Memory Mapped Interface Between Host Computer and Graphics System

This invention relates in general to the field of graphics systems. More particularly, the present invention relates to apparatus and method for graphics software development.

The computer age has brought forth microprocessors that can be accommodated in one integrated circuit package or on a chip. Some of these microprocessors serve general purpose functions that are capable of executing a myriad of software, such as a central processing unit in a personal computer. Certain other microprocessors have specialized applications, such as those which compute the Fourier Transform and the like.

One specialized microprocessor is the processor dedicated to generation of graphics. It may have a particular architecture, specialized logic, or micro instructions that facilitates the production and display of graphics designs.

One such graphics processor is the TM34020 manufactured by Texas Instruments Incorporated of Dallas, Tex. The TM34020 features specialized hardware and an instruction set that provide special graphics operations. However, in order to thoroughly exploit the features of the TM34020, a software development support tool must be provided to users who wish to develop application graphics software for the TM34020 graphics system processor.

Accordingly, a need has arisen for apparatus and method for graphics software development in support of a specialized graphics processor.

In accordance with the present invention, apparatus and a method for are provided which substantially eliminate or reduce disadvantages and problems associated with prior circuits.

In one aspect of the present invention, an apparatus for graphics software development for execution in a graphics processor is provided. The graphics processor resides in a host computer, which has a system bus and a video graphics array (VGA) for video enhanced display. The apparatus comprises memory mapping means coupled to the system bus for providing interface between the host computer and the apparatus, and a memory storage circuit coupled to the graphics processor for providing local memory space for storing data and access to the stored data. A video pass through means further allows either the output of the memory storage circuit or the output of the VGA to be viewed on one monitor.

In another aspect of the present invention, a method for interface between a host computer and a processor residing therein is provided. The method comprises the steps of determining the address of a block of predetermined amount of memory space in the system bus memory and storing the determined address. Thereafter, the local memory space is mapped into the system bus memory space in response to the determined address stored in the base register.

In yet another aspect of the present invention, a method for video pass through in a graphics system having a video graphics array (VGA) and local memory space is provided. The method comprises the steps of connecting a monitor to a palette circuit, and multiplexing one of the output from the VGA and the local memory space to the palette circuit.

An important technical advantage of the present invention provides apparatus and method for graphics software development in support of a specialized graphics processor.

Another important technical advantage of the present invention provides for apparatus and method for memory mapped interface.

Yet another important technical advantage of the present invention provides for apparatus and method for using one monitor to display both VGA and video data.

For a better understanding of the present invention, reference may be made to the accompanying drawings, in which:

FIG. 1 is a top level block diagram of the preferred embodiment of the present invention; and

FIG. 2 is a more detailed block diagram of the present invention.

With reference to the drawings, FIG. 1 illustrates a preferred embodiment of apparatus for graphics software development, indicated generally at 10 and constructed according to the teaching of the present invention. Graphics software apparatus 10 is typically embodied in a printed circuit board coupled to the host central processing unit (CPU; not shown) of a personal computer (not shown) via an expansion slot connection (not shown).

Apparatus 10 is therefore coupled to a system bus 12, such as an ISA bus in an IBM personal computer, for communication with the host CPU. A host interface 14 is coupled directly with system bus 12 and a graphics processor 18. A coprocessor 16 is further connected to graphics processor 18. Both graphics processor 18 and coprocessor 16 are coupled to a local bus 20. Graphics processor 18 and coprocessor 16 may be implemented with integrated circuits TM34020 and TM34082, both manufactured by Texas Instruments.

A connector 22 which provides coupling between graphics processor 18 and an emulator (not shown) is coupled to graphics processor 18. Additionally, a clock circuit 24 is coupled to graphics processor 18 and a number of other circuit blocks to be described below.

Local bus 20 is coupled to memory components DRAM (dynamic random access memory) 26 and VRAM (video random access memory) 28. A connection between DRAM 26 and VRAM 28 is also provided. A logic circuit 30 is further coupled to VRAM 28, which is in turn connected to a VGA connector 36. Logic circuit 30 is also coupled to a palette circuit 32. Palette circuitry 32 is also coupled to local bus 20 and clock circuit 24. A video connector 34 is coupled to palette circuit 32 and clock circuit 24.

