WO2001001575A1 - System memory access system and method for reconfigurable chip - Google Patents

System memory access system and method for reconfigurable chip Download PDF

Info

Publication number
WO2001001575A1
WO2001001575A1 PCT/US2000/017892 US0017892W WO0101575A1 WO 2001001575 A1 WO2001001575 A1 WO 2001001575A1 US 0017892 W US0017892 W US 0017892W WO 0101575 A1 WO0101575 A1 WO 0101575A1
Authority
WO
WIPO (PCT)
Prior art keywords
memory
units
local memory
data
local
Prior art date
Application number
PCT/US2000/017892
Other languages
French (fr)
Other versions
WO2001001575A9 (en
Inventor
Bradley L. Taylor
Original Assignee
Chameleon Systems, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chameleon Systems, Inc. filed Critical Chameleon Systems, Inc.
Priority to AU58991/00A priority Critical patent/AU5899100A/en
Priority to EP00944986A priority patent/EP1198888A4/en
Publication of WO2001001575A1 publication Critical patent/WO2001001575A1/en
Publication of WO2001001575A9 publication Critical patent/WO2001001575A9/en

Links

Classifications

    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03K—PULSE TECHNIQUE
    • H03K19/00—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits
    • H03K19/02—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components
    • H03K19/173—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components
    • H03K19/177—Logic circuits, i.e. having at least two inputs acting on one output; Inverting circuits using specified components using elementary logic circuits as components arranged in matrix form
    • H03K19/17748—Structural details of configuration resources
    • H03K19/1776—Structural details of configuration resources for memories

Definitions

  • the present invention relates to reconfigurable logic chips.
  • Reconfigurable logic chips such as field programmable gate arrays (FPGAs) have become increasingly popular. Such chips allow logic to implement different circuits at different times.
  • FPGAs field programmable gate arrays
  • FPGAs are also being increasingly used because they offer greater flexibility and shorter development cycles than traditional Application Specific
  • ASICs Integrated Circuits
  • reconfigurable computing One growing popular use of FPGAs is referred to as reconfigurable computing.
  • reconfigurable computing hardware logic functions are loaded into the FPGA as needed to implement different sections of a computationally intensive code.
  • advantages are obtained over dedicated processors.
  • Reconfigurable computing is being pursued by university researchers as well as FPGA companies.
  • a problem with typical FPGAs concerns memory access.
  • One common FPGA memory layout uses a central memory. Unfortunately, accesses to this central memory can form a bottleneck. A large number of access lines are required and, even then, only a certain level of concurrent access is supported.
  • the present invention uses dedicated local memory units distributed throughout the reconfigurable logic.
  • the local memory units have system bus ports which allow for quick and easy access with an external system memory.
  • the system bus ports are in addition to local bus ports of the local memory units.
  • the system bus ports allow for system memory transfers done by a direct memory access (DMA) controller on the reconfigurable chip.
  • DMA direct memory access
  • data path units on the reconfigurable chip can initiate the transfer of memory between the system memory and the local memory units.
  • Data path units are reconfigurable elements that execute a number of functions. The initiation of the block data transfer can be the result of a data path unit instruction.
  • the data path units can calculate when to transfer data between the system memory and the local memory units, for example, upon an overflow or underflow of a local memory unit.
  • the data path units and local memory units are associated with circuitry to do an automatic cache-like transfer of data between the local memory units and the system memory.
  • This transfer can greatly simplify the programming of the reconfigurable chips. With the cache-like transfer of data, the programming of the reconfigurable chips can substantially ignore the small size of the local memory units. Portions of the system memory can be automatically loaded in and out of the local memory units without requiring programmed steps. This cache-like system is aided by the use of the system memory ports on the local memory units.
  • Another embodiment of the present invention involves implementing a memory structure using the reconfigurable chip.
  • the implemented memory structure uses local memory units on the reconfigurable chip along with an external memory.
  • data is swapped between the local memory unit and the system memory. This swapping is made easier by the use of the system memory port on the local memory units.
  • the implemented memory structure can be made quite large, and is not limited by the size of the local memory units on the reconfigurable chip.
  • the swapping of data in and out of local memory units can occur concurrently with the operation of other local memory units in the implemented memory structure.
  • data path units can calculate when to swap out or swap in data to and from the local memory units.
  • One example of an implemented memory structure is a First-In-First-Out
  • the FIFO buffer is implemented using local memory units acting as a head and tail of the FIFO buffer.
  • the middle of the implemented FIFO buffer can be stored in the system memory.
  • Figure 1 is an overview of the reconfigurable chip of one embodiment of the present invention.
  • Figure 2 is a diagram showing the arrangement of a slice for the reconfigurable chip of one embodiment of the present invention.
