Computer graphics processing and selective visual display system – Computer graphic processing system – Plural graphics processors
Reexamination Certificate
1999-01-08
2001-12-11
Chauhan, Ulka J. (Department: 2671)
Computer graphics processing and selective visual display system
Computer graphic processing system
Plural graphics processors
C345S503000
Reexamination Certificate
active
06329996
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to systems for computer graphics. More specifically, the present invention includes a method and apparatus for synchronizing the sequential operation of a group of graphics pipelines.
BACKGROUND OF THE INVENTION
Computer systems (and related devices) typically create three-dimensional images using a sequence of stages known as a graphics pipeline. During early pipeline stages, images are modeled using a mosaic-like approach where each image is composed of a collection of individual points, lines and polygons. These points, lines and polygons are know as primitives and a single image may require thousands, or even millions, of primitives. Each primitive is defined in terms of its shape and location as well as other attributes, such as color and texture.
The graphics pipeline maps each primitive into a memory storage device know as a frame buffer. Each storage location within the frame buffer defines one pixel within the image being produced. The graphics pipeline performs the mapping process by determining which pixels (i.e., which frame buffer storage locations) are included within each primitive. Each pixel is then initializes to reflect the attributes of the primitive, or primitives in which it is included. In many cases, the graphics pipeline will further modify the pixel values in the frame buffer to apply texture, lighting and other effects to the graphics primitives.
The task of rendering primitives to pixels can be very time consuming. This is especially true for complex images that include many primitives or require complex lighting, shading or other effects. The time consumed transforming primitives becomes problematic for applications, such as flight simulators and virtual reality environments, where rapid image generation is required.
One method for improving the speed of the rendering process is to use multiple graphics pipelines. For this method, the task of rendering primitives is subdivided so that each pipeline performs some portion of the total task. One way of subdividing the rendering task is to assign a fixed portion of each image to a different pipeline. The pipelines operate in parallel and each pipeline processes the primitives that are included in its image portion.
The parallel output of the various pipelines is combined for display. To be effective, this combination must be accomplished seamlessly, as though the image were being created by a single graphics pipeline. In general, this seamless integration requires that the output of each pipeline be synchronized. Otherwise, the pipelines produce output at somewhat random intervals. This can make the resulting images appear to be inaccurate or even incoherent.
For some systems, the synchronization problem may be overcome through the use of a global ready signal. SILICON GRAPHICS INC. provides a signal of this type, known as SWAP_READY, for use with certain types of SGI computers. The SWAP_READY signal is physically connected to each graphics pipelines that is to be synchronized. Systems that are ready to output their image portion assert local ready signals. SWAP_READY functions as a logical AND of these local ready signal. As a result, SWAP_READY becomes asserted when all systems are ready to output their image portions. The graphics pipelines sample the state of the SWAP_READY signal on a periodic, synchronized basis. The pipelines output their image portions when the discover that SWAP_READY is high.
A second way to combine multiple graphics pipelines is to subdivide the rendering task so that the pipelines operate in round-robin or sequential fasion. For this method, each pipeline contributes a succeeding image to a sequence of images. The sequential method has certain advantages when compared to the method where each image is partitioned between pipelines. One of these advantages is the ability to use less powerful graphics pipelines to produce a sequence of images at an accelerated rate. This is especially beneficial for flight simulators and other real time environments.
To ensure image quality, sequential pipeline operation requires synchronization of the pipelines that are used to generate an image. Otherwise, there is a chance that the pipelines will output images out of order. For animation and simulation environments, this can cause the display to move unpredictably forward and backward in time. Unfortunately, a synchronization method, such as a global ready signal, is not available for systems using the sequential pipeline operation. As a result, complex software synchronization is required, increasing the difficulty of implementing systems of this type.
For some applications it is desirable to use a combination of both of the previously described techniques for combining multiple graphics pipelines. This allows images to be partitioned into regions with each region using one or more pipelines connected in sequence. The sequential operation of each group of one or more pipelines proceeds in parallel with the operation with the remaining pipeline groups. This provides a flexible method for arranging graphics pipelines to match the image generation requirements of specific applications. Unfortunately, the lack of a synchronization method for sequential pipeline operation makes combination of these systems with other graphics pipelines difficult.
Based on the preceding it may be appreciated that a need exists for an effective method for synchronizing multiple graphics pipelines that are configured to operate in a sequential fashion. A preferable solution to this problem would allow sequential graphics pipeline groups to operate in parallel with other pipeline groups.
SUMMARY OF THE INVENTION
An embodiment of the present invention includes a method and apparatus for synchronizing the operation of graphics pipelines. This method and apparatus allows graphics pipelines to be grouped for sequential operation and allows parallel execution between these pipeline groups. For a representative embodiment of the present invention, one or more graphics pipelines are connected as a daisy-chain sequence. This means that the output of each pipeline (except the last) is connected as an input to a subsequent pipeline. The output of the last pipeline is connected to a display device.
Each pipeline may be configured to operate in two different modes: local and pass-through. Pipelines operating in local mode output their own digital video data (i.e., the pixel values generated by the graphics pipelines). Pipelines operating in pass-through mode output the video data that they receive from preceding pipelines.
In operation, an application program configures the sequence of pipelines so that one pipeline operates in local mode. The remaining pipelines operate in pass-through mode. This causes the digital video output produced by the one pipeline to be passed through to the output device. At periodic intervals, the application program selects the next pipeline for operation in local mode. Shortly after the selected next pipeline becomes ready a swap event occurs. During the swap event, the selected next pipeline switches to local mode and the remaining pipeline either enter or stay in pass-through mode. In this way, the pipelines cooperate in sequential fashion to produce a sequence of images.
The timing of swap events is based on the state of a global ready signal. The global ready signal represents a logical AND of a set of local ready signals. One of these local ready signals is produced by the selected next pipeline. This local ready signal functions as the local ready signal for the entire sequence of graphics pipelines. The remaining local ready signals are provided by any graphics pipelines that are configured to work in parallel with the sequence of graphics pipelines. The global ready signal becomes asserted when each of these local ready signals is asserted. This means that the selected next pipeline, and all pipelines that are configured to work in parallel with the selected next pipeline, are ready to begin outputting video data. All of the
Arnaud Remi Simon Vincent
Bowen Andrew D.
Bowman James
Buchner Gregory C.
Chian Daniel T.
Chauhan Ulka J.
Silicon Graphics Inc.
Squire Sanders & Dempsey LLP
LandOfFree
Method and apparatus for synchronizing graphics pipelines does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for synchronizing graphics pipelines, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for synchronizing graphics pipelines will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2587416