In operation, host interface 14 provides handshaking, address decoding, and a bidirectional path for the transfer of data between system bus 12 and apparatus 10. In particular, graphics processor 18 transmits and receives data to and from the host CPU via host interface 14. Clock circuit 24 generates and provides clock and timing signals of predetermined frequencies to graphics processor 18. Graphics processor 18 and coprocessor 16 may also communicate with an emulator board via emulator connector 22. DRAM 26 and VRAM 28 provide storage and random access of stored data to graphics processor 18, coprocessor 16 and palette circuit 32 (via logic circuit 30). VGA connector 36 provides a path between apparatus 1? and a conventional VGA board (not shown). Palette circuit 32 may receive video data from VRAM 28 or VGA via logic circuit 30. A video monitor may be coupled to video connector 34 to display the graphics images.

Referring to FIG. 2, where like numerals refer to like elements, host interface 14 is shown having several functional components. An I/O (input/output) data buffer 40 is coupled to system bus 12. I/O data buffer 40, in turn, is coupled to a status register 41, a map A register 42, a map B register 44, an interface register 46, and a base register 48. Map A and B registers 42 and 44 are also coupled to graphics processor 18.

Data transceivers 50 and host control logic 52 are also coupled to system bus 12. Host control logic 52 in turn has a control data path to graphics processor 18. A set of bus address latches 54 are coupled to system bus 12 and an address comparator 56. Address comparator 56 is also coupled to host control logic 52 and base register 48.

An address decoder 58 is coupled between graphics processor 18 and DRAM 26. Also coupled to local bus 20 is a hardware space register 60, which is also coupled to status register 41 and clock circuit 24.

Logic circuit 30 contains 4-bit shift registers 62, coupled to VRAM 28, and 8-bit shift registers 64 also coupled to VRAM 28. Further included is a multiplexor (MUX) 66, the inputs of which are coupled to the serial outputs of both shift registers 62 and 64, and to VGA connector 36. The output of MUX 66 is coupled to palette circuit 32.

In operation, host interface circuits 14 provide a control and data path for accesses to and from system bus 12 and support either 8-bit or 16-bit I/O (input/output) data transfers.

Host interface 14 includes I/O data buffers 40, which are adapted to receive and transmit I/O read and write data between system bus 12 and apparatus 10. Similarly, bus address latches 54 receive and latch the current bus cycle address on system bus 12. The current bus cycle address is used by apparatus 10 to derive the address used locally for the purpose of accessing VRAM 28 and DRAM 26. In addition, data transceivers 50 provide a bidirectional memory data path between apparatus 10 and system bus 12.

Host interface 14 also includes address comparator 56, which receives and compares the programmed I/O address against the address provided on system bus 12 to generate a confirmation signal indicative of a match therebetween, which is provided to host control logic 52.

Interface register 46 of host interface 14 determines how apparatus 10 may respond to system bus cycles. In the preferred embodiment of the present invention, interface register 46 is an 8-bit register. The logic value of one predetermined bit position in the register indicates whether apparatus 10 is enabled to respond to memory accesses. Another bit is indicative of whether 8- or 16-bit data transfers are allowed. A third bit disables direct palette accesses on write operations. A fourth bit position resets hardware space register 60, interface register 46 and graphics processor 18. A fifth bit position value determines whether apparatus 10 responds to the upper or lower half of the memory block with base address determined by base register 48 (function thereof described below). These bit values are made available to host control logic 52.

In the preferred embodiment of the present invention, host control logic 52 is implemented with programmable array logic (PAL) technology. Host control logic 52 provides control logic for host interface for graphics processor 18 and generates control signals required by graphics processor 18 and other functional blocks. Host control logic 52 generates signals in order to select the appropriate input to apparatus 10, such as one of the host processor, system bus I/O space and VGA palette. Host control logic 52 also generates control signal sequences for registers A 42 and map register B 44, interface register 46, base register 48 and status register 41 in response to system bus I/O cycles. In addition, handshake signals between apparatus 10 and system bus 12 are generated herein.

