DESC:TECHNICAL FIELD
[0001] The present invention relates generally to computer graphics and specifically to Graphics Processing Units (GPU).

BACKGROUND
[0002] The need from market for better qualities in computer graphics, particularly for three-dimension (3D), GIS and real-time computer graphics, has increased. This increase inbandwidth has been realized in industry using PCI Express based GPU cards in various form factors to meet bandwidth hungry applications for future generations of applications. The flexible aspect of PCIe technology allows scaling of speeds.
[0003] In a conventional arrangement disclosed in US 5611057, daughter card is mounted on an adapter card. The daughter card includes adapter card connectors for mounting to the adapter card and also an edge connector for insertion directly into a computer slot so that the daughter card may also function as a stand-alone card. The daughter card is mechanically and electrically compliant as an independent PCI add-in card and includes a PCI edge connector for insertion directly into a PCI slot.
[0004] In another conventional system disclosed in US 7074051B2, a first board is in electrical communication with a second board. Electrical communication connectors are positioned on one side of the first boardand slots constructed in said the boardfor allowing the electrical communication connector positioned on said second board to mate through said slots with said first board electrical communication connector.
[0005] In yet another conventional system disclosed in US 20150294434 A1, a graphics card adapter includes a printed circuit board(PCB) having a PCI Express (PCI-E) interface for transferring graphics information via a PCI-E bus and MXM connectors coupled to the PCB for matingly engaging with graphics cards having a corresponding MXM interface. The graphics card adapter includes a switch arranged on the PCB which configures a graphics bus between the switch and each of the graphics cards via the corresponding MXM interface. Each of the graphics bus is configured to have a substantially equal bandwidth, and wherein the switch multiplexes the PCI-E bus between each of the graphics busses.
[0006] In yet another conventional system disclosed in US 7248479 B2, a method includes providing a heat generating component disposed on a first side of a first circuit boardand transferring heat from the heat generating component through the first circuit board to a second side of the first circuit board. The method includes slidingly thermally coupling the second side of the first circuit board to a thermal solution disposed on a second circuit board.
[0007] However, the conventional GPU modules are not suitable for VNX form factor due to the size constraint.
[0008] Therefore, there is a need for an improved GPU module that increases efficiency and achieves durability and maintainability.

SUMMARY
[0009] This summary is provided to introduce concepts related to a Graphics Processing Unit (GPU) module. This summary is neither intended to identify essential features of the present invention nor is it intended for use in determining or limiting the scope of the present invention.
[0010] In an embodiment of the present invention, a Graphics Processing Unit (GPU) module is provided. The GPU module include a base card, a daughter card, a top cover, and a bottom cover. The base card has a Low-Power GPU (LPGPU). The daughter card has a High-Power GPU (HPGPU). The base card has an interface for a VNX backplane, an interface for the daughter card, and a power management unit. The top cover is in direct contact with the daughter card. The bottom cover is in contact with the top cover,thereby forming an enclosed cassette.
[0011] In an exemplary embodiment, the GPU module further includes a baseplate and a plurality of heat pipes. The heat pipes are configured to transfer heat from the top cover to the baseplate.
[0012] In an exemplary embodiment, the base card is connected to the daughter card by way of a mezzanine connectorto form a unitary GPU module.
[0013] In another exemplary embodiment, the base card has a VNX edge connector to interface with a computing system through the VNX backplane.
[0014] In yet another exemplary embodiment,the LPGPU is configured to operate as a standalone GPU.
[0015] In yet another exemplary embodiment, the HPGPU is configured to operate as a standalone GPU.
[0016] In yet another exemplary embodiment,the LPGPU and the HPGPU are configured to be upgraded without redesigning the GPU module.
[0017] In yet another exemplary embodiment,the GPU module includes an enclosure for heat dissipation to an external cooling mechanism.
[0018] In yet another exemplary embodiment, the unitary GPU module includes the heat pipes overlaid on a heat sink for removal of heat from the unitary GPU module.

BRIEF DESCRIPTION OF ACCOMPANYING DRAWINGS
[0019] The detailed description is described with reference to the accompanying figures.
[0020] FIG. 1 shows a connection between a GPU and a computing module in accordance with an embodiment of the present invention.
[0021] FIG. 2shows a mezzanine connector arrangement in accordance with an embodiment of the present invention.
[0022] FIG. 3 shows a High-Power GPU (HPGPU)mounted on a Low-Power GPU (LPGPU) in accordance with an embodiment of the present invention.
[0023] FIG. 4 shows a High-Power GPU (HPGPU) mounted on a Low-Power GPU (LPGPU) along with top and bottom covers in accordancewith an embodiment of the present invention.
[0024] FIG. 5 shows top and bottom covers of a unitary GPU in accordance with an embodiment of the present invention.
[0025] FIG. 6 shows an exploded view of heat pipes and a baseplate in accordance with an embodiment of the present invention.
[0026] FIG. 7 shows a side view of a GPU along with heat sink assembly in accordance with an embodiment of the present invention.
[0027] It should be appreciated by those skilled in the art that any block diagram herein represents conceptual views of illustrative systems embodying the principles of the present invention. Similarly, it will be appreciated that any flow chart, flow diagram, and the like represent various processes which may be substantially represented in computer readable medium and so executed by a computer or processor, whether or not such computer or processor is explicitly shown.

