A control system for full-automatic testing of a display card

Abstract

The application relates to a control system for full-automatic testing of a display card and belongs to the technical field of display card testing. The control system comprises a double-layer circulating transportation module, a feeding module, four groups of independently-operated test units and a distribution scheduling module. The feeding module is used for mounting a display card to be tested on a test fixture and feeding the display card to the double-layer circulating transportation module. The four groups of test units are used for testing the functions of the display card to be tested. The double-layer circulating transportation module is used for feeding and circulating the display card to be tested from a feeding area to the test units and recycling the display card after testing from a test area to a discharging area. The distribution scheduling module is connected with the test units and is used for collecting real-time state information of the test equipment and distributing test tasks according to a preset scheduling algorithm. The double-layer circulating transportation module conveys the test fixture to the corresponding test unit according to the test task. The application solves the problems of low space utilization and operation efficiency of a display card test production line in the full-automatic testing process of the display card.

Description

Technical Field

[0001] This invention belongs to the field of graphics card testing technology, specifically relating to a control system for fully automated graphics card testing. Background Technology

[0002] As a core component of a computer system, the graphics card is responsible for graphics processing and computing tasks. During the mass production phase of graphics cards, it is necessary to place the graphics card in a complete system to simulate high-load application scenarios, so as to fully expose system-level defects such as compatibility, stability, and heat dissipation that may exist under complex interactions. Therefore, mass production testing is a critical step in ensuring the reliability of graphics card products.

[0003] Traditional graphics card testing typically involves manual connection checks, which is slow, inefficient, and labor-intensive. Employee fatigue can lead to errors, and the separate, time-consuming manual testing processes prevent simultaneous testing of various functions. Furthermore, some specialized testing equipment is bulky and has low single-station efficiency. Therefore, to avoid physical damage caused by human error and to achieve fully automated, high-precision, multi-station parallel testing of graphics cards, a solution is needed.

[0004] In the prior art, patent CN212256291U discloses a high-precision fully automatic adjustable graphics card testing fixture, including a workbench, a conveyor belt, and a cross arm. This patent automatically feeds and positions the graphics card by setting up a conveyor belt and a limiting partition. It uses dual motors to drive positive and negative lead screws to move the left and right slide plates, so that the test plug can be accurately connected to the graphics card for functional testing. At the same time, it integrates a temperature detector and a telescopic rod structure to realize temperature measurement feedback during operation and improve the comprehensiveness of testing.

[0005] However, while the aforementioned patent achieves fully automated testing of graphics cards, it still has the following shortcomings: When testing a large number of graphics cards, the graphics card testing production line uses a single-layer conveyor structure, which leads to the risk of interference between the shared tracks of the graphics cards' inbound and outbound paths; when a testing module malfunctions and stops, the graphics cards contained in that module will block the operation of subsequent graphics cards, causing the entire testing production line to stop, seriously affecting the testing efficiency of the production line. In addition, the solution uses fixed workstation binding and does not consider dynamic factors such as the physical transfer distance of the graphics cards from the installation station to the testing station, the real-time load differences of each testing area, and the heat dissipation requirements of graphics cards with different power consumptions. It cannot flexibly switch scheduling strategies according to different production scenarios and is difficult to meet diverse production needs.

[0006] Therefore, how to improve the space utilization and operating efficiency of the graphics card testing production line, avoid mutual interference between incoming and outgoing materials, and improve the fault tolerance, resource utilization and testing smoothness of the production line are problems that need to be solved by those skilled in the art. Summary of the Invention

[0007] To address the aforementioned problems in the existing technology, this invention provides a control system for fully automated graphics card testing, which solves the problems of low space utilization and low operating efficiency in existing fully automated graphics card testing production lines.

[0008] The objective of this invention can be achieved through the following technical solution: a control system for fully automated testing of graphics cards, comprising a dual-layer circulating transport module, a loading module, four independently operating test units, and an allocation and scheduling module. The loading module is used to install the graphics card under test onto the test fixture and load it into the dual-layer circulating transport module. The four test units are used to test the functionality of the graphics card under test. The dual-layer circulating transport module is used to transport the graphics card under test from the loading area to the test units and to retrieve the tested graphics card from the test area to the unloading area. The allocation and scheduling module is connected to each test unit and is used to collect real-time status information of the test equipment and allocate test tasks according to a preset scheduling algorithm. The dual-layer circulating transport module transports the test fixture to the corresponding test unit according to the test task.

