WO1996041325A1 - Computer system with display controller including arbitration of display memory access according to a multi-tiered priority scheme - Google Patents
Computer system with display controller including arbitration of display memory access according to a multi-tiered priority scheme Download PDFInfo
- Publication number
- WO1996041325A1 WO1996041325A1 PCT/US1996/008145 US9608145W WO9641325A1 WO 1996041325 A1 WO1996041325 A1 WO 1996041325A1 US 9608145 W US9608145 W US 9608145W WO 9641325 A1 WO9641325 A1 WO 9641325A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- dram
- display
- tier
- access
- data
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
- G09G5/00—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators
- G09G5/001—Arbitration of resources in a display system, e.g. control of access to frame buffer by video controller and/or main processor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/14—Digital output to display device ; Cooperation and interconnection of the display device with other functional units
- G06F3/1423—Digital output to display device ; Cooperation and interconnection of the display device with other functional units controlling a plurality of local displays, e.g. CRT and flat panel display
- G06F3/1431—Digital output to display device ; Cooperation and interconnection of the display device with other functional units controlling a plurality of local displays, e.g. CRT and flat panel display using a single graphics controller
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
- G09G5/00—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators
- G09G5/36—Control arrangements or circuits for visual indicators common to cathode-ray tube indicators and other visual indicators characterised by the display of a graphic pattern, e.g. using an all-points-addressable [APA] memory
- G09G5/363—Graphics controllers
Definitions
- the present invention relates generally to a computer system with one or more display devices, such as a cathode ray tube (CRT) or liquid crystal display (LCD) for example.
- the display devices provide a user of the computer system with a visible display of computer data, such as text or graphics.
- the present invention is in the field of a computer system having a graphics generator, and a video display controller (VDC) for such a computer system.
- VDC video display controller
- the VDC receives image information, such as text or graphics generated by a processor (CPU) or retrieved by the CPU from another facility (such as a CD-ROM) of the computer system, and provides signals driving one or both of the CRT or LCD displays.
- the present invention is in the field of a VDC having a sequencer and controller (SEQC) for a dynamic random access memory (DRAM) of the VDC.
- Image information to be displayed on the CRT or LCD is stored in the DRAM in preparation to being transferred to a video memory of the first-in-first-out (FIFO) type.
- FIFO first-in-first-out
- other devices of the computer system such as a bit block transfer engine (bit-BLT) (i.e., a graphics generator) also request access to the DRAM.
- bit-BLT bit block transfer engine
- the SEQC arbitrates requests for access to the DRAM by the various devices of the computer system and prioritizes these requests for access to insure both that the display FIFO is not denied data for display, and that most efficient access to the DRAM is provided to the other devices of the computer system.
- a conventional bus arbitrating circuit is known in accord with United States patent No. 4,453,214 (hereinafter, the '214 patent), issued 5 June 1984 to Ralph L. Adcock.
- a bus arbitrator and memory manager (BAMM) establishes a priority among competing operating units of a computer system.
- the BAMM sorts requests for access to the memory according to a priority, and allows the device with the highest priority access ahead of the other devices. It appears that once a device is allowed access to the memory, an interrupt of this access is not allowed when a request for access from another device with a higher priority is received by the BAMM of the '214 patent. When a device which has had memory access is finished with this access, it provides a "sign off' signal, thus allowing the BAMM to permit memory access to the device requesting access and having the highest priority.
- a display FIFO of a computer system could conceivably be denied access to the DRAM at a time when a display FIFO is nearly or completely out of information for display.
- continuity of operation of the display of the computer system could be interrupted. Understandably, this type of display interrupt would be concerning and confusing for a user of the computer system.
- a DRAM controller receives refresh requests and requests for access to the DRAM in bursts, and arbitrates among the requests.
- the '112 patent receives refresh requests and requests for access to the DRAM in bursts, and arbitrates among the requests.
- the display stage after a preparatory read, the greatest priority is given to the display FIFO.
- a read of the DRAM in bursts may be interrupted when the FIFO is full. In this case, priority is given to a possible preparatory read.
- a graphics controller such as a VDC generally described above
- arbitrating DRAM interface (access) among the several devices of the system is the most critical portion of the controller.
- Access to the DRAM dictates how and when devices such as the bit-BLT engine, display FIFO, and the local bus (that is, the CPU) have access to the DRAM.
- Access requests by the CPU and bit-BLT engine are mutually exclusive, and will not occur simultaneously.
- a new page of the DRAM must be accessed. That is, the DRAM may be visualized as a two-dimensional array of memory locations.
- This memory uses rows and columns of memory locations (or memory cells) with a row pointer and a column pointer. As long as memory access is made to a single row of the memory, with the column pointer simply moving along the row as data is written to or read from address locations of the row, then a single- page access to the memory is effected, and no page break is necessary. However, when another row (i.e., another page) of the memory must be accessed, a pre-charge sequence must be run in preparation to accessing the next row of memory locations. This pre-charge sequence takes time so that a multiple-page access to the memory is not nearly as efficient as a single-page access in terms of the amount of data written into or read from the memory during the time interval of such a memory access.
- page-mode access to the DRAM is much more efficient in terms of time utilization than is random access to the DRAM because of the many page breaks required for random access.
- page-mode is not maintained for the DRAM, then at least one preparatory pre-charge cycle must be conducted to allow access to another different page of the DRAM in addition to the time interval required to write the data to or read the data from the memory cells.
- access is allowed to the DRAM for the bit-BLT, these accesses will ordinarily be multi-page accesses which consume considerable time, but a request for this access does not require that immediate access to the DRAM be granted.
- CPU local bus access to the DRAM is usually a single-page access, requires considerably less time than a bit-BLT access, and also does not require that a request result in immediate access.
- the system throughput is decreased and the WINMARKS (industry standard performance bench marks) for the computer system also are decreased.
