险，另一方面，通过将芯片的允许热阻值与单相冷板液冷散热模式的热阻阈值进行比较从而确定芯片采用设计成本更低的单相冷板液冷散热模式或采用散热能力更好的单相微通道冷板液冷散热模式。由此，通过复合液冷散热方案对单板上的多个芯片进行散热，有效兼顾了单板上多个芯片的散热效率和设计成本。
可以理解的是，由于单板上的各个芯片对应的通过串联设计使得冷却液需要流经多个芯片对应的散热器件并带走热量，而带走热量后冷却液也会发生升温，由此，若处于冷却液流 动的末端的芯片采用单相换热模式，会存在由于冷却液温度过高，芯片与冷却液之间温差较小，冷却液无法带走芯片所散发的热量或只能带走少量芯片所散发的热量，由此导致散热效率低，散热能力无法满足芯片散热需求的问题。在本申请实施例中，通过计算换热功耗阈值，将换热功耗阈值与前N+1个芯片的功耗总和进行比较，当前N+1个芯片的功耗总和大于换热功耗阈值时，即说明从第N+1芯片开始，冷却液过冷度较低，芯片所散发的热量足以使冷却液发生相变，由此，从第N+1芯片开始，后续芯片可以采用两相换热模式，在两相换热模式下，冷却液在带走芯片所散发的热量后，会发生相变从液态汽化为气态，再在管路中进行冷凝重新液化，从而实现对冷却液进行降温，有效避免了冷却液流至管路末端时，由于前N个芯片所散发的热量导致冷却液温度较高，对末端芯片散热效果不佳的问题。基于此，再将芯片的允许热阻值与两相冷板液冷散热模式的热阻阈值进行比较，当芯片的允许热阻值大于该热阻阈值时，说明采用两相冷板液冷散热模式即可满足该芯片的散热需求，而当芯片的允许热阻值小于该热阻阈值时，即说明两相冷板液冷散热模式无法满足该芯片的散热需求，该芯片需采用两相微通道冷板液冷散热模式。基于本申请实施例所提出的方法，一方面通过将前N+1个芯片的功耗总和与换热功耗阈值进行比较，从而确定从第N+1个芯片开始，可以采用两相换热模式，使冷却液发生相变，从而降低冷却液温度，规避了单相换热模式中冷却液流至末端管路时由于前端芯片所散发的热量导致冷却液温度较高，散热效率低，从而导致单板上各芯片温度不均匀的问题；另一方面，通过将芯片的允许热阻值与两相冷板液冷散热模式的热阻阈值进行比较从而确定芯片采用设计成本更低的两相冷板液冷散热模式或采用散热能力更好的两相微通道冷板液冷散热模式。基于此，通过复合液冷散热方案对单板上的多个芯片进行散热，有效兼顾了单板上多个芯片的散热效率和设计成本。
如图4a至图4e所示，本申请实施例还提供一种单板，单板上设置有多个芯片，并应用上述实施例所提出的芯片散热方法确定各个芯片的散热模式，以下结合图4a至图4e进行说明。
参照图4a，在一些实施例中，芯片1至芯片5对应的散热器件之间均采用串联连接，其中芯片1和芯片2的功耗总和小于换热功耗阈值，芯片1至芯片3的功耗总和大于换热功耗阈值，芯片1的允许热阻大于第一热阻阈值，芯片2的允许热阻小于第一热阻阈值，芯片3和芯片4的允许热阻均大于第二热阻阈值，芯片5的允许热阻小于第二热阻阈值，基于此，芯片1采用单相冷板液冷散热模式，芯片2采用单相微通道冷板液冷散热模式，芯片3和芯片4均采用两相冷板液冷散热模式，芯片5采用两相微通道冷板液冷散热模式，冷却液从左端连接头进入管路，依次流经芯片1至芯片5对应的散热器件。
参照图4b，在一些实施例中，芯片1至芯片5对应的散热器件之间均采用串联连接，其中芯片1和芯片2的功耗总和小于换热功耗阈值，芯片1至芯片3的功耗总和大于换热功耗阈值，芯片1的允许热阻均大于第一热阻阈值，芯片2的允许热阻小于第一热阻阈值，芯片3至芯片5的允许热阻均小于第二热阻阈值，基于此，芯片1采用单相冷板液冷散热模式，芯片2采用单相微通道冷板液冷散热模式，芯片3、芯片4和芯片5均采用两相冷板液冷散热模式，冷却液从左边连接头进入管路，依次流经芯片1至芯片5对应的散热器件。