[0009] As a preferred embodiment of the present invention, the double-layer circulating transport module includes an upper feeding line module and a lower recycling line module. The upper feeding line module and the lower recycling line module are connected by a vertical lifting module. The vertical lifting module is located in the testing area and is used to move the completed test fixture from the upper feeding line module to the lower recycling line module. The feeding direction of the feeding line module is opposite to the recycling direction of the lower recycling line module, forming a closed-loop circulation path.

[0010] As a preferred embodiment of the present invention, the scheduling algorithm includes: The shortest path first algorithm is used to assign graphics cards to available test points with the shortest transfer distance from the loading area to the inspection and finished product scanning process. The balanced round-robin allocation algorithm is used to sequentially and cyclically allocate the graphics cards to be tested in each test area, so as to evenly distribute the overall test task. The regional load priority algorithm is used to prioritize scheduling graphics cards to test areas with fewer graphics cards currently under test and lower load. The device idle priority algorithm is used to prioritize scheduling graphics cards to test areas with more idle test devices; The allocation and scheduling module obtains the current status of all test stations and the occupancy status of each section of the conveying path of the upper feeding line and the lower recycling line. The allocation and scheduling module selects the test station whose conveying path is not occupied from the test stations in the idle state as the target test station.

[0011] As a preferred embodiment of the present invention, the dynamic switching method of the four scheduling algorithms is as follows: The shortest path first algorithm is the default scheduling algorithm under normal production conditions. The balanced polling allocation algorithm is switched on and enabled when the single-card test takes a long time and it is necessary to make full use of all area test equipment. The regional load priority algorithm is switched on and enabled in high-power graphics card production scenarios. The equipment idle priority algorithm is switched on and activated under the requirement of balanced control of equipment utilization rate in each region. The allocation and scheduling module obtains the characteristic parameters of the current test task, and dynamically switches between four scheduling algorithms based on the characteristic parameters. The characteristic parameters include whether the graphics cards in the current batch are high-power graphics cards, whether the estimated test time for a single card exceeds the system's preset time threshold, and whether the device utilization rate balancing control switch is manually turned on.

[0012] As a preferred technical solution of the present invention, it also includes a fixture detection module, which is located in the unloading area. The fixture detection module is used to bind and archive the complete test data of the tested graphics card with the corresponding test fixture, and to calculate the overall test yield of each fixture based on the bound data.

[0013] As a preferred embodiment of the present invention, the fixture detection module is used to lock and mark fixtures that have not reached the preset test yield threshold, and the feeding module is used to stop feeding and conveying the marked test fixtures.

[0014] As a preferred embodiment of the present invention, each test unit has an independent power supply interface, an independent communication interface and independent control logic. Several independent test benches are evenly arranged in the test area of ​​the four test units. The test benches are used to connect with the test fixture to test the graphics card under test.

[0015] As a preferred embodiment of the present invention, the test bench is used to perform Linux Diag testing, Windows 3D testing, and I / O interface testing on the graphics card under test.

[0016] As a preferred embodiment of the present invention, the test bench is used to mark graphics cards whose test duration exceeds a preset threshold.

[0017] The beneficial effects of this invention are as follows: By setting up an allocation and scheduling module connected to each test unit and collecting real-time status, the system can dynamically allocate tasks according to the real-time load of each unit, thereby making full use of test resources. The setting of the double-layer circulating transport module realizes the physical isolation of feeding and recycling, allowing the upper layer to feed while the lower layer can recycle simultaneously, with feeding and recycling not interfering with each other. At the same time, the setting of four independent test units ensures that the remaining units can still operate normally when a single unit fails, thereby avoiding mutual interference between feeding and recycling and improving the fault tolerance, resource utilization and test smoothness of the production line. This solves the problem of low space utilization and low operating efficiency of graphics card test production lines in the existing fully automatic graphics card testing process. Attached Figure Description

[0018] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0019] Figure 1 This is a block diagram of the control system operation logic of the present invention. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0021] In existing fully automated graphics card testing production lines, graphics cards are typically transported along a single assembly line. Testing equipment is set up on both sides of the assembly line to test the graphics cards. Therefore, after one batch of graphics cards completes testing, it must wait for another batch to be loaded before the line can continue operating, resulting in a significant amount of idle waiting time. Furthermore, existing production lines typically treat the entire testing line as an inseparable unit. When the feeding and discharging lines share the same track, if congestion or abnormalities occur at the discharging end, the feeding end will be simultaneously obstructed, forcing the feeding module to pause. Conversely, feeding abnormalities will also affect discharging. Therefore, if a testing station malfunctions, the entire production line will be forced to stop operating. This assembly line design makes the testing process inflexible, results in long fault recovery times, and severely impacts production efficiency.