- the display FIFO of a graphics controller also requests DRAM access, and may be envisioned as a storage tank of water (data) draining at a uniform rate from the bottom, and only occasionally being refilled from the top.
- the display FIFO stores image information to be sent to the display devices (i.e., to the CRT or LCD, for example).
- the rate of drainage of the data from the display FIFO depends on the mode of display operation. If the display is being operated in a grey-scale mode which requires four bits per pixel, then the display FIFO will not drain very fast. On the other hand, if the user is operating the display in a color mode, then each pixel of the display may require eight bits, or sixteen bits, or possibly more than sixteen bits of information; and the display FIFO will drain correspondingly faster.
- the refilling rate of the display FIFO is much higher than the draining rate. But, refilling may be intermittent and interrupted for the allowance of other activities requiring access to the DRAM. Further, it must be understood that while the FIFO is being refilled, complete double-words of data must be input from the DRAM. If there is insufficient room at the top of the display FIFO to accept all of the last complete double-word of data being input at a particular time, then some of the existing data will be overwritten and lost.
- a FIFOLO request (a low priority request for DRAM access) is issued by the display FIFO to the DRAM controller as soon as the display FiFo has room at the top for at least one double-word of new data without overwritin existing data waiting to be sent to the display device. Consequently, one or more accesses to the DRAM may be granted to the display FTJTO in response to the FIFOLO request. This request is not cleared until the FIFO is filled. If the display FIFO is not adequately refilled in response to the FIFOLO request, then as soon as the display FIFO starts to write its last double- word of data to the display a FIFOHI request for access to the DRAM will be issued. This FIFOHI request will be honored immediately.
- FIFOHI request will not be cleared until the FIFO is filled completely. Consequently, a conventional DRAM controller will clear both FIFOLO and FIFOHI simultaneously after a FIFOHI request has been issued. Again, these requests for DRAM access would conventionally not be cleared until the FIFO is completely filled with fresh data.
- Figure 1 depicts a timing diagram showing an idealized sequence of accesses to a DRAM of a VDC alternating between a display FIFO and a bit-BLT engine.
- Line 2 of this Figure 1 also shows an idealized sequence of accesses to the DRAM by a display FIFO and the CPU.
- These idealized timing diagrams show that neither the bit-BLT or CPU is required to wait for DRAM access, that the DRAM has no idle time, and that the accesses granted are relatively long for the bit-BLT so that multi-page accesses can be accomplished.
- Conventional computer system graphics controllers do not achieve such idealized management of DRAM access.
- FIG. 2 depicts a timing diagram as might be experienced in an actual conventional computer system graphics controller. Viewing Figure 2, the first of the three time lines of this graph respectively shows requests for access to the DRAM from the CPU.
- the next two lines show access requests from the display FIFO: first on a low priority basis (FIFOLO) - indicating that the display FIFO is sufficiently depleted of display information that at least one double-word of new information can be written to this FIFO without overwriting existing data; and secondly, on a high priority basis (FIFOHI) - indicating that the display FIFO is using its last double-word of information and is at risk of running out of information to be provided to the user by means of the display device (i.e., the CRT or LCD, for example).
- FIFOLO and FIFOHI requests are not cleared (i.e., removed or discontinued) until the display FIFO is granted DRAM access and the FIFO is completely filled with data.
- Interval #4 indicates waiting time for the CPU.
- the end of interval #4 indicates the simultaneous clearing of both FIFOLO and FIFOHI, and the beginning of interval #5 during which the CPU is finally granted DRAM access.
- Interval #9 indicates the issuance of a FIFOLO request from the display FIFO for DRAM access. Because insufficient data is provided to the display FIFO in response to the FIFOLO request (another device, such as the bit-BLT, for example, may be making DRAM access so that the FIFOLO request in not sufficiently honored), the display FIFO issues a FIFOHI request at the beginning of interval #10. This FIFOHI request is honored immediately during interval #10. However, another interval (interval #10) results during which the CPU is denied access to the DRAM. At the end of interval #10 the FIFOLO and FIFOHI requests are both cleared simultaneously, and the CPU is granted DRAM access.
- Figure 2 shows that the conventional arbitration scheme results in the DRAM sometimes being idle (intervals 2, 6, and 8), and in either the CPU or display FIFO waiting for access to the DRAM (intervals 4, 9, and 10).
- the CPU is be required to wait until the display FIFO is completely filled before access to the DRAM can be granted to the CPU, even though the display FIFO may have received enough data that there in no longer an immediate risk of its running out of data for the displays.
- This conventional graphics controller both fails to maintain page mode for the DRAM, and also decreases the throughput rate for the computer system. Accordingly, a long-felt need has been recognized for a more efficient and effective way of arbitrating access to the DRAM of a graphics controller.
- an object for thi invention is to provide a computer system with a graphics controller avoiding one o more of these deficiencies.
- Another object for this invention is to provide such a computer system with a video display controller (VDC) receiving image information, such as text or graphics, and providing signals driving a display.
- VDC video display controller
- the present invention has as an object the provision of a VDC having a sequencer and controller (SEQC) arbitrating access to a dynamic random access memory (DRAM) of the VDC among a CPU, a video memory of the first-in-first-out (FIFO) type, a bit-BLT, and other devices of the computer system.
- SEQC sequencer and controller
- the present invention provides a computer system with a graphics controller (i.e., a video display controller) having a sequencer-controller (SEQC) implementing a two-tiered sequence of priorities among requests for access to a DRAM of the VDC, with each request in the upper tier being allowed to effect an interrupts of any lower-priority access to the DRAM, and requests in the lower tier not being allowed to effect such an interrupt. Interrupted accesses and requests awaiting access to the DRAM are placed in a que according to the priority they enjoy in the two-tiered priority scheme.
- a graphics controller i.e., a video display controller
- SEQC sequencer-controller
- a result is that the sequence and logic of the priorities assigned to requests for access to the DRAM by the various devices of the computer system, along with the enabling or denial of interrupting access to the DRAM by other devices of the computer system, causes the DRAM to be accessed more effectively.