参照图4c，在一些实施例中，芯片1和芯片2对应的散热器件基于单板结构安装要求采用了并联结构，并与芯片3、芯片4和芯片5各自对应的散热器件之间串联连接，其中芯片1和芯片2的功耗总和小于换热功耗阈值，芯片1至芯片3的功耗总和大于换热功耗阈值，芯 片1的允许热阻大于第一热阻阈值，芯片2的允许热阻小于第一热阻阈值，芯片3和芯片4的允许热阻大于第二热阻阈值，芯片5的允许热阻小于第二热阻阈值，基于此，芯片1采用单相冷板液冷散热模式，芯片2采用单相微通道冷板液冷散热模式，芯片3和芯片4均采用两相冷板液冷散热模式，芯片5采用两相微通道冷板液冷散热模式，冷却液从左端连接头进入管路，同时流经芯片1和芯片2对应的散热器件，再依次流经芯片3至芯片5对应的散热器件。在并联设计中，当冷却液流经芯片1和芯片2时，流阻较低。
参考图4d，在一些实施例中，芯片3和芯片4对应的散热器件采用并联结构，并与芯片1、芯片2、芯片5各自对应的散热器件之间串联连接，其中芯片1和芯片2的功耗总和小于换热功耗阈值，芯片1、芯片2和芯片3的功耗总和大于换热功耗阈值，芯片1、芯片2和芯片4的功耗总和大于换热功耗阈值，芯片1的允许热阻大于第一热阻阈值，芯片2的允许热阻小于第一热阻阈值，芯片3和芯片4的允许热阻大于第二热阻阈值，芯片5的允许热阻小于第二热阻阈值，基于此，芯片1采用单相冷板液冷散热模式，芯片2采用单相微通道冷板液冷散热模式，芯片3和芯片4均采用两相冷板液冷散热模式，芯片5采用两相微通道冷板液冷散热模式，冷却液从左端连接头进入管路，依次流经芯片1和芯片2对应的散热器件，再同时流经芯片3和芯片4对应的散热器件，再流经芯片5对应的散热器件。在并联设计中，当冷却液流经芯片3和芯片4时，流阻较低。
参考图4e，在一些实施例中，芯片1至芯片5对应的散热器件之间均采用串联连接，芯片1至芯片5的功耗总和小于换热功耗阈值，芯片2和芯片5的允许热阻小于第一热阻阈值，芯片1、芯片3和芯片4的允许热阻大于第一热阻阈值，基于此，芯片2和芯片5采用单相微通道冷板液冷散热模式，芯片1、芯片3和芯片4采用单相冷板液冷散热模式，冷却液从左端连接头进入管路，依次流经芯片1至芯片5对应的散热器件。
可以理解的是，在本申请实施例中，根据各个芯片对应的散热器件的连接关系、各个芯片的功耗值以及功耗阈值确定芯片的换热模式，再根据芯片的允许热阻以及对应散热模式下的热阻阈值确定芯片的散热模式，从而为单板上的多个芯片分别采取不同的散热模式，通过复合液冷散热方案对单板上的多个芯片进行散热，由此有效兼顾了单板上的多个芯片的散热效率和设计成本。
可以理解的是，图4a至图4e及说明书相关部分仅起说明作用，不对本申请实施例构成限制，本领域技术人员在此基础上对散热器件的连接关系、对换热功耗阈值与芯片功耗总和之间的大小关系以及对芯片允许热阻与对应的热阻阈值之间的大小关系所作出的任何变形均在本申请的保护范围之内。
本申请实施例还提供一种电子设备，包括有上述实施例所提供的单板。
如图5所示，本申请实施例还提供了一种计算机设备500。
在一些实施例中，该计算机设备500包括：一个或多个处理器520和存储器510，图5中以一个处理器520及存储器510为例。处理器520和存储器510可以通过总线530或者其他方式连接，图5中以通过总线连接为例。
存储器410作为一种非暂态计算机可读存储介质，可用于存储非暂态软件程序以及非暂态性计算机可执行程序，如上述本申请实施例中的芯片散热方法。处理器420通过运行存储在存储器410中的非暂态软件程序以及程序，从而实现上述本申请实施例中的芯片散热方法。例如执行以上图1中的步骤S101至S104，图2中步骤S201至步骤S203，图3中的步骤S301 至步骤S304。根据多个芯片的第一特征参数确定各个芯片的允许热阻值；确定各个芯片的换热模式，换热模式包括第一换热模式和第二换热模式；在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式；在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，其中，第一热阻阈值大于第二热阻阈值。通过确定芯片的换热模式，再比较芯片的允许热阻值与该种换热模式下的预设热阻阈值的大小关系，基于该比较结果，判断芯片应当采用哪种散热模式，由此，为各个芯片选择合理的散热模式，通过复合液冷散热方案对单板上的多个芯片进行散热，以实现在满足散热需求的前提下，降低芯片散热的设计成本。