  • Figure 3 is a diagram which illustrates the operation of the data processing and a local memory unit block for the present invention.
  • Figure 4 is an alternate embodiment of a system using the data path unit and local memory unit of the present invention.
  • Figure 5 is a flow chart that illustrates the operation of a portion of the apparatus in Figure 4.
  • Figure 6 is a diagram that illustrates the interconnection between the local memory units and the data path units.
  • Figure 7 is a diagram that illustrates the interconnection between the local memory units and the system memory busses.
  • Figures 8A and 8B are illustrations that show the arrangement of the local and global address.
  • Figure 9 is a diagram illustrating the interconnection between the local memory units and the system memory units in one embodiment of the present invention.
  • Figure 10 is a diagram that shows a top-level view of the use of local memory units and system memory to implement a FIFO buffer.
  • Figures 11 and 12 are flow charts which illustrate the operation of the implemented FIFO buffer of figure 10.
  • Figure 13 illustrates one method of interconnecting local memory units in order to implement a FIFO buffer.
  • Figures 14 and 15 illustrate two embodiments of the data path unit interacting with the DMA controller.
  • FIG. 1 is a diagram which illustrates a reconfigurable chip 20 and external memory 22.
  • the external memory 22 is connected to the memory controller 24.
  • the address and data from the external memory is sent to the system data bus 26 and system address bus 28.
  • These busses are interconnected to a Direct Memory Access (DMA) control/bus interface 30 connected to the local memory units.
  • the local memory units are arranged in slices interleaved with the data path units. Slices 32, 34, 36, and 38 are shown. Each of the slices has a configuration memory 32a, 34a, 36a, and 38a. This configuration memory is loaded from the configuration buffer 40.
  • the central processing unit 42 and the fabric setup and control logic 44 is also shown in this system, and the fabric setup and control logic 44.
  • the configuration memory stores the arrangement of configurable logic and the interconnections between the configurable logic and the local memory units. This diagram does not show the input and output to the slices or the interconnection between the data path units of the slices.
  • FIG. 2 illustrates one slice 50 of the reconfigurable chip of Figure 1.
  • the slice 50 is divided into tiles 50a, 50b, 50c, and 50d.
  • Each of the tiles contains a number of local memory units 52, and data path units 54 and 56.
  • Some of the data path units are adapted to generate local read/write addresses.
  • the data path units can execute a number of functions on the reconfigurable chip.
  • Figure 3 is a diagram illustrating the interconnection of a data path unit element with the local memory unit.
  • the local memory unit is attached to the system bus 62, local read bus 64, and local write bus 66.
  • the system bus 62 allow the local memory unit to be read from and write to a system memory independent from the local read and write accesses.
  • Reconfigurable connections 68 on the read and write busses 64 and 66 allow the local memory units and the data path units to be interconnected together in a variety of fashions.
  • Reconfigurable connections on the local bus lines allow segments of the bus line to be electrically connected or isolated as desired.
  • the reconfigurable connections can be implemented as transmission gates or pairs of directional gated buffers. Such reconfigurable connections are provided for each bit of the bus lines.
  • the data path unit includes an operation unit 70 that can implement a number of functions.
  • the input data to the operation unit comes from registers 72 and 74.
  • the output of operation unit 70 is stored into register 76.
  • Data path units have a connection to the local memory unit 60.
  • the operation unit 70 can provide a write address to the local memory unit, a write address to the write address portion of the write bus 66, data to the write data 68 portion of the local memory unit, or data to the write data portion of the write bus line 66.
  • the operation unit 70 can also provide the read address to the read address portion of the local memory unit 60 or to the read address portion of the read bus 64.
  • the read data from the local memory unit can be sent out to the read bus 64 or to the mutiplexer 78 to provide the output of the data path unit.
  • a number of gated buffers 80, 82, 84, and 86 are used to arrange the interconnection between the operation unit and the local busses 64 and 66.
  • the gated buffer 88 is used to send a request on line 90 to DMA controller on the reconfigurable chip in order to send memory to local memory unit 60 or another local memory unit on the reconfigurable chip.
  • the DMA transfer requires the source address (the local memory unit or system memory), the destination address (system memory or local memory unit), and the size of the data transfer. This information can be provided in three cycles. After this information is sent to the DMA controller, the DMA controller transfers data between the local memory unit and the system memory. The DMA transfer need not be to or from a local memory unit adjacent to the data path unit making the request; local memory units anywhere in the computing fabric can be accessed in the DMA transfer.
  • Figure 14 shows another embodiment for the data unit transfer.
  • the data path unit 300 is connected to the DMA controller 302.
  • DMA request goes to DMA controller 302.