[0022] Please see Figure 1This embodiment provides a control system for fully automated graphics card testing, including a dual-layer circulating transport module, a loading module, four independently operating test units, and an allocation and scheduling module. The loading module is used to install the graphics card under test onto the test fixture and load it into the dual-layer circulating transport module. The four test units are used to test the functions of the graphics card under test. Each test unit is equipped with complete testing equipment, including a graphics card power supply, a signal generator, a video signal acquisition card, and a temperature sensor, etc., and can independently complete all test items of the graphics card. The dual-layer circulating transport module is used to transport the graphics card under test from the loading area to the test unit and to retrieve the tested graphics card from the test area to the unloading area. The allocation and scheduling module is connected to each test unit via TCP / IP protocol to collect the real-time occupancy status of the test stations, the number of graphics cards currently under test, and the operating status of each device, and to allocate test tasks according to a preset scheduling algorithm. The dual-layer circulating transport module transports the test fixture to the corresponding test unit for testing according to the test task.

[0023] The entire testing system of this invention consists of four independently operating test units, arranged in parallel. A double-layer circulating transport module connects the test stations to the loading module. Specifically, the loading module offers four transport paths for the graphics cards, and similarly, four paths are provided for unloading and recycling. This ensures that the loading and recycling paths have multiple parallel routes during the testing and recycling process, preventing the entire test line from shutting down due to a single test failure. Furthermore, the loading and recycling paths are arranged vertically. The layer-separated design ensures that each tested graphics card has its own independent transport path, without occupying the loading module's transport path. This physical separation of the loading and unloading / recycling paths improves space utilization and operational efficiency in the graphics card testing production line. Furthermore, the loading and unloading transport paths are cyclical, allowing for efficient testing. First, the loading module installs each graphics card onto a test fixture, a dedicated carrier that supports the graphics card and provides electrical connections. The fixture has an interface at its bottom for docking with the test bench. After installation, the loading module feeds the graphics card and fixture together to the upper feed line of the double-layer cyclic transport module. At this time, the allocation and scheduling module collects the operating status of each test device in the four test units in real time through the TCP / IP protocol, and calculates which test unit the graphics card should be allocated to according to the preset scheduling algorithm. The allocation instruction is then sent to the double-layer circulation transport module. The double-layer circulation transport module transports the test fixture to the corresponding test unit according to the instruction and the specified loading and conveying path. The test unit tests the various functions of the graphics card. After the test is completed, the test unit transports the fixture to the recycling path. The graphics card is then recycled from the test area to the unloading area through the recycling path. The test content of the graphics card is then acquired. The graphics card is classified by the acquired content, and then the graphics card is removed from the fixture. A new graphics card to be tested is then installed on the fixture for a new round of testing and circulation.

[0024] By setting up an allocation and scheduling module connected to each test unit and collecting real-time status data, the system can dynamically allocate tasks according to the real-time load of each unit, thus making full use of test resources. The dual-layer cyclic transport module and the four independently operating test units allow the graphics cards to be transported to different test units during the testing process, preventing the other units from continuing to operate normally when a single unit fails. At the same time, it avoids interference between material input and output, thereby improving the fault tolerance, resource utilization, and test smoothness of the production line. Moreover, the upper layer material feeding process can be synchronized with the lower layer recycling process, with material feeding and recycling not interfering with each other, achieving physical isolation between material feeding and recycling. This solves the problem of low space utilization and low operating efficiency in the existing fully automated graphics card testing production line.