- those access to the DRAM which need to be relatively long multi-page accesses will be implemented in this way; while under other operating conditions, the accesses will be briefer accesses always insuring that priority requests to the DRAM are implemented in a timely manner. That is, requests for accesses by a display FIFO on a high priority basis and for drawing a mouse image, for example, will always be allowed immediate access to the DRAM.
- the display access requests and the mouse image requests never occur simultaneously.
- the present invention provides a computer system including an input device for receiving inputs from a user of the computer system; a central processing unit (CPU) interfacing with the input device and responding to the inputs by performing a processing function producing an output including display data; a memory device interfacing with the CPU to provide memory storage of commands and data used by the CPU; an output device receiving the output response from the CPU and providing a sensible output response; the output device including a display device providing a visible image to the user in response to the display data; a video display controller (VDC) including a graphics generator (GG) interfacing with the processor and with the output device to provide additional display data providing portions of the visible image; a dynamic random access memory (DRAM) interfacing with the CPU via the VDC and receiving the display data including the additional display data for temporary storage, the DRAM also interfacing with the CPU and the GG to provide temporary storage of data used in the processing function and in the providing of the additional data; the VDC including a display first-in-first-out (FIFO) memory interfacing with the
- Figure 1 provides a graphical representation of idealized accesses to a dynamic random access memory by a bit block transfer (bit-BLT) engine and by a central processing unit (CPU).
- bit-BLT bit block transfer
- CPU central processing unit
- Figure 2 presents a hypothetical timing diagram which may be experienced in a prior art computer system having a conventional graphics controller.
- Figure 3 provides a pictorial presentation of a computer system, including a notebook type computer having a LCD display to display a first image, and also a television which being used as a display device for the computer system 10 to display a different second image;
- Figure 4 is a schematic functional block diagram of the computer system seen in Figure 1;
- FIG. 5 provides a schematic functional block diagram of the video display controller (VDC) of the computer system seen in the preceding figures
- Figure 6 is a graphical representation of a first-in-first-out display memory of the computer system seen in preceding Figures;
- Figure 7 is a schematic functional block diagram of a sequencer and controller (SEQC) of the present computer system
- Figure 8 provides a tabulation of a two-tiered prioritized arbitration scheme implemented for allowing access to a DRAM of the present computer system
- Figure 9 provides two simultaneously running flow charts implemented by the SEQC of the present invention in arbitrating access to the DRAM;
- Figure 10 provides a tabulation of a three-tiered prioritized arbitration scheme implemented for allowing access to a DRAM of an alternative embodiment of the present computer system.
- Figures 11 and 12 provide timing diagrams illustrating the result of the arbitration for access to the DRAM performed by the SEQC of the present computer system.
- a computer system 10 includes a notebook computer 12, and an additional display device 14 interfaced with the notebook computer via a cable 16.
- the additional display device 14 is illustrated as a conventional television. Those ordinarily skilled in the pertinent arts will recognize that the television accepts signals in NTSC format and displays an interlaced image. Alternatively, the computer system 10 may be interfaced with a conventional CRT monitor using RGB signals and providing a non-interlaced image.
- the notebook computer includes various input devices, such as a keyboard 18, a floppy disk drive 20, and a track ball 22. Those ordinarily skilled in the pertinent arts will recognize that the track ball is essentially a stationary mouse input device.
- the computer system 10 may include additional input devices, such as a hard disk drive, a CD- ROM, and a serial input-output I/O) port. Several of these devices also function as output devices for the computer system 10 in addition to a liquid crystal display 24. As described hereinbelow, the display 24 is presented as being of dual panel type. As depicted, the notebook computer is being used to perform a multi-task operation. For example, the notebook computer 12 may be used to conduct a financial analysis, the data for which is displayed on LCD 24, and a graphical depiction of which is displayed on CRT 14.
- FIG. 4 provides a schematic block diagram of the computer system 10, with the input devices all subsumed within one representative block 26.
- the input devices are interfaced with a microprocessor 28, which also has an interface with a memory facility 30.
- the memory facility 30 will include the floppy disk drive 20, and may include a hard disk drive, CD-ROM, and other devices.
- a data bus 32 interfaces with the microprocessor 28 and provides an interface with the output devices, including the LCD and CRT image display devices 14 and 24.
- the other output devices for the computer system 10 are subsumed in a representative block 34.
- the computer system 10 includes a video display controller (VDC) 36 interfacing with the bus 32, and providing driving signals for the LCD 24 and CRT 14.
- the VDC has an interface with dynamic random access memory (DRAM), represented on Figure 4 with the schematic blocks 38.
- the VDC has an interface with a power management facility 40 of the computer system 10.
- a dedicated clock 42 provides a reference clock rate to the VDC 36.
- the VDC 36 includes an internal clock 44 referenced to the clock signal from the dedicated clock 42, and providing clock signals to a video section 46 of the VDC.
- the clock signals provided by internal clock 44 may include a pixel clock (Pclk) and a memory clock (Mclk), the use of which will be further explained below.
- the video section 46 includes a programmable host interface 48.
- the host interface 48 is programmable to configure the VDC 36 for interface with a number of conventional bus configurations.
- host interface 48 may be configured for interface with a conventional Intel 486DX local bus, with a VL-Bus, and with a PCI interface bus.
- the host interface 48 interfaces the bus 32 with a VGA core portion 50 of the VDC 36.
- This VGA core portion 50 includes a sequencer, to be further described below, a cathode ray tube controller (CRTC), a graphics controller, an attribute controller, and conventional VGA circuitry.
- CRTC cathode ray tube controller
- the VGA core 50 In order to allow the VGA core 50 to generate and control the text, graphics and other visual characters to be displayed on the CRT and LCD (such as a cursor and icons, for example), the VGA core is interfaced with a hardware cursor generator 52, a bit-BLT engine 54, and a display FIFO 56. An additional two display FIFO's 56', and 56" are also interfaced with the VGA core 50.