存储器510可以包括存储程序区和存储数据区，其中，存储程序区可存储操作***、至少一个功能所需要的应用程序；存储数据区可存储执行上述本申请实施例中的芯片散热方法所需的数据等。此外，存储器510可以包括高速随机存取存储器，还可以包括非暂态存储器，例如至少一个磁盘存储器件、闪存器件、或其他非暂态固态存储器件。在一些实施方式中，存储器可包括相对于处理器远程设置的存储器，这些远程存储器可以通过网络连接至该计算机设备。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
实现上述本申请实施例中的芯片散热方法所需的非暂态软件程序以及程序存储在存储器510中，当被一个或者多个处理器执行时，执行上述本申请实施例中的芯片散热方法。
此外，本申请实施例还提供了一种计算机可读存储介质，该计算机可读存储介质存储有计算机可执行程序，该计算机可执行程序被一个或多个控制处理器执行，例如，执行以上图1中的步骤S101至S104，图2中步骤S201至步骤S203，图3中的步骤S301至步骤S304。根据多个芯片的第一特征参数确定各个芯片的允许热阻值；确定各个芯片的换热模式，换热模式包括第一换热模式和第二换热模式；在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式；在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，其中，第一热阻阈值大于第二热阻阈值。通过确定芯片的换热模式，再比较芯片的允许热阻值与该种换热模式下的预设热阻阈值的大小关系，基于该比较结果，判断芯片应当采用哪种散热模式，由此，为各个芯片选择合理的散热模式，以实现在满足散热需求的前提下，降低芯片散热的设计成本。
本申请实施例提供了一种芯片散热方法、单板、电子设备、计算机设备和计算机可读存储介质，方法包括：根据多个芯片的第一特征参数确定各个芯片的允许热阻值；确定各个芯片的换热模式，换热模式包括第一换热模式和第二换热模式；在换热模式为第一换热模式的情况下，根据允许热阻值与第一热阻阈值确定对应芯片的散热模式；在换热模式为第二换热模式的情况下，根据允许热阻值与第二热阻阈值确定对应芯片的散热模式，其中，第一热阻阈值大于第二热阻阈值。通过确定芯片的换热模式，再比较芯片的允许热阻值与该换热模式下对应的预设热阻阈值的大小关系，基于该比较结果，判断芯片应当采用哪种散热模式，通过复合液冷散热方案对单板上的多个芯片进行散热，实现在满足单板上各个芯片散热需求的前提下，降低芯片散热设计的成本，能有效兼顾设计成本和散热效率。
本领域普通技术人员可以理解，上文中所公开方法中的全部或某些步骤、***可以被实施为软件、固件、硬件及其适当的组合。某些物理组件或所有物理组件可以被实施为由处理器，如中央处理器、数字信号处理器或微处理器执行的软件，或者被实施为硬件，或者被实施为集成电路，如专用集成电路。这样的软件可以分布在计算机可读介质上，计算机可读介 质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。如本领域普通技术人员公知的，术语计算机存储介质包括在用于存储信息(诸如计算机可读程序、数据结构、程序模块或其他数据)的任何方法或技术中实施的易失性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于RAM、ROM、EEPROM、闪存或其他存储器技术、CD-ROM、数字多功能盘(DVD)或其他光盘存储、磁盒、磁带、磁盘存储或其他磁存储装置、或者可以用于存储期望的信息并且可以被计算机访问的任何其他的介质。此外，本领域普通技术人员公知的是，通信介质通常包含计算机可读程序、数据结构、程序模块或者诸如载波或其他传输机制之类的调制数据信号中的其他数据，并且可包括任何信息递送介质。
以上是对本申请的一些实施进行了说明，但本申请并不局限于上述实施方式，熟悉本领域的技术人员在不违背本申请范围的共享条件下还可作出种种等同的变形或替换，这些等同的变形或替换均包括在本申请权利要求所限定的范围内。

