Electrical computers and digital data processing systems: input/ – Intrasystem connection – Bus access regulation
Reexamination Certificate
1998-10-14
2002-08-13
Sheikh, Ayaz R. (Department: 2781)
Electrical computers and digital data processing systems: input/
Intrasystem connection
Bus access regulation
C710S120000, C710S120000
Reexamination Certificate
active
06434649
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to a data processor, and more specifically to a data transfer arrangement mechanism employed to transfer data to various components of the data processor.
BACKGROUND OF THE INVENTION
In many data processing chip sets data is transferred from one or many processors to memory devices and input/output, I/O, subsystems, or other chip components known as functional units, via an appropriate bus structure. Typically, the bus structure includes a processor bus, a system bus and a memory bus. Thus, when there is a memory operation wherein data is required to be moved to or from a memory location to a processor, the system bus would cease to operate until the data movement from the memory location to the processor is completed. Similarly, when there is a data movement from an external device to a memory location, the processor bus would cease to operate until the data is moved to its intended location.
In order to alleviate the under utilization of bus subsystems as described above, U.S. Pat. No. 5,668,965 issued on Sep. 16, 1997, teaches the use of a controller that forms a three-way connection of three kinds of buses including a processor bus linked to at least one processor, a memory bus connected to a main memory, and a system bus linked to at least one connected device such as an input/output, I/O, device, thereby establishing interconnections between various buses. The controller includes data path switch means for transferring control signals and addresses through the control and address buses respectively of the three kinds of buses, and for generating a data path control signal to be supplied to the data switch means.
This arrangement allows the use of the buses on an independent basis. For example, when a processor on the processor bus conducts a processor/main memory access to access the main memory on the memory bus, data is transferred only via the processor and memory buses, allowing the system bus to operate independently.
However, the arrangement disclosed in the '965 patent does not provide for a priority based data movement. Furthermore, it does not disclose a mechanism to handle data transfers between endpoints that exhibit mismatched bandwidth requirements.
Additionally, conventional data movement arrangements have failed to address application-specific requirements. For example, when a data processor is employed for handling graphical images and displaying them on a screen, considerable throughput efficiency may be gained by taking into account the memory address patterns that are inherent with such graphical images.
Another disadvantage with conventional systems is that the resources employed by the data movement arrangements cannot be flexibly specified based on a corresponding data transfer between two end points. For example, some data movement arrangements employ fixed buffers to accommodate separate input/output, I/O, data transfers.
Thus, there is a need for a data movement arrangement that overcomes the advantages discussed above , and specifically accommodates data transfers for an integrated media processor chip set that contains various system components such as processors, data cache, three dimensional graphics units, memory and input/output devices.
SUMMARY OF THE INVENTION
In accordance with one embodiment of the invention, an information processing system includes a plurality of modules including a processor, a data cache memory, a main memory and a plurality of I/O devices. A data transfer switch is provided to handle a plurality of data transfer operations simultaneously on a priority based scheduling. A split level bus allows for such simultaneous data transfer operations between the processor, main memory and I/O devices. The split level bus comprises a first request bus having a request bus arbiter for receiving read and write requests from each one of said plurality of modules.
A processor memory bus is configured to receive address and data information from a predetermined number of said modules including said processor. The processor memory bus includes a data bus arbiter for receiving data read and write requests from each one of the predetermined number of modules coupled to it. Furthermore, an internal memory bus is configured to receive address and data information from a predetermined number of the modules including the memory and I/O devices. The memory bus has a data bus arbiter for receiving data read and write requests from each one of the predetermined number of modules coupled to the internal memory bus.
For data transfers that originate from a module coupled to one of the data buses and is intended for a module that is coupled to the other data bus a transceiver system is employed which is coupled to both processor memory bus and the internal memory bus for transferring data between the two data buses.
In accordance with another embodiment of the invention, the request bus arbiter is configured to receive a plurality of read and write requests each having a specifiable priority level, wherein the requests are served in a descending order of priority.
In accordance with yet another embodiment of the invention, a data streamer is employed in the information processing system, which allows for simultaneous data transfers between modules that may have disparate data transfer rates. The data streamer comprises a channel state memory configured to store a first allocated channel information corresponding to a data transfer operation from a source module to the data streamer. The channel state memory is further configured to store a second allocated channel information corresponding to the data transfer operation from the data streamer to a destination module. The data streamer further includes a buffer memory allocated to the data transfer operation for receiving data provided by the source module in accordance with the first allocated channel information and providing the received data to the destination module in accordance with the second allocated channel information.
In accordance with another embodiment of the invention, the channel state memory stores information corresponding to a plurality of data transfer operations between the modules. Furthermore, a buffer memory is allocated for each one of the data transfer operations and the size of the buffer memory variably changes in accordance with the size of data in a corresponding data transfer operation.
In accordance with yet another embodiment of the invention, the data transfer operations occur in accordance with a program referred to as a channel descriptor. Thus, transfers to the buffer memory occur in accordance with a corresponding channel descriptor, as well as transfers from buffer memory to a destination module occur in accordance with a corresponding channel descriptor.
The data streamer also conducts a data cache operation with its data transfer operations for a data cache having a coherent allocation policy. The data streamer may also conduct the data cache operation for a data cache having a coherent no-allocation policy, or having a non-coherent no-allocation policy.
REFERENCES:
patent: 4951232 (1990-08-01), Hannah
patent: 5005117 (1991-04-01), Ikumi
patent: 5010515 (1991-04-01), Torborg, Jr.
patent: 5111425 (1992-05-01), Takeuchi et al.
patent: 5276836 (1994-01-01), Fukumaru et al.
patent: 5301351 (1994-04-01), Jippo
patent: 5303339 (1994-04-01), Ikuma
patent: 5392392 (1995-02-01), Fischer et al.
patent: 5412488 (1995-05-01), Ogata
patent: 5440752 (1995-08-01), Lentz et al.
patent: 5442802 (1995-08-01), Brent et al.
patent: 5448702 (1995-09-01), Garcia et al.
patent: 5461266 (1995-10-01), Koreeda et al.
patent: 5483642 (1996-01-01), Okazawa et al.
patent: 5493644 (1996-02-01), Thayer et al.
patent: 5506973 (1996-04-01), Okazawa et al.
patent: 5561820 (1996-10-01), Bland et al.
patent: 5594882 (1997-01-01), Bell
patent: 5630094 (1997-05-01), Hayek et al.
patent: 5646651 (1997-07-01), Spannaus et al.
patent: 5655131 (1997-08-01), Davies
patent: 5655151 (1997-08-01), Bowes et al.
patent: 5664218 (1997-09-01)
Baker David
Basoglu Christopher
Cutler Benjamin
Gervasio Gregorio
Lee Woobin
Hitachi , Ltd.
Poole John
Sheikh Ayaz R.
Sofer & Haroun LLP
Strans X. Chung
LandOfFree
Data streamer does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data streamer, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data streamer will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2954743