- An alternative embodiment of the VDC 36 supporting only a single display devic (either LCD or CRT) will include only a single display FIFO, and is further explained below.
- Another ⁇ Jternative embodiment supporting two display devices one LCD and one CRT will include two display FIFO's 56 and 56'.
- this embodiment will also support a single display device of LCD or CRT type.
- the embodiment including display FIFO 56 and the additional two display FIFO's 56' and 56" is employed to support the dual display operation of the computer system 10, as was explained with reference to Figure 3, using a standard television as the second display device.
- VDC 36 When a display device providing an interlaced image is used to display non- interlaced imagery, the image ordinarily includes a lot of flicker.
- the computer system 10 includes the two additional display FIFO's 56' and 56" which are employed to store alternate lines of the non-interlaced imagery, and to sequentially supply these alternate lines of imagery to the television 24 for display as an interlaced image without flicker.
- this reference includes also display FIFO's 56' and 56".
- the alternative embodiment of the invention having only a single display FIFO may implement a simplified decisional scheme when deciding on allowing access to the DRAM 38.
- the hardware cursor generator 52 selectively provides a cursor of increased size (i.e., twice as large as normal, for example), which is easier to visually follow as it moves across a display screen, in response to detection of a certain preselected speed of movement of the cursor provided by a software program running on microprocessor 28.
- a cursor of increased size i.e., twice as large as normal, for example
- a software program running on microprocessor 28.
- the bit-BLT engine performs read, write, and block transfers of bits representing these characters, solid fills, destination inversions, and pattern fills.
- the bit-BLT performs all data alignment and masking at the boundaries of block transferred characters, as well as text expansions to accelerate the writing of monochrome images.
- the display FIFO temporarily stores bits of information, in integer multiples of double-word size units or levels, awaiting the writing of these bits to pixels of the displays 14 and 24.
- the display FIFO 56 is an eight-stage FIFO, storing eight 32-bit double- words of display information for sending to the CRT and LCD 14 and 24.
- Each of the hardware cursor generator 52, bit-BLT 54, and display FIFO 56 are also interfaced with a DRAM controller 58.
- This DRAM controller 58 implements the functions of the DRAM controller/sequencer described in general terms above to arbitrate and implement requests for access to the DRAM by various functional units of the computer system 10, including other portions of the VDC 36.
- the DRAM controller 50 has an interface with the DRAM 38.
- the DRAM 38 is shown in Figure 5 as a single functional block. However, those ordinarily skilled in the pertinent arts will recognize that this DRAM may comprise one or several DRAM chips.
- the display FIFO 56 has an interface (via the VGA controller 50 and DRAM controller 58) with both a palette controller 60, and with a liquid crystal display (LCD) interface controller 62.
- the palette controller implements the standard 256-by-18 VGA palette, while the LCD interface controller performs frame modulation and dithering for 64 shades of grey in monochrome mode operation; and 4K colors, with dithering for a full 256K colors in color mode operation.
- the VDC 36 includes a power down controller 64.
- This power down controller has an interconnection with a power down register 65, which itself has a generalized interconnection within the VDC 36.
- This generalized interconnection of the power down register 65 is indicated on Figure 5 with the plurality of arrows leaving the register 65.
- These interconnections of the power down register 65 permeate the VDC 36 and allow it to be configured for various modes of operation and for various corresponding power down modes.
- the power down controller 64 has an interface with the LCD 24 in order to facilitate such power saving functions as LCD back light "off', and LCD display "off', under control of parameters set by the user of the computer system 10.
- this display FIFO 56 has a capacity of 8, 32-bit double-words, or levels. These levels are indicated by the numerals 1-8 along the left side of Figure 6.
- Other memory capacities may be employed for a display FIFO without departing form the spirit and scope of the present invention. From the bottom of this display FIFO, while the display 14 or 24 is active, data is continuously drained to the display units at a rate which is mode dependent, as was explained above. At the top, the display FIFO is intermittently refilled at a rate dependen upon the speed of the DRAM 38 and of the memory clock of the VDC (as well a other parameters of the computer system 10).
- sw are placed two movable pointers.
- One of these pointers indicates that when the data level in the FIFO falls below this pointer, then a FIFOLO request is issued for additional data from the DRAM.
- This pointer 66 has a permissible position from 4 to 7.
- the other pointer 68 indicates the issuance of a FIFOHI request for additional data from the DRAM 38.
- Pointer 68 has a permissible position from 0 to 7. In each case the issuance of a FIFOLO or FIFOHI request indicates that the FIFO 56 can accept at least one additional double-word level of data.
- the position of the pointer 68 along the FIFO 56 is dependent upon the mode of operation of the displays 14 and 24 (indicative of the rate of drainage of the FIFO 56), and the rate of possible filling of this FIFO (determined by the interval of the pixel and memory clocks of the VDC, the speed of the DRAM 38, and other interconnect intervals and data transfer intervals of the VDC) such that a FIFOHI request can be issued at a time as early as level 7 or as late as zero level of data in the FIFO 56 in order to insure that the FIFO does not run out of data.
- FIFOHI when FIFOHI is issued, other accesses to the DRAM 38 are interrupted. Accordingly, if the display FIFO is draining slowly and the computer system can refill the FIFO quickly, the pointer 68 can be set at zero and the display will still not run out of data.
- the pointer 66 is dependent for its position on the mode of display operation and on similar parameters of the computer system 10. This pointer will be set in the range from 4 to 7 in order both to facilitate early filling of the FIFO 56 with a minimal number of FIFOHI requests being issued, and to allow other devices of the computer system with best access to the DRAM 38. Understandably, the set point for the FIFOLO request (pointer 68) is not as critical as that for the FIFOHI request, and as will be seen, the level for this FIFOLO request pointer fits into a lower- tier prioritization scheme implemented by the DRAM controller 58. However, the FIFOLO request is issued at a level of the FIFO attempting to obtain sufficient access to the DRAM that a FIFOHI request will not be issued, or that the intervals between FIFOHI requests will be maximized.
