Multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode
By designing a multi-source information fusion FPGA heterogeneous acceleration system, combined with the systolic array structure and block matrix multiplication, the problem of multi-source information fusion in vertical agriculture is solved and efficient environmental monitoring and calculation are achieved.
Patent Information
- Application Number
- CN202411375788.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-29
- Publication Date
- 2025-01-21
AI Technical Summary
The existing technology is difficult to achieve multi-source information fusion, resulting in the inability to effectively monitor and manage the vertical agricultural greenhouse environment, affecting plant growth.
Design a multi -source information fusion FPGA heterogeneous acceleration system. Through a three -dimensional cultivation rack and acceleration module, combining the pulse array structure and block matrix multiplication, the high -speed computing and resource optimization of matrix multiplication are realized.
It improves the accuracy and real-time performance of environmental monitoring, reduces algorithm complexity and resource occupancy, and achieves efficient calculation of multi-source information fusion.
Smart Images
Figure CN119336702A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of stereoscopic cultivation, and in particular to a multi-source information fusion FPGA heterogeneous acceleration system for a stereoscopic cultivation mode. Background Art
[0002] In vertical agricultural greenhouses, multi-layer shelves are used to densely plant plants in a three-dimensional cultivation mode to maximize the yield per unit area. Creating a suitable growth environment is a necessary condition to ensure that plants grow well in the three-dimensional cultivation racks of the facility environment. However, the greenhouse system is a complex time-varying system, and a single sensor cannot accurately and effectively describe the various interactions between various indoor environmental factors and plant growth. In order for plants to grow in the best environment, real-time monitoring of environmental parameters must be achieved, including environmental parameters such as temperature, humidity, and light intensity; as well as some intelligent devices for monitoring plant growth, such as depth cameras and hyperspectral cameras. However, there is currently a lack of a three-dimensional cultivation equipment based on multi-source information fusion, which makes it impossible to introduce multi-source information fusion technology into vertical agriculture for more comprehensive and effective environmental monitoring.
[0003] As a delay-sensitive application, the facility cultivation environment monitoring system has high requirements for the real-time performance of data fusion. In addition, the higher the complexity of the fusion algorithm, the greater the computational burden of the fusion center, which leads to a decrease in the overall system performance, thereby affecting the accuracy of data fusion. The efficiency optimization of matrix multiplication is the key to determining the performance of the fusion algorithm. Matrix multiplication operations are widely used in fusion algorithms, but a large number of matrix multiplication operations reduce the real-time performance of the data fusion process and occupy a lot of resources.
[0004] Therefore, technicians in this field are committed to developing a multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode, which can realize multi-source information fusion technology, achieve high-speed calculation of matrix multiplication and reduce resource utilization. Summary of the invention
[0005] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is how to realize multi-source information fusion, how to improve the real-time performance of the fusion process and reduce the complexity of the algorithm.
[0006] To achieve the above-mentioned purpose, the present invention provides a multi-source information fusion FPGA heterogeneous acceleration system for a three-dimensional cultivation mode, comprising a three-dimensional cultivation frame, a host and an acceleration module, wherein the three-dimensional cultivation frame comprises two cultivation areas and a hyperspectral data acquisition area; the cultivation area is used to plant cultivated plants; the cultivation area and the hyperspectral data acquisition area are used to obtain plant information during the cultivation process; the three-dimensional cultivation frame sends the acquired plant information to the host, the host uses a data fusion algorithm to calculate according to the received plant information, and sends the plant information to the acceleration module through a 10 Gigabit optical fiber network card; the acceleration module adopts a pulsating array structure combined with block matrix multiplication; each of the cultivation areas comprises an LED light group, a first binocular depth camera, a second binocular depth camera, an RGB camera, and a three-in-one sensor; a hyperspectral camera is arranged above the hyperspectral data acquisition area; the acceleration module is based on FPGA.