[0025] To improve the conveying efficiency of the dual-layer circulating transport module while avoiding the interference risk of shared inlet and outlet tracks, this embodiment includes an upper-layer feeding line module and a lower-layer recycling line module, connected by a vertical lifting module. The feeding direction of the upper-layer feeding line module is opposite to the recycling direction of the lower-layer recycling line module, forming a closed-loop circulation path. In actual operation, the test fixture carrying the graphics card enters the upper-layer feeding line from the loading area and is transported to the corresponding test unit according to the assigned test task. The test fixture is connected to the test unit to test the graphics card. After the graphics card in the test unit completes all test items, the test unit transports the test fixture to the vertical lifting module, which then moves the test fixture to the lower-layer recycling line. The lower-layer recycling line runs in the opposite direction to the upper-layer feeding line, transporting the tested graphics card back to the upper layer. At the front end of the production line, in the unloading area, vertical lifting modules are located in the loading area and various testing areas. The vertical lifting modules in the loading area are used to lift the completed test fixtures from the lower recycling line to the loading area, remove the completed graphics cards from the test fixtures, and then reinstall the graphics cards to be tested onto the test fixtures to achieve fixture recycling. The vertical lifting modules in the testing areas are used to feed the test fixtures to be tested from the upper feeding line to the corresponding test units for testing, and at the same time, lower the completed test fixtures from the layer where the test units are located to the lower recycling line for recycling. The vertical lifting modules connect the upper feeding, lower recycling, and testing stations, realizing the separation of feeding and recycling to different layers and forming a closed loop. This allows feeding and recycling to be completely parallel and do not interfere with each other, avoiding the situation where the entire testing production line stops when a test unit fails.

[0026] Since different specifications of graphics cards are tested under different standards, in order to improve production line efficiency and test quality, the allocation and scheduling module in this embodiment has four built-in scheduling algorithms: shortest path first algorithm, balanced round-robin allocation algorithm, regional load first algorithm, and device idle first algorithm. Shortest Path First (SCF) Algorithm: When a graphics card enters the production line from the loading area, the allocation and scheduling module obtains the location information of all available test stations in the four test units, calculates the total physical transport distance from the loading station to the test station, and then from the test station to the inspection area, and selects the test point with the shortest total distance for allocation. The total transport distance of the SCF algorithm is defined as the total physical transport distance from the loading station to the target test station and then to the inspection area. The allocation and scheduling module pre-stores the physical distance matrix between each test station in the four test units. This matrix is ​​generated based on the CAD layout of the production line and records the transport path length between any two locations in the production line. When a graphics card needs to be allocated, the allocation and scheduling module first obtains a list of all currently available (idle and fault-free) test stations. Then, for each available station, it looks up the distance d1 from the loading station to the station and the distance d2 from the station to the inspection area from the distance matrix, calculates the total distance D = d1 + d2, and finally selects the station with the smallest D value as the target test point. For example, if test unit 1 is close to the loading area and its test station is empty, the graphics card will be preferentially assigned to test unit 1.

[0027] Balanced Round Robin Allocation Algorithm: The balanced round robin allocation algorithm maintains a global round robin pointer, which cycles through test units 1, 2, 3, and 4. Each time a graphics card needs to be allocated, the system assigns the graphics card to the test unit currently pointed to by the pointer, and then moves the pointer to the next unit. For example, when the first graphics card is assigned to unit 1, the second to unit 2, the third to unit 3, the fourth to unit 4, the fifth to unit 1 again, and so on, the system ensures that each test unit receives approximately the same amount of testing tasks per unit of time.

[0028] Regional load priority algorithm: The number of graphics cards executing test tasks in each test unit is collected in real time via TCP / IP protocol. The current number of cards under test in each unit is compared, and the unit with the fewest cards is selected for allocation. When multiple units have the same number of cards under test, a secondary selection can be made based on the temperature or historical load of each unit. For example, if there are 5 graphics cards under test in unit 1, 3 in unit 2, 4 in unit 3, and 2 in unit 4, then the new graphics card will be preferentially allocated to unit 4.

[0029] The device idle priority algorithm collects the operating status (running / idle / faulty) of all test devices in each test unit in real time through the TCP / IP protocol, counts the number of idle devices in each unit, and selects the unit with the most idle devices for allocation.

[0030] Therefore, when a graphics card under test needs to be tested, the allocation and scheduling module obtains the current status of all test stations and the occupancy status of each section of the upper feeding line and lower recycling line. Based on the obtained status of the available stations and the allocation method of the four scheduling algorithms, the graphics card under test is transported to the designated test station through the upper feeding line module. During the transportation process, it will not interfere with the graphics card that is already in the process of transportation. That is, if the newly allocated graphics card would interfere with the graphics card that is already in the process of transportation, the allocation and scheduling module will reassign it to a new test station to avoid it. This ensures that the algorithm switching will not cause the running test task to be interrupted, and avoids the loss of test data or equipment abnormality caused by the switching.