  • the DMA controller then obtains the source address, destination address, count and mode from registers in the data path unit. Once the data transfer is finished the DMA controller 302 sends an acknowledge signal to the data path unit 300.
  • Figure 15 shows an alternate embodiment.
  • Data path unit 304 is connected to the DMA controller 306. When a DMA request is sent from the data path unit 304, the DMA controller 306 checks in one of the local memory units 308 to obtain the source address, data address, count and mode.
  • the mode can contain an indicator which can control the DMA controller to move on to the next group of source address, data address, count and mode without requiring the intervention of the data path unit.
  • the DMA controller continues to do the data transfers to the local memory units until each of the requests in the local memory are finished.
  • each line of the local memory unit can store a source address, data address, count or mode. In this instance, four lines will store all of the information for a transfer and the local memory unit can store up to 16 groups of the transfer information.
  • FIG. 4 illustrates an alternate embodiment that uses the cache-like controller 92.
  • the cache-like controller 92 when data is stored in the local system memory from the system memory, an indication of the higher bits of the system memory address are stored in the tag memory 94.
  • the operation unit 70' accesses memory, the memory address is compared to the address or addresses stored in the tag memory using comparator 96. If there is a match, the system accesses the local memory unit in the normal fashion.
  • the tag memory can store a single tag value or it can be arranged to store multiple tag values.
  • FIG. 5 is a flow chart illustrating the operation of the apparatus in Figure 4.
  • a data path unit instruction is obtained.
  • the instruction is checked to see whether the data processing unit accesses data stored in the memory. If not, no further operation occurs with respect to the cache unit. If the data path unit instruction accesses the local memory, step 104 checks whether the top bits of the address match the tag memory values stored in the tag region. If these bits match, the access to the memory access is done in step 106. If not, execution is halted, data is swapped in from the system memory and the execution restarted in step 108. In step 110, after the execution is restarted, the memory access is repeated in the local memory.
  • Figure 6 illustrates the local bus connections between the local memory units 120, 122, and 124, and the data path units 126, 128, and 130.
  • Reconfigurable connections 132 are used to interconnect the local memory units.
  • Gated buffers 134 are also used to aid in interconnection. Details of the interconnection system of local memory units is described in the patent application "A Hierarchical Distributed Memory System for Realization of Localization of Variable Requirements for an Adaptive Computing Fabric", inventors Chris Phillips, et al. , filed 16 June 1999, Serial No. 09/333,977 (corresponding to attorney docket no. 032001-015), which is incorporated herein by reference.
  • Figure 7 shows one embodiment of a system interconnect between the local memory units and a system bus 140.
  • the local memory units 142 and 144 are connected to the system bus using buffers 146 and multiplexers 148; and multiplexers 150 and buffers 152.
  • the connection of the local memory units to the system memory is such that more than one element in the local memory unit can be loaded at a time. Such a system speeds up the switching of memory in and out to the local memory unit.
  • FIG. 8A and 8B illustrate the arrangement of the local and global addresses.
  • the local address 150 includes a Field 150a, contains the position within an element of a local memory unit.
  • the element ID bits 150b indicate an element within the N x M local memory unit.
  • Bits 150c corresponds to local memory ID bits. These bits are compared with identity bits stored in the local memory units. This allows the local memory units to be connected together as described in the patent application "A Hierarchical Distributed Memory System for Realization of Localization of Variable Requirements for an Adaptive Computing Fabric", discussed above.
  • the global address 152 is shown in figure 8B.
  • the global address 152 includes Field 152a which indicates a position within the multiple element block in the global address.
  • the block ID bits 152b show the block number within the local memory.
  • the local memory ID field 150c can be the fixed local memory identity when used with the memory scheme of Figure 3, or can be the tag bits which are used in the memory scheme of Figure 4.
  • Figure 9 illustrates use of the system of the present invention.
  • the system memory 160 is connected by the system bus 162 to a local memory unit 164.
  • This local memory unit 164 has local bus 166 connections to configured logic 168, which then is connected on local bus 170 to local memory 172.
  • Local memory 172 is connected out of the system bus 162 to the system memory 160.
  • Figure 9 illustrates the use of the local memory units to act as buffers for the system memory.
  • the system bus 162 allows quick access between the local memory units 164 and 172 and the system memory.
  • Figures 10-13 show how one memory structure, such as a FIFO buffer, can be implemented using the local memory units and the system memory.
  • Figure 10 is a diagram that illustrates an implemented first-in-first-out (FIFO) buffer 180.
  • the local memory 184 has a number of local memory units which are used in the FIFO buffer.
  • Local memory units A and B are used to store the data as it enters the FIFO buffer.
  • the head address shows the local address of the position to write into the buffer.