Claims (15)

  1. 一种芯片散热方法，所述方法包括：
    根据多个芯片的第一特征参数确定各个所述芯片的允许热阻值，其中，所述多个芯片集成于同一单板上；
    确定各个所述芯片的换热模式，其中，所述换热模式包括第一换热模式和第二换热模式；
    在所述换热模式为所述第一换热模式的情况下，根据所述允许热阻值与第一热阻阈值确定对应所述芯片的散热模式；
    在所述换热模式为所述第二换热模式的情况下，根据所述允许热阻值与第二热阻阈值确定对应所述芯片的散热模式，其中，所述第一热阻阈值大于所述第二热阻阈值。
  2. 根据权利要求1所述的方法，其中，所述第一换热模式为单相换热模式，所述第二换热模式为两相换热模式。
  3. 根据权利要求1所述的方法，其中，所述第一特征参数包括：芯片温度要求、单板进液温度、芯片的功耗值、芯片与单板之间的界面热阻、芯片的结壳热阻。
  4. 根据权利要求1所述的方法，其中，所述在所述换热模式为所述第一换热模式的情况下，根据所述允许热阻值与第一热阻阈值确定对应所述芯片的散热模式，包括:
    当所述允许热阻值大于所述第一热阻阈值时，确定对应芯片采用第一散热模式；
    当所述允许热阻值小于所述第一热阻阈值时，确定对应芯片采用第二散热模式。
  5. 根据权利要求4所述的方法，其中，所述第一散热模式为单相冷板液冷散热模式，所述第二散热模式为单相微通道冷板液冷散热模式。
  6. 根据权利要求1所述的方法，其中，所述在所述换热模式为所述第二换热模式的情况下，根据所述允许热阻值与第二热阻阈值确定对应所述芯片的散热模式，其中，所述第一热阻阈值大于所述第二热阻阈值，包括：
    当所述允许热阻值大于所述第二热阻阈值时，确定对应芯片采用第三散热模式；
    当所述允许热阻值小于所述第二热阻阈值时，确定对应芯片采用第四散热模式。
  7. 根据权利要求6所述的方法，其中，所述第三散热模式为两相冷板液冷散热模式，所述第四散热模式为两相微通道冷板液冷散热模式。
  8. 根据权利要求1所述的方法，其中，所述确定各个所述芯片的换热模式，包括：
    获取第二特征参数；
    根据所述第二特征参数确定换热功耗阈值；
    根据所述换热功耗阈值确定多个芯片的换热模式。
  9. 根据权利要求8所述的方法，其中，所述第二特征参数，包括：工质物性潜热、单板的总功耗、干度要求、单板进液过冷度、工质物性比热容。
  10. 根据权利要求9所述的方法，其中，所述根据所述第二特征参数确定所述换热功耗阈值，包括：
    根据所述工质物性潜热、所述单板的总功耗、所述干度要求确定流量；
    根据所述流量，所述单板进液过冷度、所述工质物性比热容确定所述换热功耗阈值。
  11. 根据权利要求8所述的方法，其中，所述第一特征参数包括多个芯片的功耗值，所述根据所述换热功耗阈值确定多个芯片的换热模式，包括：
    根据各个芯片对应的散热器件的连接关系确定芯片排序；
    根据所述芯片排序以及所述芯片的功耗值计算前N个芯片的第一功耗总和与前N+1个芯片的第二功耗总和，其中，N为大于0的正整数；
    将所述第一功耗总和与所述第二功耗总和分别与所述换热功耗阈值进行比较；
    当所述第一功耗总和小于所述换热功耗阈值且所述第二功耗总和大于所述换热功耗阈值时，确定前N个芯片的换热模式为第一换热模式，确定所述芯片排序中处于第N个芯片之后的芯片的换热模式为第二换热模式。
  12. 一种单板，所述单板上设置有多个芯片，并应用如权利要求1至11任意一项所述的芯片散热方法确定各个所述芯片的散热模式。
  13. 一种电子设备，包括如权利要求12所述的单板。
  14. 一种计算机设备，包括：存储器、处理器及存储在存储器上并可在处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现如权利要求1至11任意一项所述的芯片散热方法。
  15. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机可执行程序，所述计算机可执行程序用于使计算机执行如权利要求1至11任意一项所述的芯片散热方法。
PCT/CN2023/079590 2022-07-27 2023-03-03 芯片散热方法、单板、电子设备、计算机设备和存储介质 WO2024021606A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210896127.0A CN117435009A (zh) 2022-07-27 2022-07-27 芯片散热方法、单板、电子设备、计算机设备和存储介质
CN202210896127.0 2022-07-27

