Multiple parallel processor computer graphics system
An accelerated graphics processing subsystem that significantly increases the processing speed of computer graphics commands. The preferred embodiment of this invention presents a first-of-its-kind graphics processing subsystem that combines the processing power of multiple, off-the-shelf, video cards, each one having one or more graphic processor units. The video cards can be used without substantial modification. Under the preferred embodiment, each video card processes instructions for drawing a predetermined portion of the screen which is displayed to the user through a monitor or other visual output device. The invention harnesses the power of multiple video cards without suffering from the high bandwidth constraints affecting prior attempts at parallel graphics processing subsystems.
1. A method for accelerating the processing of graphics images by a computer for display on a single visual output device screen through use of a plurality of video cards, comprising the steps of:
intercepting graphics commands issued by an application;
generating multiple, modified graphics command streams;
assigning each of said multiple, modified graphics command streams to different video cards selected from said plurality of video cards;
processing at each of said video cards the assigned modified graphics command stream and generating an output signal;
synchronizing the output signals by said plurality of video cards;
combining the output signal from said plurality of video cards into a single graphics output signal, wherein said combining comprises sequentially routing selected portions of said output signals from said plurality of video cards, wherein routing is determined by a vertical refresh rate of said output signals from said plurality of video cards, a vertical resolution of said output signals from said plurality of video cards and a load balancing ratio assigned to each of said plurality of video cards; and
displaying said single graphics output signal on a single visual output device whereby said visual output device displays images defined by said intercepted graphics commands with accelerated performance.
2. The method of claim 1 wherein the output signal of each of said multiple, modified graphics command streams draws only a portion of a graphics screen.
3. The method of claim 2 wherein said intercepting and generating steps are performed by a graphics command replicator.
4. The method of claim 2 wherein said combining step is performed by a video merger hub.
5. The method of claim 4 wherein said video merger hub is comprised of:
a video switch;
a video switch controller;
a microcontroller; and
a video output.
6. The method of claim 5 wherein said video switch is controlled by said video switch controller.
7. The method of claim 6 wherein said video switch controller controls said video switch by triggering routing switches at appropriate intervals.
8. The method of claim 7 wherein said load balancing ratio is transmitted by said microcontroller to said video switch controller.
9. The method of claim 7 wherein said load balancing ratio is assigned by a user through software.
10. The method of claim 7 wherein said load balancing ratio is equal for each of said video cards.
11. The method of claim 7 wherein said load balancing ratio is based on each of said video cards's graphics throughput.
12. The method of claim 7 wherein said load balancing ratio is dynamically adjusted to maximize the throughput of said subsystem by utilizing a test feedback loop program which measures the load on each of said video cards and makes appropriate adjustments.
13. The method of claim 2 wherein said visual output device is an output device selected from the group consisting of: cathode ray tube displays, liquid crystal displays, plasma screen displays, projection displays, OLED displays, head-mounted displays and hybrids thereof.
14. The method of claim 2 wherein said generation of multiple, modified graphics command streams is accomplished by incorporating a clipping command in each stream.
15. The method of claim 14 wherein said clipping command is a 2 D clipping command.
16. The method of claim 14 wherein said clipping command is a 3 D clipping command.
17. The method of claim 2 wherein the sum of all said portions of said graphics screen combine to generate a full graphics screen.
18. The method of claim 17 wherein said portions of said graphics screen are non-overlapping.
19. The method of claim 17 wherein said portions of said graphics screen have overlapping regions.
20. The method of claim 2 wherein the sum of all said portions of said graphics screen combine to generate a partial graphics screen.
21. The method of claim 20 wherein said portions of said graphics screen are non-overlapping.
22. The method of claim 20 wherein said portions of said graphics screen have overlapping regions.
23. The method of claim 2 wherein each of said plurality of video cards is equipped with a single GPU.
24. The method of claim 2 wherein each of said plurality of video cards is equipped with a plurality of GPUs.
25. The method of claim 2 wherein said plurality of video cards is comprised of a combination of video cards equipped with a plurality of GPUs and video cards equipped with a single GPU.
26. The method of claim 2 wherein said synchronizing step is accomplished through use of a genlock mechanism.
27. The method of claim 2 wherein said synchronizing step is accomplished by designating the timing regulating device in one of said plurality of video cards as a master timing regulating device and designating timing regulating devices in the remainder of said plurality of video cards as slaves of said master timing regulating device.
28. The method of claim 27 wherein the timing reference sources for said master and slave timing regulating devices are timing reference sources selected from the group consisting of piezoelectric crystals, programmable crystals, and combinations thereof.
29. The method of claim 1 wherein said intercepted graphics commands are API commands and said multiple, modified graphics command streams are multiple, modified API command streams.
30. The method of claim 29 wherein said multiple, modified API command streams are each received by a separate instance of an API module and wherein each of said API module instances generates a command stream which is processed by a separate video card selected from said plurality of video cards.
31. A method for accelerating the processing of graphics instructions on a computer through use of a plurality of video cards, comprising the steps of:
intercepting graphics commands issued by an application;
generating multiple, modified graphics command streams;
assigning each of said multiple, modified graphics command streams to a different video cards selected from said plurality of video cards;
processing at each of said video cards the assigned modified graphics command stream and generating an output signal;
synchronizing the output signals by said plurality of video cards;
combining the output signals from said plurality of video cards into a single graphics output signal by receiving said output signals from said plurality of video cards and sequentially routing selected portions of said output signals from said plurality of video cards to said video output; and
displaying said single graphics output signal on a single visual output device;
wherein said routing is performed at appropriate intervals, and wherein said intervals are determined on the basis of a vertical refresh rate of said output signals from said plurality of video cards, a vertical resolution of said output signals from said plurality of video cards, and a load balancing ratio assigned to each of said plurality of video cards.
32. A method for accelerating the processing of graphics images by a computer for display on a single visual output device screen through use of a plurality of video cards, comprising the steps of:
intercepting graphics commands issued by an application;
generating multiple, modified graphics command streams;
assigning each of said multiple, modified graphics command streams to different video cards selected from said plurality of video cards;
processing at each of said video cards the assigned modified graphics command stream and generating an output signal;
synchronizing the output signals by said plurality of video cards;
combining the output signal from said plurality of video cards into a single graphics output signal with a video merger hub comprising a video switch, a video switch controller, and a microcontroller, wherein the video switch controller controls the video switch based at least on inputs received from the plurality of video cards, wherein said combining is performed by routing and said routing is performed at appropriate intervals, and wherein said intervals are determined on the basis of a vertical refresh rate of said output signals from said plurality of video cards, a vertical resolution of said output signals from said plurality of video cards, and a load balancing ratio assigned to each of said plurality of video cards;
displaying said single graphics output signal on a single visual output device whereby said visual output device displays images defined by said intercepted graphics commands with accelerated performance.