  • the local memory unit B is filled up, the system switches to write memory into the local memory unit A.
  • the contents of the local memory unit B is swapped out to the system memory block CII. This swapping step can be done concurrently with data being written into local memory unit A.
  • the local memory unit B By the time local memory unit A is filled with data, the local memory unit B will again be available to have data written into it.
  • the local memory units, C and D contain the data to be written out of the FIFO buffer.
  • the local memory unit C When the local memory unit C is completely emptied of data, data is written out of the local memory unit D. Thereafter, data from the system memory block I can be swapped into the local memory unit C.
  • the use of the system memory port in the local memory units allows for a quick swapping in and out of data between the local memory and the system memory.
  • figure 10 shows an implementation of a FIFO buffer
  • additional memory structure implementations such as tree structures, linked lists, or large tables can also be implemented.
  • These implemented memory structures also use local memory units and the system memory, where the reconfigurable chip causes data to be swapped in and out of the local memory units from and to the system memory.
  • FIGS 11 and 12 are flow charts illustrating the operation of the FIFO buffer.
  • step 190 indicates the inputting of data to the FIFO buffer.
  • Step 192 checks whether the FIFO buffer has more than 4 blocks of data to determine if the FIFO buffer is completely contained within the local memory units. If the buffer is not using more than 4 blocks, the data is added to the local memory units in the conventional manner in step 194. If the buffer is using more than 4 blocks, in step 194, it is checked whether the data will fill the current input unit. If not, data is added to the current input local memory unit at the head address, in step 196 and the head address is updated, in step 198. If the data will fill the current local memory unit, the data is added to the current input local memory unit in step 200.
  • step 202 the data in the current input local memory unit is swapped out to the system memory. Step 202 can be done concurrently with other steps at any time before the next branch 199.
  • step 204 data is checked whether the current input local memory unit is local memory unit A. If not, the current input local memory unit is set to a local memory unit A in step 206. If the current input local memory unit is local memory unit A, then the current input local memory is set to local memory unit B, in step 208. In step 210, the head address is updated.
  • Figure 12 is a flow chart illustrating the output operation.
  • step 214 it is checked whether the FIFO buffer is using more than 4 blocks. If not, data is read normally from the local memory units, in step 216.
  • Step 218 checks whether the output of data will empty the current output local memory unit. If not, data is read out of the current local memory unit at the tail address in step 220. In step 222, the tail address is updated. If the output data will empty the current local memory unit, data is read from the current local memory unit in step 224.
  • step 226 data is swapped-in from the system memory to the current output local memory unit. Step 226 can be done concurrently with the other steps until the next time the system goes to branch 223.
  • the step 228 checks whether the current output local memory unit is local memory unit C.
  • step 230 the current output local memory unit is set to local memory unit C. If the current output local memory unit is local memory unit C, the current output local memory unit is switched to local memory unit D. In step, 234, the head address is updated. Note that the actions described in Figures 11 and 12 of inputting and outputting of data from the FIFO could be implemented independently in the data path units, as long as the head and tail address do not interfere with one another. This can speed the operation of the FIFO buffer. The operation of many local memory units and many data path units concurrently within the same reconfigurable chip can produce considerable computation time advantages for reconfigurable computing.
  • Figure 13 illustrates one implementation of a FIFO buffer with the system of the present invention.
  • the data and local memory units are connected using a system and data bus to a data memory access controller 240, which connects to the system memory 242. Other local connections are interconnected as shown.
  • the write data can be sent to any of the local memory units shown.
  • the output data from any of the units shown, and the write and read addresses are sent to all the units.
  • local memory units A and B are used for writing data in
  • local memory units C and D are used for writing data out of the FIFO buffer.
  • the local memory units A, B, C, and D are given different identity bits so that the local write and read addresses can point to specific local memory units.
  • the input address logic 244 is implemented with data path units to provide the input address.
  • a data path unit can store the current input address and while the address is being input to the same local memory unit increment the write address after each write to the FIFO buffer. Each time, a new local memory unit is set as the current input local memory unit, the identity bits for the current input local memory unit are masked onto the write address. This can be done by data path units which implement a state machine, storing which memory unit is the current memory unit and indicating which input will cause the current input local memory to switch.
  • Data path units can also implement the system swap logic 246.
  • the system swap logic 246 causes data to be swapped from a full local memory unit A or B to the system memory and to be loaded to an empty local memory unit C or D from the system memory.