DETAILED DESCRIPTION
[0028] The various embodiments of the present inventionprovide a Graphics Processing Unit (GPU) module.
[0029] In the following description, for purpose of explanation, specific details are set forth in order to provide an understanding of the present invention. It will be apparent, however, to one skilled in the art that the present inventionmay be practiced without these details. One skilled in the art will recognize that embodiments of the present invention, some of which are described below, may be incorporated into a number of systems.
[0030] However, the systems and methods are not limited to the specific embodiments described herein. Further, structures and devices shown in the figures are illustrative of exemplary embodiments of the present inventionand are meant to avoid obscuring of the present invention.
[0031] Furthermore, connections between components and/or modules within the figures are not intended to be limited to direct connections. Rather, these components and modules may be modified, re-formatted or otherwise changed by intermediary components and modules.
[0032] References in the present inventionto “embodiment” or “embodiment” mean that a particular feature, structure, characteristic, or function described in connection with the embodiment or the embodiment is included in at least one embodiment or embodiment of the invention. The appearances of the phrase “in anembodiment” in various places in the specification are not necessarily all referring to the same embodiment.
[0033] Referring now to FIGS. 1-7, a GPU module (105) is shown in accordance with an embodiment of the present invention. The GPU module includes a Low-Power GPU (LPGPU) (103) and a High-Power GPU (HPGPU) (104) and a thermal management system. The LPGPU (103) is in a base card and the HPGPU (104) is in a daughter card.The base card has an interface for a VNX backplane, an interface for the daughter card, and a power management unit. A top cover (400) is in direct contact with the daughter card. The bottom cover (401) is in contact with the top cover (400). The heat pipes are configured to transfer heat from the top cover (400) to a baseplate (600).
[0034] In an embodiment, each of the GPUs (103 and 104) is connected to a processor through respective PCIe interfaces. For example, the LPGPU (103) uses x1 PCIe Gen1 interface (101) while the HPGPU (104) uses a x16 PCIe Gen3 interface (102). The HPGPU (104) is interfaced to the LPGPU (103) through a mezzanine connector (201). The LPGPU (103) and the HPGU (104) are connected to the VNX backplane through a VNX connector (203) present on the LPGPU (103). Power for both, the daughter card and the base card, is supplied from the VNX backplane. The power management for the base and daughter cards is controlled from the LPGPU (103). Based on the application requirement, either the LPGPU (103) or the HPGPU (104) can be selected. The LPGPU (103) may be used in 2D low resolution graphics where low power is required. The HPGPU (104) may be used in 3D high performance graphics intensive applications. The DVI-I (107) and HDMI (106) signals are provided by the LPGPU (103) and the HPGPU (104) cards respectively. These signals are routed to the backplane.
[0035] Thus, an embodiment of the present invention provides a multi-GPU module communicating with any computing system over PCIe interface. Also, the LPGPU (103), i.e., the base card acts a motherboard for the HPGPU (104), i.e., the daughter card to interface with the VNX backplane as depicted in FIG. 3.
[0036] A conduction cooling system is used for removal of heat from the GPU module (105). The arrangement of the base card and the daughter card is such that the high heat dissipating HPGPU (104) is in direct contact with the top cover (701). The top cover (400) and the bottom cover (401) provide a two-level maintenance system whose requirements are to provide both mechanical and ESD protection. The GPU cards (103 and 104) along with the top and bottom covers (400 and 401) form a unitary GPU module (500). The unitary GPU module (500) is mounted to a heat pipe (601) with the baseplate (600) before plugging into the chassis. Therefore, the heat is transferred from the top cover (400) to the baseplate (600) with the help of heat pipes (601).
[0037] In an embodiment, the LPGPU (103) is configured to operate as a standalone GPU without requiring the baseplate (600) and the heat pipes (601). This results into a very compact cassette that is suitable for applications with space constraint. The cassette has a small form factor.
[0038] Similarly, the HPGPU (104) is also configured to operate as a standalone GPU for high power applications. This provides a choice to the user to select any one of the LPGPU (103) and the HPGPU (104) depending on the power requirements. The LPGPU (103) may be selected for standard graphics applications and the HPGPU (104) may be selected for high performance graphics applications. In standard graphics applications, the HPGPU (104) may be shut off to save power as there is no requirement for high performance.
[0039] The foregoing description of the invention has been set merely to illustrate the invention and is not intended to be limiting. Since modifications of the disclosed embodiments incorporating the spirit and substance of the invention may occur to person skilled in the art, the invention should be construed to include everything within the scope of the invention.
,CLAIMS:
1. A Graphics Processing Unit (GPU) module(105) comprising:
a daughter card having a High-Power GPU (HPGPU) (104);
a base card having a Low-Power GPU (LPGPU)(103), said base card including:
an interface for a VNX backplane;
an interface for the daughter card; and
a power management unit;
a top cover (400) in direct contact with the daughter card; and
a bottom cover(401) in contact with the top cover (400), thereby forming an enclosed cassette.

2. The GPU module (105) as claimed in claim 1, further comprising:
a baseplate (600); and
a plurality of heat pipes configured to transfer heat from the top cover (400) to the baseplate (600).

3. The GPU module (105) as claimed in claim 1, wherein the base card is connected to the daughter card by way of a mezzanine connector(201) to form a unitary GPU module (500).

4. The GPU module (105) as claimed in claim 1, wherein the base card has a VNX edge connector to interface with a computing system through the VNX backplane.

5. The GPU module (105) as claimed in claim 1, wherein the LPGPU (103) is configured to operate as a standalone GPU.

6. The GPU module (105) as claimed in claim 1, wherein the HPGPU (104) is configured to operate as a standalone GPU.

7. The GPU module (105) as claimed in claim 1, wherein the LPGPU(103) and the HPGPU (104) are configured to be upgraded without redesigning the GPU module.

8. The GPU module (105) as claimed in claim 2, comprising an enclosure for heat dissipation to an external cooling mechanism.

9. The GPU module (105) as claimed in claim 3, wherein the unitary GPU module (500) includes the heat pipes overlaid on a heat sink for removal of heat.