- FIG 7 a functional block diagram of the interconnections of the DRAM controller 58 and DRAM with the various devices of the computer system 10 is depicted.
- the numeral 70 within a block indicates a possible request for a DRAM refresh cycle, which request is issued on a regular repeating time interval by a clock in the VDC 38 (indicated on Figure 5 as "DRclk"). Thus, the receipt of this request is a certainty. The time-sequencing of this request with the other requests is uncertain.
- the numeral 72 within a block indicates a possible request for access to the DRAM by a half-frame buffer of the LCD controller 62 (indicated on Figure 5 with the numeral 62').
- This half-frame buffer receives and temporarily stores in the DRAM 38 pixel values which are written to the panels of the LCD 24.
- one of the panels receives fresh information from its associated display FIFO via the LCD controller 62.
- the other panel receives a repeat of previous pixel values which had been previously stored temporarily in the DRAM by the half-frame buffer 62'.
- the panels of the LCD 24 alternate in receiving fresh image data from the display FIFO and from the half-frame buffer 62', with the half-frame buffer temporarily storing the fresh pixel values in the DRAM 38 for use in refreshing the particular panel of the display 24 while the other of the two panels is receiving fresh image data.
- This half-frame buffer has a limited amount of internal memory. Accordingly, during a memory access to the DRAM 38, the half-frame buffer 62' will receive enough pixel values to provide refreshing of several pixels on the display 24.
- the half-frame buffer issues requests for access to the DRAM on a FRAMELO (low priority), or FRAMEHI (high priority) basis dependent upon the amount of data remaining in the limited memory capacity of the half-frame buffer for use in refreshing pixels of the LCD 24.
- the numeral 74 within block 28 indicates a possible request made by the
- the numeral 76 within block 54 indicates a possible request for access to the DRAM 38 issued by the bit-BLT engine 54, while the numeral 78 within block 56 indicates possible requests (FIFOLO or FIFOHI) issued by the display FIFO 56 seeking fresh data to be temporarily held for sending the display devices 14 or 24.
- the FIFO requests may include a FIFOLO or FIFOHI request from and identified with each of the FIFO's 56, 56', and 56". Requests once make are continued (remain as pending) until satisfied, or until otherwise cleared in the case of the FIFOLO and FIFOHI requests.
- the numeral 80 within a block indicates a possible request received from the mouse image generator circuits (i.e., introduced and explained above) for access to the DRAM 38 to draw a mouse image.
- the numeral 82 within a block represents an address generator servicing the display FIFO 56 (or FIFO's 56, 56', and 56") by generating addresses for use in reading data from the DRAM 38 to the display FIFO in response to a request for such data.
- FIFO 56 or FIFO's 56, 56', and 56
- a DRAM address multiplexer 84 provides the generated addresses to the DRAM 38.
- This address multiplexer also includes a facility for recognizing when generated addresses require a page break in the DRAM 38, and provides a page break signal (indicated with numbered arrow 88) to the SEQC 86 which is used in initiating the pre-charge sequence necessary in the DRAM 38 in order to allow a different page of this memory to be accessed.
- This page break signal is indicated when new data has a different row address from the last previous data input into the DRAM 38. In the event that the new data has the same row address, a page break signal is not issued, and page mode operation of the DRAM is maintained with no lost time for a pre-charge sequence even when the last previous data and the new data are from different devices of the computer 10. That is, a change of device accessing the DRAM 38 does not necessarily cause a page break in the DRAM 38.
- a priority logic unit 90 implementing a logical selection process among the pending requests for access to the DRAM 38, as is illustrated in Figures 8 and 10.
- Figure 8 represents the simpler alternative of a DRAM controller having only a single display FIFO 56, and will be considered first.
- the pending requests for access to the DRAM 38 are first of all assigned to one of two tiers (an upper tier and a lower tier), as will be further explained.
- pending requests are ranked in order of priority (numbered lu through 5u).
- pending requests are ranked in order of priority (indicated as 11 through 31).
- each upper- tier request may interrupt any existing access to the DRAM 38 granted in response to another upper-tier request with a lower rank, and may also interrupt an access granted in response to all lower-tier requests.
- a FIFOLO request ranked 11
- a request to refresh the DRAM is received by the SEQC (ranked 4u)
- the display FIFO access is interrupted.
- the DRAM is then refreshed.
- Figure 9 graphically depicts an additional function performed by the DRAM controller 58 by use of a level counter 92 (seen in Figure 8), recalling that display FIFO 56 and DRAM controller 58 are both within the VDC 36 and are interfaced with one another.
- the level counter 92 continuously monitors addresses generated for accessing data within DRAM 38, and which are used in writing this data to the display FIFO 56. Every Nth level of data ("N" representing a selected integer multiple of levels of data provided to the display FIFO 56), the level counter 92 resets a flag or register in the SEQC 86.
- N representing a selected integer multiple of levels of data provided to the display FIFO 56
- the level counter 92 resets a flag or register in the SEQC 86.
- the DRAM controller 58 simultaneously and independently tests the result of two separate questions.
- One question is whether the level of data in display FIFO 56 is below the FIFOLO pointer. IF the answer is "no”, the question is continued. If the answer is "yes”, then the FIFOLO request is issued. Every Nth level of data writte to the display FIFO 56 (as indicated by the reset flag or register explained above) the question is ask again, and if the answer is "no", the FIFOLO request is cleared Thus, FIFOLO may be cleared without the display FIFO being completely fille with data.
- FIFOHI may also be cleared without the display FIFO being completely fille or even filled to the level of the FIFOLO pointer.
- the display FIFO 56 need only be filled to a level above the FIFOHI pointer at the completion of writing N levels of data into the display FIFO.