  • the address logic for the output address 248 can also be implement with the data path units. As described above, the output address will decrement as data is read from the local memory unit. When a local memory unit is empty of data, the current output local memory unit is switched and the read address is masked with byte data bits for the new current output address. Since the swap logic 246 can cause the data in the local memory units to be swapped to or from the system memory 242, the implemented FIFO buffer can be as large as the system memory. The swapping of data between the system memory and the local memory units can be done behind the scenes which allows for very fast access to the memory structure. Even though the implemented memory structure can be very large, it will operate very fast.

Landscapes

  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Logic Circuits (AREA)

Abstract

A memory access system is described in which local memory units (60) on a reconfigurable chip can be used in conjunction with the system memory. Data path units (90) on the reconfigurable chip can cause data to be swapped in and out of the local memory units (60) as a result of calculations within the reconfigurable fabric. Alternately, a cache-like system can be used so that the data can be read into the local memory unit (60) from the system memory units automatically.

Description

SYSTEM MEMORY ACCESS SYSTEM AND METHOD FOR RECONFIGURABLE CHIP
Background of the Invention
The present invention relates to reconfigurable logic chips. Reconfigurable logic chips, such as field programmable gate arrays (FPGAs) have become increasingly popular. Such chips allow logic to implement different circuits at different times.
FPGAs are also being increasingly used because they offer greater flexibility and shorter development cycles than traditional Application Specific
Integrated Circuits (ASICs), while providing most of the performance advantages of a dedicated hardware solution.
One growing popular use of FPGAs is referred to as reconfigurable computing. In reconfigurable computing, hardware logic functions are loaded into the FPGA as needed to implement different sections of a computationally intensive code. By using the FPGAs to do the computational intensive code, advantages are obtained over dedicated processors. Reconfigurable computing is being pursued by university researchers as well as FPGA companies.
A problem with typical FPGAs concerns memory access. One common FPGA memory layout uses a central memory. Unfortunately, accesses to this central memory can form a bottleneck. A large number of access lines are required and, even then, only a certain level of concurrent access is supported.
Other systems create memory units from reconfigurable logic. A disadvantage of these systems is that a large amount of chip "real estate" is required for the reconfigurable logic to implement a memory system. Additionally, programming the reconfigurable logic to implement the memory can be quite complicated.
It is desired to have an improved memory system for a reconfigurable chip.
Summary
The present invention uses dedicated local memory units distributed throughout the reconfigurable logic. The local memory units have system bus ports which allow for quick and easy access with an external system memory. The system bus ports are in addition to local bus ports of the local memory units. The system bus ports allow for system memory transfers done by a direct memory access (DMA) controller on the reconfigurable chip.
In one embodiment, data path units on the reconfigurable chip can initiate the transfer of memory between the system memory and the local memory units. Data path units are reconfigurable elements that execute a number of functions. The initiation of the block data transfer can be the result of a data path unit instruction. The data path units can calculate when to transfer data between the system memory and the local memory units, for example, upon an overflow or underflow of a local memory unit.
In an alternate embodiment, the data path units and local memory units are associated with circuitry to do an automatic cache-like transfer of data between the local memory units and the system memory. This transfer can greatly simplify the programming of the reconfigurable chips. With the cache-like transfer of data, the programming of the reconfigurable chips can substantially ignore the small size of the local memory units. Portions of the system memory can be automatically loaded in and out of the local memory units without requiring programmed steps. This cache-like system is aided by the use of the system memory ports on the local memory units.
Another embodiment of the present invention involves implementing a memory structure using the reconfigurable chip. The implemented memory structure uses local memory units on the reconfigurable chip along with an external memory. During the operation of the memory structure, data is swapped between the local memory unit and the system memory. This swapping is made easier by the use of the system memory port on the local memory units. The implemented memory structure can be made quite large, and is not limited by the size of the local memory units on the reconfigurable chip. The swapping of data in and out of local memory units can occur concurrently with the operation of other local memory units in the implemented memory structure. In the preferred embodiment, data path units can calculate when to swap out or swap in data to and from the local memory units. One example of an implemented memory structure is a First-In-First-Out
(FIFO) buffer. The FIFO buffer is implemented using local memory units acting as a head and tail of the FIFO buffer. The middle of the implemented FIFO buffer can be stored in the system memory.
Brief description of the drawings
Figure 1 is an overview of the reconfigurable chip of one embodiment of the present invention.
Figure 2 is a diagram showing the arrangement of a slice for the reconfigurable chip of one embodiment of the present invention.
Figure 3 is a diagram which illustrates the operation of the data processing and a local memory unit block for the present invention. Figure 4 is an alternate embodiment of a system using the data path unit and local memory unit of the present invention.
Figure 5 is a flow chart that illustrates the operation of a portion of the apparatus in Figure 4. Figure 6 is a diagram that illustrates the interconnection between the local memory units and the data path units.
Figure 7 is a diagram that illustrates the interconnection between the local memory units and the system memory busses.