[0031] All of the above algorithms must be switched under unified pre-constraints, namely, they must not interfere with the normal loading, unloading and testing processes of other graphics cards. For example, if Unit 1 is currently testing a batch of graphics cards and it is expected to take another 10 minutes to complete, the switched algorithm will not assign the new graphics card to the workstation currently in use in Unit 1, nor will it assign the graphics card to the conveyor path that is about to be occupied.

[0032] Because different batches of graphics cards vary in power consumption, test duration, and customer requirements, different scheduling algorithms are used to maximize production line efficiency under different conditions. In this embodiment, the dynamic switching of the four scheduling algorithms is as follows: the shortest path first algorithm is the default scheduling algorithm under normal production conditions; the balanced round-robin allocation algorithm is switched on when single-card testing takes a long time and all testing equipment in all areas needs to be fully utilized; the area load first algorithm is switched on in high-power graphics card production scenarios; and the device idle first algorithm is switched on under the requirement of balanced equipment utilization rate management in each area. The allocation and scheduling module obtains the characteristic parameters of the current test task and dynamically switches between the four scheduling algorithms based on these characteristic parameters. The characteristic parameters include the current batch of graphics cards... Whether the card is a high-power graphics card, whether the estimated test time of a single card exceeds the system's preset time threshold, and whether the equipment utilization rate balancing control switch is manually turned on are all factors to consider. In actual production, when a batch of graphics cards to be tested needs to be tested, the requirements for the graphics cards to be tested are first input, and then the allocation and scheduling module obtains the characteristic parameters of the current test task and determines whether the characteristic parameters of the current test task meet one of the following conditions: whether it is a high-power graphics card, whether the estimated test time of a single card exceeds the system's preset time threshold, and whether the equipment utilization rate balancing control switch is manually turned on. If none of these conditions are met, then shortening the total turnaround time of the graphics card from entering the production line to completing the test is the most important goal. At this time, the shortest path first algorithm is the most suitable, which is the standard configuration under most daily working conditions. When the production plan changes, specifically when the estimated test time for a single card exceeds the system's preset time threshold, if the shortest path priority algorithm is still used, units closer to the loading area will continue to be assigned tasks while units further away will remain idle for extended periods. In this case, the allocation and scheduling module switches to a balanced round-robin allocation algorithm to ensure a balanced workload across the four test units, preventing situations where Unit 1 is congested while Unit 4 remains idle for a long time. It also prevents a large number of graphics cards from completing tests simultaneously in the same area, causing material unloading congestion. However, if the characteristic parameter is a high-power graphics card, using the shortest path priority algorithm would lead to a large number of graphics cards being concentrated in one area, resulting in excessively high local temperatures, affecting test accuracy and equipment lifespan. In this case, the allocation and scheduling module switches to a regional load priority algorithm, distributing high-power graphics cards to different test units to prevent multiple high-power graphics cards from running simultaneously in the same area, causing localized heat accumulation and overloading the cooling system. When the equipment utilization rate balancing control switch is manually enabled, the allocation and scheduling module uses an idle-priority algorithm for allocation based on the customer's specific requirements for balanced equipment utilization rate control in each area.

[0033] By executing different scheduling algorithms according to different test scenarios and through a dynamic switching mechanism, the system can select the most suitable scheduling strategy based on the characteristics of the current test task, thereby maximizing production line efficiency.

[0034] In mass production, each graphics card corresponds to unique test data, while each fixture carries the testing tasks for multiple graphics cards. Therefore, when a batch of graphics cards fails in mass testing, it is only possible to determine which testing station or time period caused the problem based on the test data corresponding to the graphics card, but it is impossible to determine which fixture itself malfunctioned. Therefore, to avoid the inability to accurately troubleshoot when fixtures malfunction, this embodiment also includes a fixture detection module. The fixture detection module is located in the unloading area. The fixture detection module is used to bind and archive the complete test data of the tested graphics cards with the corresponding test fixtures, and to calculate the overall test yield of each fixture based on the bound data. In actual operation, the fixtures that have completed testing return to the unloading area through the lower recycling line. The inspection area overlaps with the unloading area. At the entrance of the inspection area, a barcode scanner is installed to scan the product barcode on the surface of the graphics card to obtain its unique identification code. At the same time, another barcode scanner scans the fixture code on the surface of the test fixture that carries the graphics card to obtain the fixture's unique number. The fixture inspection module binds the graphics card identification code, the fixture number, and all test data generated by the graphics card during the testing process, including the test status of each test item, test time, and parameter curves recorded during the test, to form a complete test record and archive it to the database. Through the dual scanning and binding archiving of the graphics card and the fixture, the test data of each graphics card can be accurately matched with the specific fixture that carries it, so that the problematic fixture can be located when a defect rate occurs.