Publications (1)

Publication Number Publication Date
WO2024021606A1 true WO2024021606A1 (zh) 2024-02-01

Family

ID=89546847

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/079590 WO2024021606A1 (zh) 2022-07-27 2023-03-03 芯片散热方法、单板、电子设备、计算机设备和存储介质

Country Status (2)

Country Link
CN (1) CN117435009A (zh)
WO (1) WO2024021606A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120279686A1 (en) * 2011-05-06 2012-11-08 International Business Machines Corporation Cooled electronic system with liquid-cooled cold plate and thermal spreader coupled to electronic component
CN105682434A (zh) * 2016-04-20 2016-06-15 电子科技大学 一种结合热电制冷和微通道液冷的复合散热装置
CN212988425U (zh) * 2020-09-17 2021-04-16 温州泛波激光有限公司 一种冷却效率测量装置
CN112902715A (zh) * 2019-12-03 2021-06-04 中兴通讯股份有限公司 一种液冷板及散热设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120279686A1 (en) * 2011-05-06 2012-11-08 International Business Machines Corporation Cooled electronic system with liquid-cooled cold plate and thermal spreader coupled to electronic component
CN105682434A (zh) * 2016-04-20 2016-06-15 电子科技大学 一种结合热电制冷和微通道液冷的复合散热装置
CN112902715A (zh) * 2019-12-03 2021-06-04 中兴通讯股份有限公司 一种液冷板及散热设备
CN212988425U (zh) * 2020-09-17 2021-04-16 温州泛波激光有限公司 一种冷却效率测量装置

Also Published As

Publication number Publication date
CN117435009A (zh) 2024-01-23

Similar Documents

Publication Publication Date Title
US11291136B2 (en) Liquid-cooled cold plate device
US8437129B2 (en) Server cabinet
US20130342987A1 (en) Method, Apparatus, and System for Dissipating Heat of Memory with Liquid Cooling
CN107743354B (zh) 数据中心机房的制冷***和数据中心
US10260781B2 (en) Liquid cooling device having diversion mechanism
WO2017028512A1 (zh) 单板液冷散热***及机柜
US9681591B2 (en) Heat-receiving device, cooling device, and electronic device
US20090272144A1 (en) Computer cooling apparatus
US11612083B2 (en) System and method for phase-change cooling of an electronic rack
CN206075227U (zh) 一种计算机散热器
WO2011137798A1 (zh) 一种散热方法和装置
US20150124406A1 (en) Heat-receiving device, cooling device, and electronic device
WO2024021606A1 (zh) 芯片散热方法、单板、电子设备、计算机设备和存储介质
CN215601717U (zh) 一种算力板和数据处理设备
Brunschwiler et al. Direct waste heat utilization from liquid-cooled supercomputers
JP2009182313A (ja) 部品冷却構造
CN206118269U (zh) 一种换流阀的循环冷却装置
CN114007381A (zh) 一种浸没式液冷***
US20120175088A1 (en) Container data center and waste heat utilizing system therefor
CN112799489A (zh) 一种液冷散热***和服务器
WO2020135311A1 (zh) 散热装置及方法
CN106993393B (zh) 一种散热设备及终端
Sun et al. A novel concept methodology of in-direct cold plate liquid cooling design for data center in cpu socket area
CN114710931A (zh) 一种数据中心的制冷***及液冷机柜
CN208188773U (zh) 一种电脑机箱水冷散热器

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23844825

Country of ref document: EP

Kind code of ref document: A1