Figures 8A and 8B are illustrations that show the arrangement of the local and global address.
Figure 9 is a diagram illustrating the interconnection between the local memory units and the system memory units in one embodiment of the present invention.
Figure 10 is a diagram that shows a top-level view of the use of local memory units and system memory to implement a FIFO buffer.
Figures 11 and 12 are flow charts which illustrate the operation of the implemented FIFO buffer of figure 10.
Figure 13 illustrates one method of interconnecting local memory units in order to implement a FIFO buffer. Figures 14 and 15 illustrate two embodiments of the data path unit interacting with the DMA controller.
Detail Description of the Preferred Embodiment
Figure 1 is a diagram which illustrates a reconfigurable chip 20 and external memory 22. The external memory 22 is connected to the memory controller 24. The address and data from the external memory is sent to the system data bus 26 and system address bus 28. These busses are interconnected to a Direct Memory Access (DMA) control/bus interface 30 connected to the local memory units. In the embodiment shown, the local memory units are arranged in slices interleaved with the data path units. Slices 32, 34, 36, and 38 are shown. Each of the slices has a configuration memory 32a, 34a, 36a, and 38a. This configuration memory is loaded from the configuration buffer 40. Also shown in this system, is the central processing unit 42 and the fabric setup and control logic 44. The configuration memory stores the arrangement of configurable logic and the interconnections between the configurable logic and the local memory units. This diagram does not show the input and output to the slices or the interconnection between the data path units of the slices. The connection to the external memory through the DMA control/ bus interface 30, as will be discussed below, allows for the transferring of blocks of data between the external memory and the local memory units.
Figure 2 illustrates one slice 50 of the reconfigurable chip of Figure 1. The slice 50 is divided into tiles 50a, 50b, 50c, and 50d. Each of the tiles contains a number of local memory units 52, and data path units 54 and 56. Some of the data path units are adapted to generate local read/write addresses. The data path units can execute a number of functions on the reconfigurable chip.
Figure 3 is a diagram illustrating the interconnection of a data path unit element with the local memory unit. The local memory unit is attached to the system bus 62, local read bus 64, and local write bus 66. The system bus 62 allow the local memory unit to be read from and write to a system memory independent from the local read and write accesses. Reconfigurable connections 68 on the read and write busses 64 and 66 allow the local memory units and the data path units to be interconnected together in a variety of fashions.
Reconfigurable connections on the local bus lines allow segments of the bus line to be electrically connected or isolated as desired. The reconfigurable connections can be implemented as transmission gates or pairs of directional gated buffers. Such reconfigurable connections are provided for each bit of the bus lines.
The data path unit includes an operation unit 70 that can implement a number of functions. The input data to the operation unit comes from registers 72 and 74. The output of operation unit 70 is stored into register 76. Data path units have a connection to the local memory unit 60. The operation unit 70 can provide a write address to the local memory unit, a write address to the write address portion of the write bus 66, data to the write data 68 portion of the local memory unit, or data to the write data portion of the write bus line 66. The operation unit 70 can also provide the read address to the read address portion of the local memory unit 60 or to the read address portion of the read bus 64. The read data from the local memory unit can be sent out to the read bus 64 or to the mutiplexer 78 to provide the output of the data path unit. A number of gated buffers 80, 82, 84, and 86 are used to arrange the interconnection between the operation unit and the local busses 64 and 66.
The gated buffer 88 is used to send a request on line 90 to DMA controller on the reconfigurable chip in order to send memory to local memory unit 60 or another local memory unit on the reconfigurable chip.
In one embodiment, the DMA transfer requires the source address (the local memory unit or system memory), the destination address (system memory or local memory unit), and the size of the data transfer. This information can be provided in three cycles. After this information is sent to the DMA controller, the DMA controller transfers data between the local memory unit and the system memory. The DMA transfer need not be to or from a local memory unit adjacent to the data path unit making the request; local memory units anywhere in the computing fabric can be accessed in the DMA transfer.
Figure 14 shows another embodiment for the data unit transfer. The data path unit 300 is connected to the DMA controller 302. DMA request goes to DMA controller 302. The DMA controller then obtains the source address, destination address, count and mode from registers in the data path unit. Once the data transfer is finished the DMA controller 302 sends an acknowledge signal to the data path unit 300. Figure 15 shows an alternate embodiment. Data path unit 304 is connected to the DMA controller 306. When a DMA request is sent from the data path unit 304, the DMA controller 306 checks in one of the local memory units 308 to obtain the source address, data address, count and mode. In this system the mode can contain an indicator which can control the DMA controller to move on to the next group of source address, data address, count and mode without requiring the intervention of the data path unit. In this embodiment the DMA controller continues to do the data transfers to the local memory units until each of the requests in the local memory are finished. In one embodiment, each line of the local memory unit can store a source address, data address, count or mode. In this instance, four lines will store all of the information for a transfer and the local memory unit can store up to 16 groups of the transfer information.