[0035] Because the wiring harness of the test fixture needs to be frequently plugged and unplugged, it is a high-loss and easily damaged component. In actual production, a single fixture may be plugged and unplugged hundreds of times a day. The metal contacts of the wiring harness will gradually wear and oxidize, increasing the contact resistance and causing test signal attenuation or interruption, leading to various test anomalies. These anomalies are not problems with the graphics card itself, but rather false defects caused by fixture wear. Therefore, to avoid the continuous use of severely worn fixtures in production, resulting in a constant stream of false defects and wasted testing capacity, in this embodiment, a fixture detection module is used to detect fixtures that have not reached the preset test yield threshold. The locking and marking module is used to stop feeding test fixtures marked with tags. Before each installation of the graphics card under test onto the fixture, the loading module first checks the locking status of the fixture. If the fixture is locked and marked, the loading module skips the fixture and does not feed it, instead using other unlocked fixtures. Locked fixtures need to be manually removed for inspection. The inspection personnel check whether the fixture's wiring harness is worn, whether the probes are oxidized, and whether the positioning structure is deformed. After the repair and verification are completed, the locking mark is manually removed, and the fixture can be put back into testing use.

[0036] To prevent the entire test line from being forced to shut down when a test station fails, in this embodiment, each test unit has an independent power supply interface, an independent communication interface, and independent control logic. Several independent test benches are evenly arranged in the test area of ​​the four test units. The test benches are used to connect with the test fixture to test the graphics card under test. By giving each test unit an independent rack structure, an independent power distribution box, an independent PLC controller, and an independent network interface, and by arranging the test benches in each test unit in multiple layers, with multiple test stations in each layer, each station is equipped with an electrical interface and positioning structure for interfacing with the test fixture, when a test unit fails, because the unit has independent power supply, communication, and control, the fault is isolated within that unit and will not spread to other units through a shared power supply or network. The other three test units continue to operate normally and continue to receive test tasks assigned by the allocation and scheduling module. Moreover, the repair of the faulty unit can be carried out without interrupting production, thereby significantly reducing the production capacity loss due to unplanned downtime.

[0037] To better test the functionality of graphics cards and ensure that the cards leaving the factory are fully functional and meet performance standards, in this embodiment, a test bench is used to perform Linux Diag testing, Windows 3D testing, and I / O interface testing on the graphics card under test. These three tests cover the main functions of the graphics card: basic functions (Linux Diag), graphics performance (Windows 3D), and interface integrity (I / O interface). The reason for performing all three tests on the test bench is that defects in any functional dimension of the graphics card may render the product unusable for end users. Without Linux Diag testing, hidden faults at the basic hardware level cannot be detected; without Windows 3D testing, the performance of the graphics card in real-world graphics application scenarios cannot be verified; and without I / O interface testing, poor interface contact or signal integrity issues will result in the card entering the market.

[0038] In automated testing production lines, testing stations are valuable resources. If a graphics card's testing time exceeds the normal value due to a malfunction or anomaly and is not detected in time, the station will be occupied ineffectively for a long period, preventing subsequent graphics cards from entering the station and causing production capacity loss. Therefore, to improve the utilization rate of testing stations and the overall production capacity of the line, in this embodiment, the testing station is used to mark graphics cards whose testing time exceeds a preset threshold. In actual operation, the system starts a timer when each graphics card enters the testing station to begin testing. Different types of test items correspond to different preset timeout thresholds. For example, the preset threshold for Linux Diag testing is 5 minutes, the preset threshold for Windows 3D testing is 15 minutes, and the preset threshold for I / O interface testing is 3 minutes. When the actual execution time of a test exceeds the corresponding preset threshold, the test bench automatically stops the test, marks the graphics card as having timed out, and pushes an alarm notification to the central control console and the mobile terminals of on-site management personnel. Graphics cards marked as having timed out are not directly judged as defective and sent to the lower recycling line for recycling. Instead, they are marked and await manual intervention. After receiving the alarm, on-site personnel go to the corresponding test station to check the status of the graphics card. This is because a test timeout may be caused by a hardware failure of the graphics card itself causing the test program to freeze, a poor contact of the test fixture causing communication interruption, or an abnormality in the test program itself. Personnel take appropriate measures according to the specific situation: if it is a graphics card failure, it is sent to the defective product channel; if it is a fixture problem, the fixture is replaced and the test is repeated; if it is a program abnormality, the test is restarted.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A control system for fully automated testing of graphics cards, characterized in that: The system includes a dual-layer circulating transport module, a loading module, four independently operating test units, and an allocation and scheduling module. The loading module is used to install the graphics card under test onto the test fixture and load it into the dual-layer circulating transport module. The four test units are used to test the functionality of the graphics card under test. The dual-layer circulating transport module is used to transport the graphics card under test from the loading area to the test units and to retrieve the tested graphics card from the test area to the unloading area. The allocation and scheduling module is connected to each test unit and is used to collect real-time status information of the test equipment and allocate test tasks according to a preset scheduling algorithm. The dual-layer circulating transport module transports the test fixture to the corresponding test unit according to the test task.