- the value of "N" is selected to be four (4).
- This value for N is convenient with a display FIFO having eight levels as described, and with these levels each being 32 bits.
- each pixel takes 4 bits, and the frame buffer refreshes the LCD display every 32 pixels, so there is a beneficial correlation in these modes of operation between the sequencing of frame refreshes, and the writing of N levels of data to the display FIFO.
- the display FIFO need not be eight levels deep, and N need not be selected to be four.
- N will be selected in view of the interplay between the size of the display FIFO and the speed at which data can be accessed (DRAM speed) and written to this display FIFO, as well as the rate at which the display FIFO is depleted of data, and the requirements for other devices of a particular system to access the DRAM.
- DRAM speed the speed at which data can be accessed
- the rate at which the display FIFO is depleted of data the requirements for other devices of a particular system to access the DRAM.
- the priority logic scheme implemented by the priority logic unit 90 in the more complex alternative of a DRAM controller having two or more display FIFO's 56, 56', 56", etc. is presented graphically. Viewing Figure 10, it is seen that the pending requests for access to the DRAM 38 are first of all assigned to one of three tiers (an upper tier, a middle tier, and a lower tier), as will be further explained.
- pending requests are ranked in order of priority (numbered lu through 3u). It will be noted that this tier has several co-equal requests at one ranking level, as will be explained further. Within the middle tier, requests are ranked from lm to 2m. Similarly, within the lower tier, pending requests are ranked in order of priority from 11 through 31. Within this logical structure of ranked pending requests for access to the
- each upper-tier request is placed in a que of requests in that tier.
- These requests for access to the DRAM 38 are honored in order of their position in the que.
- the upper tier requests may not interrupt any existing access to the DRAM 38 granted in response to another upper-tier request.
- these upper-tier requests lu-3u include a rank (2u) containing several co-equal requests. That is, the upper-tier rank 2u includes plural FIFOHL requests, in which the subscript "n" indicates the one of several FIFO's 56 ⁇ -n making the request.
- the FIFO's 56, 56', and 56" each carry a numerical identifier 1-3, with this identifier being used to indicate the source of a FIFO access request in Figure 10.
- an upper tier request may interrupt any existing access to the DRAM granted in response to any middle-tier or lower-tier request.
- a middle-tier request may not interrupt any upper-tier access, and may only interrupt any DRAM access granted in response to a middle- tier request with a lower rank, or any access granted in response to a lower-tier request.
- the lower-tier requests are not able to interrupt any other access to the DRAM, and are placed in a respective que for service. It will be noted that these lower-tier requests also include a rank (11) containing several co-equal requests. That is, the lower-tier rank 11 includes plural FIFOLOn requests, in which the subscript "n" similarly indicates the one of several FIFO's 56 ⁇ - n making the request.
- Figures 11 and 12 present timing diagrams illustrating (by way of example only) a result of the two-tiered arbitration for access to the DRAM performed by the SEQC of the present computer system. A similar result can be expected with a three tiered arbitration scheme as presented in Figure 10.
- FIG. 11 it is seen that in interval #1, the bit-BLT has made a request for access to the DRAM, and that this request is granted.
- the display FIFO has made a request for access to the DRAM.
- the fact that both FIFOLO and FIFOHI are issued simultaneously indicates that the display FIFO is empty of display data.
- the display FIFO In interval #2, the display FIFO supersedes the bit-BLT request (recalling the priority scheme of Figure 8), and receives sufficient display data to result in the FIFOHI request being canceled (recalling the test conducted by the flow chart of Figure 9). In interval #3, the display FIFO is still serviced by the DRAM because a FIFOLO request (still pending) has a higher priority than the pending bit-BLT request.
- interval #4 the FIFOLO request is canceled (not an indication of a full FIFO, but of a FIFO level above the FIFOLO pointer), and the bit-BLT request is honored.
- the FIFOLO request is issued, but the bit- BLT engine retains access to the DRAM because lower-tier requests cannot interrupt one another (recalling Figure 8).
- FIFOHI is issued, and interrupts the access of the bit-BLT.
- Interval #7 indicates that FIFOHI has been canceled, but that the display FIFO retains access to the DRAM, again because the pending bit-BLT request is also a lower-tier request and cannot interrupt the still- pending FIFOLO request.
- Interval #8 indicates that when FIFOLO is canceled, then the pending bit-BLT request is honored.
- Figure 12 presents a similar arbitration episode, this time with the SEQC arbitrating between the CPU and the display FIFO.
- interval #1 indicates a CPU request for access to the DRAM. This request is granted and the CPU signs off.
- interval #2 the DRAM is idle, but interval #3 indicates that the display FIFO has made a request for access to the DRAM. Again, the fact that both FIFOLO and FIFOHI are issued simultaneously indicates that the display FIFO is empty of display data.
- the display FIFO supersedes the CPU request (recalling the priority scheme of Figure 8), and receives sufficient display data to result in the FIFOHI request being canceled (again recalling the test conducted by the flow chart of Figure 9).
- interval #4 the CPU is allowed access to the DRAM because an upper-tier request can interrupt an access granted for any lower-tier request (FIFOLO being a lower-tier request).
- interval #5 the CPU has signed off the DRAM, and the FIFOLO request is honored.
- Interval #6 indicates that the CPU has made a request for access to the DRAM, which interrupts the display FIFO access because only FIFOLO is pending.
- the CPU signs off the DRAM (interval #7) the DRAM access is returned to the display FIFO.
- the FIFOHI request is issued and is honored, however this interval #8 includes the issuance of a CPU request for access to the DRAM which is not honored because the CPU cannot interrupt a higher-ranked upper-tier request.