The advantage of calculating the source address, data address, count and mode within the data path unit is that these calculations can be done in parallel. This can reduce the time required to calculate the address which can be a limiting factor in some data protocols such as the asynchronous transfer mode (ATM). Figure 4 illustrates an alternate embodiment that uses the cache-like controller 92. In this system, when data is stored in the local system memory from the system memory, an indication of the higher bits of the system memory address are stored in the tag memory 94. When the operation unit 70' accesses memory, the memory address is compared to the address or addresses stored in the tag memory using comparator 96. If there is a match, the system accesses the local memory unit in the normal fashion. If there is not a match, the system operations halts, and a request is sent across request line 90' to cause data from the system memory to be loaded into the local memory unit. Such an operation is similar to a cache system used with microprocessors. However, the present invention uses a large number of the local memory units distributed about the reconfigurable chip. The tag memory can store a single tag value or it can be arranged to store multiple tag values.
Figure 5 is a flow chart illustrating the operation of the apparatus in Figure 4. In step 100, a data path unit instruction is obtained. In step 102, the instruction is checked to see whether the data processing unit accesses data stored in the memory. If not, no further operation occurs with respect to the cache unit. If the data path unit instruction accesses the local memory, step 104 checks whether the top bits of the address match the tag memory values stored in the tag region. If these bits match, the access to the memory access is done in step 106. If not, execution is halted, data is swapped in from the system memory and the execution restarted in step 108. In step 110, after the execution is restarted, the memory access is repeated in the local memory.
Figure 6 illustrates the local bus connections between the local memory units 120, 122, and 124, and the data path units 126, 128, and 130. Reconfigurable connections 132 are used to interconnect the local memory units. Gated buffers 134 are also used to aid in interconnection. Details of the interconnection system of local memory units is described in the patent application "A Hierarchical Distributed Memory System for Realization of Localization of Variable Requirements for an Adaptive Computing Fabric", inventors Chris Phillips, et al. , filed 16 June 1999, Serial No. 09/333,977 (corresponding to attorney docket no. 032001-015), which is incorporated herein by reference. Figure 7 shows one embodiment of a system interconnect between the local memory units and a system bus 140. The local memory units 142 and 144 are connected to the system bus using buffers 146 and multiplexers 148; and multiplexers 150 and buffers 152. The connection of the local memory units to the system memory is such that more than one element in the local memory unit can be loaded at a time. Such a system speeds up the switching of memory in and out to the local memory unit.
Figure 8A and 8B illustrate the arrangement of the local and global addresses. The local address 150 includes a Field 150a, contains the position within an element of a local memory unit. The element ID bits 150b indicate an element within the N x M local memory unit. Bits 150c corresponds to local memory ID bits. These bits are compared with identity bits stored in the local memory units. This allows the local memory units to be connected together as described in the patent application "A Hierarchical Distributed Memory System for Realization of Localization of Variable Requirements for an Adaptive Computing Fabric", discussed above.
The global address 152 is shown in figure 8B. The global address 152 includes Field 152a which indicates a position within the multiple element block in the global address. The block ID bits 152b show the block number within the local memory. The local memory ID field 150c can be the fixed local memory identity when used with the memory scheme of Figure 3, or can be the tag bits which are used in the memory scheme of Figure 4.
Figure 9 illustrates use of the system of the present invention. The system memory 160 is connected by the system bus 162 to a local memory unit 164. This local memory unit 164 has local bus 166 connections to configured logic 168, which then is connected on local bus 170 to local memory 172. Local memory 172 is connected out of the system bus 162 to the system memory 160. Figure 9 illustrates the use of the local memory units to act as buffers for the system memory. The system bus 162 allows quick access between the local memory units 164 and 172 and the system memory.
Figures 10-13 show how one memory structure, such as a FIFO buffer, can be implemented using the local memory units and the system memory. Figure 10 is a diagram that illustrates an implemented first-in-first-out (FIFO) buffer 180. The local memory 184 has a number of local memory units which are used in the FIFO buffer. Local memory units A and B are used to store the data as it enters the FIFO buffer. The head address shows the local address of the position to write into the buffer. When the local memory unit B is filled up, the system switches to write memory into the local memory unit A. The contents of the local memory unit B is swapped out to the system memory block CII. This swapping step can be done concurrently with data being written into local memory unit A. By the time local memory unit A is filled with data, the local memory unit B will again be available to have data written into it. The local memory units, C and D contain the data to be written out of the FIFO buffer. When the local memory unit C is completely emptied of data, data is written out of the local memory unit D. Thereafter, data from the system memory block I can be swapped into the local memory unit C. The use of the system memory port in the local memory units, allows for a quick swapping in and out of data between the local memory and the system memory.