[0007] Furthermore, the three-dimensional cultivation rack also includes a shading back plate, a first shading side plate, a second shading side plate, a third shading side plate, a first shading horizontal plate, a second shading horizontal plate, a third shading horizontal plate, a fourth shading horizontal plate, and a fifth shading horizontal plate; the first shading side plate, the second shading side plate, and the third shading side plate are vertically arranged, and one side of the first shading side plate, the second shading side plate, and the third shading side plate is fixedly connected to the shading back plate; the first shading horizontal plate, the second shading horizontal plate, and the third shading horizontal plate are horizontally and parallelly arranged between the first shading side plate and the second shading side plate; the fourth shading horizontal plate and the fifth shading horizontal plate are horizontally and parallelly arranged Placed between the second light-shielding side plate and the third light-shielding side plate; the first light-shielding horizontal plate is aligned with the top end of the fourth light-shielding horizontal plate; the third light-shielding horizontal plate is aligned with the bottom end of the fifth light-shielding horizontal plate; the first light-shielding side plate, the second light-shielding side plate, the first light-shielding horizontal plate, the second light-shielding horizontal plate, and the light-shielding back plate form a first cultivation area; the first light-shielding side plate, the second light-shielding side plate, the second light-shielding horizontal plate, the third light-shielding horizontal plate, and the light-shielding back plate form a second cultivation area; the second light-shielding side plate, the third light-shielding side plate, the fourth light-shielding horizontal plate, the fifth light-shielding horizontal plate, and the light-shielding back plate form the hyperspectral data acquisition area.
[0008] Furthermore, the cultivation area also includes a first sliding cross bar, a second sliding cross bar, a first sliding wheel group, a second sliding wheel group, and a third sliding wheel group; the first sliding cross bar is arranged on the side of the first shading side plate and the second shading side plate away from the shading back plate, and the second sliding cross bar is arranged above the LED light group.
[0009] Furthermore, the LED lamp group is used to provide an artificial lighting environment for plant photosynthesis.
[0010] Furthermore, the first binocular depth camera is connected to the first sliding cross bar through the first sliding wheel group; the first binocular depth camera is used to obtain depth information of the plant growth morphology at a side view angle and send the information to the host.
[0011] Furthermore, the second binocular depth camera is connected to the second sliding cross bar via the second sliding wheel group; the second binocular depth camera is used to obtain depth information of the plant growth morphology from a top-down angle, and send the information to the host; the RGB camera is connected to the second sliding cross bar via the third sliding wheel group; the RGB camera is used to obtain RGB information of the plant growth morphology, and send the information to the host.
[0012] Furthermore, the three-in-one sensor is fixedly mounted on the first light-shielding side panel; the three-in-one sensor is used to monitor and obtain the temperature, humidity and light intensity of the plant growth environment, and send the information to the host.
[0013] Furthermore, the hyperspectral camera is mounted on the fourth light-shielding horizontal plate through a fixed bracket; the hyperspectral camera is used to obtain spectral information of plants and send the information to the host.
[0014] Furthermore, the cultivation area and the hyperspectral data collection area also include blackout curtains.
[0015] Furthermore, the acceleration module is based on FPGA; the acceleration module uses block matrix multiplication to divide each first matrix composed of plant information into n second matrices, the order of the second matrix is smaller than the first matrix, and the n second matrices are multiplied using a systolic array structure to obtain a calculation result, and the acceleration module sends the calculation result back to the host through the 10 Gigabit fiber optic network card.
[0016] Compared with the prior art, the present invention has at least the following beneficial technical effects:
[0017] 1. The present invention builds a three-dimensional cultivation rack based on multi-source information fusion. By associating and combining multiple sensor information, it can effectively enhance the perception of greenhouse environmental status, improve the authenticity and comprehensiveness of environmental data collection, and thus improve the accuracy of environmental monitoring, which is of great significance to the realization of intelligent management of vertical agriculture;
[0018] 2. The present invention designs an FPGA heterogeneous matrix multiplication acceleration system to accelerate the multi-source data fusion process, introduces a systolic array structure and block matrix multiplication to optimize the performance of the multi-source information fusion algorithm, which is beneficial to solving the problems of high resource occupancy and long delay time of the existing fusion algorithm, and can provide reference value for meeting and improving the real-time performance of the fusion process, reducing the complexity of the algorithm, and speeding up the model calculation speed.