2. The control system for fully automated testing of graphics cards according to claim 1, characterized in that: The dual-layer circulating transport module includes an upper feeding line module and a lower recycling line module. The upper feeding line module and the lower recycling line module are connected by a vertical lifting module located in the testing area. The vertical lifting module is used to move the completed test fixture from the upper feeding line module to the lower recycling line module. The feeding direction of the feeding line module is opposite to the recycling direction of the lower recycling line module, forming a closed-loop circulation path.

3. The control system for fully automated testing of graphics cards according to claim 1, characterized in that: The scheduling algorithm includes: The shortest path first algorithm is used to assign graphics cards to available test points with the shortest transfer distance from the loading area to the inspection and finished product scanning process. The balanced round-robin allocation algorithm is used to sequentially and cyclically allocate the graphics cards to be tested in each test area, so as to evenly distribute the overall test task. The regional load priority algorithm is used to prioritize scheduling graphics cards to test areas with fewer graphics cards currently under test and lower load. The device idle priority algorithm is used to prioritize scheduling graphics cards to test areas with more idle test devices; The allocation and scheduling module obtains the current status of all test stations and the occupancy status of each section of the conveying path of the upper feeding line and the lower recycling line. The allocation and scheduling module selects the test station whose conveying path is not occupied from the test stations in the idle state as the target test station.

4. The control system for fully automated testing of graphics cards according to claim 3, characterized in that: The dynamic switching methods for the four scheduling algorithms are as follows: The shortest path first algorithm is the default scheduling algorithm under normal production conditions. The balanced polling allocation algorithm is switched on and enabled when the single-card test takes a long time and it is necessary to make full use of all area test equipment. The regional load priority algorithm is switched on and enabled in high-power graphics card production scenarios. The equipment idle priority algorithm is switched on and activated under the requirement of balanced control of equipment utilization rate in each region. The allocation and scheduling module obtains the characteristic parameters of the current test task, and dynamically switches between four scheduling algorithms based on the characteristic parameters. The characteristic parameters include whether the graphics cards in the current batch are high-power graphics cards, whether the estimated test time for a single card exceeds the system's preset time threshold, and whether the device utilization rate balancing control switch is manually turned on.

5. The control system for fully automated testing of graphics cards according to claim 1, characterized in that: It also includes a fixture inspection module, which is located in the unloading area. The fixture inspection module is used to bind and archive the complete test data of the tested graphics card with the corresponding test fixture, and to calculate the overall test yield of each fixture based on the bound data.

6. The control system for fully automated testing of graphics cards according to claim 5, characterized in that: The fixture detection module is used to lock and mark fixtures that have not reached the preset test yield threshold, and the feeding module is used to stop feeding and conveying the marked test fixtures.

7. The control system for fully automated testing of graphics cards according to claim 1, characterized in that: Each of the test units has an independent power supply interface, an independent communication interface, and independent control logic. Several independent test benches are evenly arranged in the test area of ​​the four test units. The test benches are used to connect with the test fixture to test the graphics card under test.

8. The control system for fully automated testing of graphics cards according to claim 7, characterized in that: The test bench is used to perform Linux Diag testing, Windows 3D testing, and I / O interface testing on the graphics card under test.

9. The control system for fully automated testing of graphics cards according to claim 7, characterized in that: The test bench is used to mark graphics cards whose test duration exceeds a preset threshold.

Images