Upon power-up, apparatus 10 configuration and initialization begins. A configuration procedure, which may be running on the host CPU, examines system bus memory space to look for a block of free space of a predetermined size. Once found, base register 48 is initialized with the base address of the free block of memory. This free space is a software selectable memory segment in system bus memory to which apparatus 10 may be mapped. Map register A 42 and map register B 44 each points to a predetermined block of local memory space in DRAM 26 or VRAM 28 and supplies either address to graphics processor 18 for access, thereof. Each block of local memory determined by map register A 42 and map register B 44 are mapped into the upper or lower half of the memory block occupied by apparatus 10 in system bus memory space. Implemented in this manner, the host processor may access the memory space of graphics processor 18 via its local memory space. Note that the local memory spaces pointed to by map register A 42 and map register B 44 may be contiguous, noncontiguous, or overlapping.

During a system bus cycle, address comparator 56 compares a predetermined set of address bits of the system bus address to the contents of base register 48. Another comparison is further made of another predetermined bit position in system bus address and the fifth predetermined bit value in interface register 46 which selects upper or lower half of memory block. If both comparisons yield matching values, then apparatus 10 is selected. Once apparatus 10 is selected, it uses yet another predetermined system bus address bit to enable either map register A 42 or map register B 44. In this manner, a predetermined block of system bus memory space is decoded and selected.

Address decoder 58 performs the function of decoding an address provided by graphics processor 18 and selects either DRAM 26, VRAM 28, hardware space register 60 or palette 32.

Hardware space register 60 may be implemented as a write-only register which serve as a graphics processor 18 configuration and mode controller programmable by graphics processor 18. In the preferred embodiment of the present invention, hardware space register 60 has four active bits and the logic level thereof sets one of four modes or configurations: VGA pass through feature, high or low resolution, 4-bit per pixel, or 8-bit per pixel. On power-up, hardware space register 60 preferably has a default setting of VGA pass through on, with low resolution, 4-bit and 8-bit per pixel selections off. The setting of hardware register 60 is reflected in the setting of status register 41, where graphics processor operation mode and configuration are made available.

Clock circuit 24 may be implemented with two oscillating crystals supplying base frequency requirements of 25 MHz and 64 MHz. The 25 MHz frequency is required by the 640.times.480 resolution and 64 MHz is required by the 1024.times.768 resolution. Dividers implemented by counter circuits are also provided to further decrease the signals to desired clock and timing signals. Such timing circuitry is known in the art.

Multiplexer 66 is adapted to receive outputs from 4-bit shift registers 62, 8-bit shift registers 64, or VGA connector 36. Upon power-up, the default setting provides multiplexer 66 to receive video signals from VGA connector 36, Which is connected to a VGA board, and pass the signals to palette 32. Palette 32 receives the signals and in turn drives monitor 68 through video connector 34. The data path from the VGA board through VGA connector 36 and multiplexer 66 to video connector 34 is a video pass through feature which enables one monitor to receive video signals to receive input from both the VGA board and graphics processor 18. This feature eliminates the need to connect two separate monitors to a graphics system having both a graphics processor and a VGA board.

In order to perform video pass through correctly, palette 32 must reflect all writes to the VGA palette (not shown). Address comparator 56 monitors system bus addresses for VGA access cycles and notifies host control logic 52 thereof. A write cycle is generated by graphics processor 18 and the write operation currently occurring on system bus 12 is copied or shadowed through data transceivers 50 and supplied to palette 32. Implemented in this manner, all VGA palette modifications are mirrored in palette 32 to ensure that colors in palette 32 match those in VGA palette. Preferably, palette 32 is implemented with the 66 MHz BT 478 manufactured by Brooktree.

Although the present invention has been described in detail, it should be understood that various changes, substitutions and alterations can be made hereto without departing from the spirit and scope of the present invention as defined by the appended claims.