[0019] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 It is a structural schematic diagram of a three-dimensional cultivation frame of a preferred embodiment of the present invention;
[0021] Figure 2 It is another structural schematic diagram of a three-dimensional cultivation frame of a preferred embodiment of the present invention;
[0022] Figure 3 is a schematic diagram of system module connection of a preferred embodiment of the present invention;
[0023] Figure 4 It is a data flow diagram of a third-order systolic array matrix multiplier according to a preferred embodiment of the present invention.
[0024] Among them, 1-shading back panel; 2-first shading side panel; 3-second shading side panel; 4-third shading side panel; 5-first shading horizontal panel; 6-second shading horizontal panel; 7-third shading horizontal panel; 8-fourth shading horizontal panel; 9-fifth shading horizontal panel; 10-LED light group; 11-first binocular depth camera; 12-second binocular depth camera; 13-RGB camera; 14-three-in-one sensor; 15-first sliding cross bar; 16-second sliding cross bar; 17-first sliding wheel group; 18-second sliding wheel group; 19-third sliding wheel group; 20-hyperspectral camera; 21-fixed bracket; 22-shading curtain. DETAILED DESCRIPTION
[0025] The following describes several preferred embodiments of the present invention with reference to the drawings in the specification, so that the technical content is clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the protection scope of the present invention is not limited to the embodiments mentioned in the text.
[0026] In the drawings, components with the same structure are indicated by the same numerical reference numerals, and components with similar structures or functions are indicated by similar numerical reference numerals. The size and thickness of each component shown in the drawings are arbitrarily shown, and the present invention does not limit the size and thickness of each component. In order to make the illustration clearer, the thickness of the components is appropriately exaggerated in some places in the drawings.
[0027] The present embodiment provides a multi-source information fusion FPGA heterogeneous acceleration system for a three-dimensional cultivation mode, including a three-dimensional cultivation frame, a host and an acceleration module. The three-dimensional cultivation frame includes two cultivation areas and a hyperspectral data acquisition area. The cultivation area is used to plant cultivated plants. The cultivation area and the hyperspectral data acquisition area can obtain plant information during the cultivation process and send the plant information to the host. The host uses a data fusion algorithm to perform calculations based on the received plant information and sends the plant information to the acceleration module through a 10 Gigabit fiber optic network card.
[0028] like Figure 1 As shown, the three-dimensional cultivation frame includes a shading back plate 1, a first shading side plate 2, a second shading side plate 3, a third shading side plate 4, a first shading horizontal plate 5, a second shading horizontal plate 6, a third shading horizontal plate 7, a fourth shading horizontal plate 8, and a fifth shading horizontal plate 9. The first shading side plate 2, the second shading side plate 3, and the third shading side plate 4 are arranged vertically, and one side of the first shading side plate 2, the second shading side plate 3, and the third shading side plate 4 is fixedly connected to the shading back plate 1. The first shading horizontal plate 5, the second shading horizontal plate 6, and the third shading horizontal plate 7 are arranged horizontally and parallelly between the first shading side plate 2 and the second shading side plate 3. The fourth shading horizontal plate 8 and the fifth shading horizontal plate 9 are arranged horizontally and parallelly between the second shading side plate 3 and the third shading side plate 4. The first shading horizontal plate 5 is aligned with the top of the fourth shading horizontal plate 8; the third shading horizontal plate 7 is aligned with the bottom of the fifth shading horizontal plate 9. The overall structure of the three-dimensional cultivation rack is built with European standard 4040 aluminum.