- Interval #9 indicates the clearing of FIFOHI, and the granting of access to the CPU, because the CPU can interrupt the display FIFO request based on the pending FIFOLO request.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Computer Graphics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- General Engineering & Computer Science (AREA)
- Controls And Circuits For Display Device (AREA)
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP50089297A JP3557619B2 (en) | 1995-06-07 | 1996-05-31 | Computer system with display |
DE69604667T DE69604667T2 (en) | 1995-06-07 | 1996-05-31 | COMPUTER SYSTEM WITH DISPLAY CONTROL INCLUDING IMAGE STORAGE ACCESS ARBITRATING ACCORDING TO A MULTI-ROW PRIORITY SCHEME |
EP96916845A EP0834168B1 (en) | 1995-06-07 | 1996-05-31 | Computer system with display controller including arbitration of display memory access according to a multi-tiered priority scheme |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US08/487,117 | 1995-06-07 | ||
US08/487,117 US5767866A (en) | 1995-06-07 | 1995-06-07 | Computer system with efficient DRAM access |
Publications (1)
Publication Number | Publication Date |
---|---|
WO1996041325A1 true WO1996041325A1 (en) | 1996-12-19 |
Family
ID=23934492
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US1996/008145 WO1996041325A1 (en) | 1995-06-07 | 1996-05-31 | Computer system with display controller including arbitration of display memory access according to a multi-tiered priority scheme |
Country Status (6)
Country | Link |
---|---|
US (1) | US5767866A (en) |
EP (1) | EP0834168B1 (en) |
JP (1) | JP3557619B2 (en) |
KR (1) | KR19990022755A (en) |
DE (1) | DE69604667T2 (en) |
WO (1) | WO1996041325A1 (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0802519A1 (en) * | 1996-04-19 | 1997-10-22 | Seiko Epson Corporation | System and method for implementing an overlay pathway |
WO1999028893A1 (en) * | 1997-12-01 | 1999-06-10 | Mediaq, Inc. | High performance, highly integrated computer architecture with extendible embedded memory |
EP0945843A1 (en) * | 1998-03-26 | 1999-09-29 | Alcatel | Method of controlling a liquid crystal display |
JPH11345165A (en) * | 1997-12-05 | 1999-12-14 | Texas Instr Inc <Ti> | Traffic controller using priority and burst control for reducing access times |
US6944696B2 (en) * | 1999-12-27 | 2005-09-13 | Renesas Technology Corp. | Data processing processor |
Families Citing this family (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5983315A (en) * | 1997-04-25 | 1999-11-09 | Rockwell Science Center, Inc. | System and method for establishing priorities in transferring data in burst counts from a memory to a plurality of FIFO stages, each having a low, intermediate, and high region |
US5953020A (en) * | 1997-06-30 | 1999-09-14 | Ati Technologies, Inc. | Display FIFO memory management system |
US5903283A (en) * | 1997-08-27 | 1999-05-11 | Chips & Technologies, Inc. | Video memory controller with dynamic bus arbitration |
US6091431A (en) * | 1997-12-18 | 2000-07-18 | Intel Corporation | Method and apparatus for improving processor to graphics device local memory performance |
US6374244B1 (en) * | 1998-04-01 | 2002-04-16 | Matsushita Electric Industrial Co., Ltd. | Data transfer device |
US6914605B2 (en) * | 2000-03-21 | 2005-07-05 | Matsushita Electric Industrial Co., Ltd. | Graphic processor and graphic processing system |
US6670959B2 (en) * | 2001-05-18 | 2003-12-30 | Sun Microsystems, Inc. | Method and apparatus for reducing inefficiencies in shared memory devices |
US6654021B2 (en) * | 2001-05-18 | 2003-11-25 | Sun Microsystems, Inc. | Multi-channel, demand-driven display controller |
US7081896B1 (en) * | 2002-08-27 | 2006-07-25 | Nvidia Corporation | Memory request timing randomizer |
US7506133B2 (en) * | 2003-08-20 | 2009-03-17 | Seiko Epson Corporation | Method and apparatus for high speed addressing of a memory space from a relatively small address space |
JP4624715B2 (en) * | 2004-05-13 | 2011-02-02 | ルネサスエレクトロニクス株式会社 | System LSI |
JP4060294B2 (en) * | 2004-07-09 | 2008-03-12 | 株式会社ソニー・コンピュータエンタテインメント | Image processing apparatus, image processing system, image processing method, computer program, semiconductor device |
WO2010016169A1 (en) * | 2008-08-07 | 2010-02-11 | 日本電気株式会社 | Multiprocessor system and method for controlling the same |
US8456478B2 (en) * | 2008-10-30 | 2013-06-04 | Microchip Technology Incorporated | Microcontroller with integrated graphical processing unit |
WO2015067295A1 (en) | 2013-11-05 | 2015-05-14 | Huawei Technologies Co., Ltd. | Method and arrangement for controlling requests to a shared electronic resource |
US10937484B2 (en) * | 2015-12-30 | 2021-03-02 | International Business Machines Corporation | Dynamic bandwidth throttling of DRAM accesses for memory tracing |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4009470A (en) * | 1975-02-18 | 1977-02-22 | Sperry Rand Corporation | Pre-emptive, rotational priority system |
EP0255436A1 (en) * | 1986-07-28 | 1988-02-03 | Bull S.A. | Graphics board with keyboard and mouse interface |
US4845661A (en) * | 1985-08-19 | 1989-07-04 | Nec Corporation | Display information processing apparatus |
US4942553A (en) * | 1988-05-12 | 1990-07-17 | Zilog, Inc. | System for providing notification of impending FIFO overruns and underruns |
US4991112A (en) * | 1987-12-23 | 1991-02-05 | U.S. Philips Corporation | Graphics system with graphics controller and DRAM controller |
EP0522697A1 (en) * | 1991-06-10 | 1993-01-13 | International Business Machines Corporation | Video memory interface to control processor access to video memory |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4453214A (en) * | 1981-09-08 | 1984-06-05 | Sperry Corporation | Bus arbitrating circuit |
DE3543894A1 (en) * | 1985-12-12 | 1987-06-19 | Wabco Westinghouse Fahrzeug | OPERATING DEVICE FOR A DIFFERENTIAL LOCK |