[0029] The first shading side plate 2, the second shading side plate 3, the first shading horizontal plate 5, the second shading horizontal plate 6, and the shading back plate 1 form a first cultivation area, wherein the first shading horizontal plate 5 serves as a shading top plate of the first cultivation area, and the second shading horizontal plate 6 serves as a bottom plate of the first cultivation area. The first shading side plate 2, the second shading side plate 3, the second shading horizontal plate 6, the third shading horizontal plate 7, and the shading back plate 1 form a second cultivation area, wherein the second shading horizontal plate 6 serves as a shading top plate of the second cultivation area, and the third shading horizontal plate 7 serves as a bottom plate of the second cultivation area. The second shading side plate 3, the third shading side plate 4, the fourth shading horizontal plate 8, the fifth shading horizontal plate 9, and the shading back plate 1 form a hyperspectral data collection area, wherein the fourth shading horizontal plate 8 serves as a shading top plate of the hyperspectral data collection area, and the fifth shading horizontal plate 9 serves as a bottom plate of the hyperspectral data collection area.
[0030] like Figure 2As shown, each cultivation area includes an LED light group 10, a first binocular depth camera 11, a second binocular depth camera 12, an RGB camera 13, a three-in-one sensor 14, a first sliding crossbar 15, a second sliding crossbar 16, a first sliding wheel group 17, a second sliding wheel group 18, and a third sliding wheel group 19, wherein the first sliding crossbar 15 and the second sliding crossbar 16 are made of European standard 2020 aluminum, and the first sliding wheel group 17, the second sliding wheel group 18, and the third sliding wheel group 19 are responsible for connecting the camera and the crossbar so that the camera slides on the crossbar. The first sliding crossbar 15 is arranged on the side of the first shading side plate 2 and the second shading side plate 3 away from the shading back plate 1; the second sliding crossbar 16 is arranged above the LED light group 10.
[0031] The LED light group 10 is used to provide an artificial lighting environment for plant photosynthesis. The first binocular depth camera 11 is connected to the first sliding cross bar 15 through the first sliding wheel group 17, and the first binocular depth camera 11 is used to obtain the depth information of the plant growth morphology at the side view angle. The second binocular depth camera 12 is connected to the second sliding cross bar 16 through the second sliding wheel group 18, and the second binocular depth camera 12 is used to obtain the depth information of the plant growth morphology at the top view angle. The RGB camera 13 is connected to the second sliding cross bar 16 through the third sliding wheel group 19, and the RGB camera 13 is used to obtain the RGB information of the plant growth morphology. The three-in-one sensor 14 is fixedly mounted on the first shading side panel 2, and the three-in-one sensor 14 is used to monitor and obtain the temperature, humidity and light intensity of the plant growth environment.
[0032] A hyperspectral camera 20 is arranged above the hyperspectral data collection area, and the hyperspectral camera 20 is mounted on the fourth light shielding horizontal plate 8 through a fixing bracket 21, and the hyperspectral camera 20 is at a top-down angle. The fixing bracket 21 is made of European standard 3030 aluminum.
[0033] Both the two cultivation areas and the one hyperspectral data collection area include a blackout curtain 22 .
[0034] like Figure 3 As shown, the three-in-one sensor 14 uploads the plant information to the host through a WiFi wireless connection for data fusion, and the first binocular depth camera 11, the second binocular depth camera 12, the RGB camera 13, and the hyperspectral camera 20 upload the collected plant information to the host through a wired connection for data fusion. The data fusion algorithm runs in the host, and the host is responsible for passing the matrix multiplication operation in the data fusion algorithm to the acceleration module through the 10G fiber optic network card. The acceleration module is based on FPGA. The acceleration module uses a systolic array structure combined with block matrix multiplication to calculate the plant information, and transmits the calculation results back to the host through the fiber optic interface on the FPGA and the 10G fiber optic network card.
[0035] Matrix multiplication is a binary operation. The result of multiplying two matrices is a new matrix. When a matrix A with a dimension of n*m and a matrix B with a dimension of m*p are multiplied, the resulting matrix C has a dimension of n*p. According to traditional matrix multiplication, the element C in the i-th row and j-th column of matrix C is ij It can be expressed as shown in formula (1).