US4953101A (en) * | 1987-11-24 | 1990-08-28 | Digital Equipment Corporation | Software configurable memory architecture for data processing system having graphics capability |
US5072420A (en) * | 1989-03-16 | 1991-12-10 | Western Digital Corporation | FIFO control architecture and method for buffer memory access arbitration |
US5345577A (en) * | 1989-10-13 | 1994-09-06 | Chips & Technologies, Inc. | Dram refresh controller with improved bus arbitration scheme |
JPH03254497A (en) * | 1990-03-05 | 1991-11-13 | Mitsubishi Electric Corp | Microcomputer |
JPH0635867A (en) * | 1992-07-13 | 1994-02-10 | Mita Ind Co Ltd | Picture data processing circuit and access method to storage means for the processing circuit |
US5450542A (en) * | 1993-11-30 | 1995-09-12 | Vlsi Technology, Inc. | Bus interface with graphics and system paths for an integrated memory system |
-
1995
- 1995-06-07 US US08/487,117 patent/US5767866A/en not_active Expired - Fee Related
-
1996
- 1996-05-31 JP JP50089297A patent/JP3557619B2/en not_active Expired - Fee Related
- 1996-05-31 WO PCT/US1996/008145 patent/WO1996041325A1/en not_active Application Discontinuation
- 1996-05-31 EP EP96916845A patent/EP0834168B1/en not_active Expired - Lifetime
- 1996-05-31 KR KR1019970709197A patent/KR19990022755A/en not_active Application Discontinuation
- 1996-05-31 DE DE69604667T patent/DE69604667T2/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4009470A (en) * | 1975-02-18 | 1977-02-22 | Sperry Rand Corporation | Pre-emptive, rotational priority system |
US4845661A (en) * | 1985-08-19 | 1989-07-04 | Nec Corporation | Display information processing apparatus |
EP0255436A1 (en) * | 1986-07-28 | 1988-02-03 | Bull S.A. | Graphics board with keyboard and mouse interface |
US4991112A (en) * | 1987-12-23 | 1991-02-05 | U.S. Philips Corporation | Graphics system with graphics controller and DRAM controller |
US4942553A (en) * | 1988-05-12 | 1990-07-17 | Zilog, Inc. | System for providing notification of impending FIFO overruns and underruns |
EP0522697A1 (en) * | 1991-06-10 | 1993-01-13 | International Business Machines Corporation | Video memory interface to control processor access to video memory |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5877741A (en) * | 1995-06-07 | 1999-03-02 | Seiko Epson Corporation | System and method for implementing an overlay pathway |
EP0802519A1 (en) * | 1996-04-19 | 1997-10-22 | Seiko Epson Corporation | System and method for implementing an overlay pathway |
WO1999028893A1 (en) * | 1997-12-01 | 1999-06-10 | Mediaq, Inc. | High performance, highly integrated computer architecture with extendible embedded memory |
JPH11345165A (en) * | 1997-12-05 | 1999-12-14 | Texas Instr Inc <Ti> | Traffic controller using priority and burst control for reducing access times |
EP0945843A1 (en) * | 1998-03-26 | 1999-09-29 | Alcatel | Method of controlling a liquid crystal display |
FR2776814A1 (en) * | 1998-03-26 | 1999-10-01 | Alsthom Cge Alcatel | METHOD FOR CONTROLLING A LIQUID CRYSTAL DISPLAY |
US6944696B2 (en) * | 1999-12-27 | 2005-09-13 | Renesas Technology Corp. | Data processing processor |
Also Published As
Publication number | Publication date |
---|---|
JPH11511263A (en) | 1999-09-28 |
US5767866A (en) | 1998-06-16 |
EP0834168A1 (en) | 1998-04-08 |
DE69604667D1 (en) | 1999-11-18 |
KR19990022755A (en) | 1999-03-25 |
JP3557619B2 (en) | 2004-08-25 |
EP0834168B1 (en) | 1999-10-13 |
DE69604667T2 (en) | 2000-03-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US5694141A (en) | Computer system with double simultaneous displays showing differing display images | |
US5767866A (en) | Computer system with efficient DRAM access | |
US5877741A (en) | System and method for implementing an overlay pathway | |
US6204864B1 (en) | Apparatus and method having improved memory controller request handler | |
JP3926417B2 (en) | Display control device | |
US5754170A (en) | Transparent blocking of CRT refresh fetches during video overlay using dummy fetches | |
EP0261751A2 (en) | Concurrent memory access system | |
US5894297A (en) | Display apparatus | |
US5001652A (en) | Memory arbitration for video subsystems | |
EP0740284B9 (en) | Liquid crystal display device | |
EP0464620B1 (en) | Image information control apparatus and display system | |
JP2000035778A (en) | Memory control system, memory control method and memory controller | |
US6967661B2 (en) | Computer system which scans lines in tiled blocks of a display area | |
EP0283565B1 (en) | Computer system with video subsystem | |
JP2667817B2 (en) | Graphic processing device and information processing system using the same | |
US6542159B1 (en) | Apparatus to control memory accesses in a video system and method thereof | |
JPH039392A (en) | Led display device | |
JPH0441832B2 (en) | ||
JPH05333829A (en) | Information processor | |
JPH04344385A (en) | Semiconductor storage device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): JP KR |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWE | Wipo information: entry into national phase |
Ref document number: 1996916845 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref country code: JP Ref document number: 1997 500892 Kind code of ref document: A Format of ref document f/p: F |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1019970709197 Country of ref document: KR |
|
WWP | Wipo information: published in national office |
Ref document number: 1996916845 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 1019970709197 Country of ref document: KR |
|
WWG | Wipo information: grant in national office |
Ref document number: 1996916845 Country of ref document: EP |
|
WWW | Wipo information: withdrawn in national office |
Ref document number: 1019970709197 Country of ref document: KR |