[0036]
[0037] Taking the multiplication of two N-order matrices as an example, using traditional matrix multiplication requires n 3 multiplication operations and n 2 (n-1) addition operations, so the time complexity of traditional matrix multiplication is 0(n 3 ), the resource complexity is 0(n 3 ).
[0038] The systolic array matrix multiplier uses a parallel computing structure. If the multiplication operation of two N-order matrices is calculated, the systolic array consists of n 2 The systolic array is composed of processing units PE, each of which contains an adder and a multiplier. The PEs in the systolic array are arranged in a mesh topology, and adjacent PEs are interconnected. When performing calculations, the data of the two matrices are input in parallel from the upper and left boundaries of the systolic array, and flow in the corresponding directions. The systolic array adopts a pipeline structure, and the calculation time required for each PE is n system clock cycles. It takes n system clock cycles from the time data flows into the systolic array to the time the first PE calculates the value of the first element in the result matrix. After that, a PE calculates the value of the element in the result matrix every system clock cycle. It takes a total of 2n+1 system clock cycles from the time data flows into the systolic array to the time the last PE outputs the value of the last element in the result matrix. Therefore, the time complexity of the systolic array matrix multiplier is 0(n), and the resource complexity is 0(n). 2 ). Taking the multiplication of two third-order matrices as an example, the data flow diagram of the third-order systolic array matrix multiplier is as follows: Figure 4 shown.
[0039] When performing multiplication operations on higher-dimensional matrices, the matrix is divided into blocks through block matrix multiplication, and each block is treated as an element of the matrix to participate in the operation, which can greatly simplify the calculation process. In combination with the systolic array structure, the resource occupancy rate of high-dimensional matrix multiplication operations can be effectively reduced. Taking two nine-order matrices D and E for matrix multiplication as an example, each nine-order matrix is divided into nine three-order matrices through block matrix multiplication, which requires a total of 27 three-order matrix multiplication operations, as shown in formula (2). Any matrix element D in formula (2) ij and E ijThey are all third-order matrices; each third-order matrix multiplication operation is implemented using a third-order systolic array matrix multiplier.
[0040]
[0041] Taking the multiplication of two ninth-order matrices as an example, the time complexity and resource complexity of the three methods are calculated respectively using traditional matrix multiplication, systolic array matrix multiplier, and systolic array structure combined with block matrix multiplication; the traditional matrix multiplication is implemented on the MATLAB 2023b platform of Intel Core i71.8GHz, and the systolic array matrix multiplier and systolic array structure combined with block matrix multiplication are carried out on the Xilinx Kintex7 XC7K325T development board with a clock frequency of 200MHz, and the actual running time of the three methods is compared.
[0042] Table 1
[0043]
[0044] As shown in Table 1, in actual experiments, the systolic array matrix multiplier is 210.53 times faster than the traditional matrix multiplication, and the resource complexity is reduced by 88.89%; the systolic array structure combined with block matrix multiplication is 21.16 times faster than the traditional matrix multiplication, and the resource complexity is reduced by 98.77%. In summary, considering the limited FPGA resources, as the matrix dimension increases, the matrix multiplication acceleration module in the FPGA heterogeneous acceleration system uses a systolic array structure combined with block matrix multiplication to better reduce resource occupancy while improving the matrix multiplication operation speed. In addition, the FPGA heterogeneous acceleration system uses multi-channel 10G fiber optic network cards to facilitate the subsequent expansion of the number of FPGAs to establish an FPGA data fusion acceleration cluster.
[0045] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.
Claims
1. A multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode, characterized in that: It comprises a three-dimensional cultivation frame, a host and an acceleration module, wherein the three-dimensional cultivation frame comprises two cultivation areas and a hyperspectral data acquisition area; the cultivation area is used for planting cultivated plants; the cultivation area and the hyperspectral data acquisition area are used for acquiring plant information during the cultivation process; the three-dimensional cultivation frame sends the acquired plant information to the host, the host uses a data fusion algorithm to perform calculations based on the received plant information, and sends the plant information to the acceleration module through a 10 Gigabit optical fiber network card; the acceleration module adopts a pulsating array structure combined with block matrix multiplication; each of the cultivation areas comprises an LED light group, a first binocular depth camera, a second binocular depth camera, an RGB camera, and a three-in-one sensor; a hyperspectral camera is arranged above the hyperspectral data acquisition area; the acceleration module is based on FPGA.
2. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode as claimed in claim 1, characterized in that: The three-dimensional cultivation frame also includes a shading back plate, a first shading side plate, a second shading side plate, a third shading side plate, a first shading horizontal plate, a second shading horizontal plate, a third shading horizontal plate, a fourth shading horizontal plate, and a fifth shading horizontal plate; the first shading side plate, the second shading side plate, and the third shading side plate are vertically arranged, and one side of the first shading side plate, the second shading side plate, and the third shading side plate is fixedly connected to the shading back plate; the first shading horizontal plate, the second shading horizontal plate, and the third shading horizontal plate are horizontally and parallelly arranged between the first shading side plate and the second shading side plate; the fourth shading horizontal plate and the fifth shading horizontal plate are horizontally and parallelly arranged The second shading side plate and the third shading side plate; the first shading horizontal plate is aligned with the top end of the fourth shading horizontal plate; the third shading horizontal plate is aligned with the bottom end of the fifth shading horizontal plate; the first shading side plate, the second shading side plate, the first shading horizontal plate, the second shading horizontal plate and the shading back plate form a first cultivation area; the first shading side plate, the second shading side plate, the second shading horizontal plate, the third shading horizontal plate and the shading back plate form a second cultivation area; the second shading side plate, the third shading side plate, the fourth shading horizontal plate, the fifth shading horizontal plate and the shading back plate form the hyperspectral data collection area.
3. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode as claimed in claim 2, characterized in that: The cultivation area also includes a first sliding cross bar, a second sliding cross bar, a first sliding wheel group, a second sliding wheel group, and a third sliding wheel group; the first sliding cross bar is arranged on the side of the first shading side plate and the second shading side plate away from the shading back plate, and the second sliding cross bar is arranged above the LED light group.
4. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode according to claim 1, characterized in that: The LED lamp group is used to provide an artificial lighting environment for plant photosynthesis.
5. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode as claimed in claim 3, characterized in that: The first binocular depth camera is connected to the first sliding cross bar through the first sliding wheel group; the first binocular depth camera is used to obtain depth information of plant growth morphology at a side view angle and send the information to the host.
6. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode as claimed in claim 3, characterized in that: The second binocular depth camera is connected to the second sliding cross bar via the second sliding wheel group; the second binocular depth camera is used to obtain depth information of plant growth morphology from a top-down angle and send the information to the host; the RGB camera is connected to the second sliding cross bar via the third sliding wheel group; the RGB camera is used to obtain RGB information of plant growth morphology and send the information to the host.
7. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode as claimed in claim 2, characterized in that: The three-in-one sensor is fixedly mounted on the first light-shielding side plate; the three-in-one sensor is used to monitor and obtain the temperature, humidity and light intensity of the plant growth environment, and send the information to the host.
8. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode as claimed in claim 2, characterized in that: The hyperspectral camera is mounted on the fourth light-shielding horizontal plate through a fixed bracket; the hyperspectral camera is used to obtain spectral information of plants and send the information to the host.
9. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode according to claim 1, characterized in that: The cultivation area and the hyperspectral data collection area also include shade curtains.
10. The multi-source information fusion FPGA heterogeneous acceleration system for three-dimensional cultivation mode according to claim 1, characterized in that: The acceleration module is based on FPGA; The acceleration module uses block matrix multiplication to divide each first matrix composed of plant information into n second matrices, the order of the second matrix is smaller than the first matrix, and the n second matrices are multiplied using a systolic array structure to obtain a calculation result. The acceleration module sends the calculation result back to the host through the 10 Gigabit fiber optic network card.