Although figure 10 shows an implementation of a FIFO buffer, additional memory structure implementations such as tree structures, linked lists, or large tables can also be implemented. These implemented memory structures also use local memory units and the system memory, where the reconfigurable chip causes data to be swapped in and out of the local memory units from and to the system memory.
Figures 11 and 12 are flow charts illustrating the operation of the FIFO buffer. Looking at figure 11, step 190 indicates the inputting of data to the FIFO buffer. Step 192 checks whether the FIFO buffer has more than 4 blocks of data to determine if the FIFO buffer is completely contained within the local memory units. If the buffer is not using more than 4 blocks, the data is added to the local memory units in the conventional manner in step 194. If the buffer is using more than 4 blocks, in step 194, it is checked whether the data will fill the current input unit. If not, data is added to the current input local memory unit at the head address, in step 196 and the head address is updated, in step 198. If the data will fill the current local memory unit, the data is added to the current input local memory unit in step 200. In step 202, the data in the current input local memory unit is swapped out to the system memory. Step 202 can be done concurrently with other steps at any time before the next branch 199. In step 204, data is checked whether the current input local memory unit is local memory unit A. If not, the current input local memory unit is set to a local memory unit A in step 206. If the current input local memory unit is local memory unit A, then the current input local memory is set to local memory unit B, in step 208. In step 210, the head address is updated.
Figure 12 is a flow chart illustrating the output operation. In step 214, it is checked whether the FIFO buffer is using more than 4 blocks. If not, data is read normally from the local memory units, in step 216. Step 218 checks whether the output of data will empty the current output local memory unit. If not, data is read out of the current local memory unit at the tail address in step 220. In step 222, the tail address is updated. If the output data will empty the current local memory unit, data is read from the current local memory unit in step 224. In step 226, data is swapped-in from the system memory to the current output local memory unit. Step 226 can be done concurrently with the other steps until the next time the system goes to branch 223. The step 228 checks whether the current output local memory unit is local memory unit C. If not, in step 230, the current output local memory unit is set to local memory unit C. If the current output local memory unit is local memory unit C, the current output local memory unit is switched to local memory unit D. In step, 234, the head address is updated. Note that the actions described in Figures 11 and 12 of inputting and outputting of data from the FIFO could be implemented independently in the data path units, as long as the head and tail address do not interfere with one another. This can speed the operation of the FIFO buffer. The operation of many local memory units and many data path units concurrently within the same reconfigurable chip can produce considerable computation time advantages for reconfigurable computing.
Figure 13 illustrates one implementation of a FIFO buffer with the system of the present invention. The data and local memory units are connected using a system and data bus to a data memory access controller 240, which connects to the system memory 242. Other local connections are interconnected as shown. Thus, the write data can be sent to any of the local memory units shown. The output data from any of the units shown, and the write and read addresses are sent to all the units. When the size of the FIFO buffer is greater than the 4 memory units, local memory units A and B are used for writing data in, whereas local memory units C and D are used for writing data out of the FIFO buffer. The local memory units A, B, C, and D are given different identity bits so that the local write and read addresses can point to specific local memory units. The input address logic 244 is implemented with data path units to provide the input address. A data path unit can store the current input address and while the address is being input to the same local memory unit increment the write address after each write to the FIFO buffer. Each time, a new local memory unit is set as the current input local memory unit, the identity bits for the current input local memory unit are masked onto the write address. This can be done by data path units which implement a state machine, storing which memory unit is the current memory unit and indicating which input will cause the current input local memory to switch. Data path units can also implement the system swap logic 246. The system swap logic 246 causes data to be swapped from a full local memory unit A or B to the system memory and to be loaded to an empty local memory unit C or D from the system memory. The address logic for the output address 248 can also be implement with the data path units. As described above, the output address will decrement as data is read from the local memory unit. When a local memory unit is empty of data, the current output local memory unit is switched and the read address is masked with byte data bits for the new current output address. Since the swap logic 246 can cause the data in the local memory units to be swapped to or from the system memory 242, the implemented FIFO buffer can be as large as the system memory. The swapping of data between the system memory and the local memory units can be done behind the scenes which allows for very fast access to the memory structure. Even though the implemented memory structure can be very large, it will operate very fast.
It will be appreciated by those of ordinary skill in the art that the invention can be implemented in other specific forms without departing from the spirit or central character thereof. The presently disclosed embodiments are therefore considered in all respects to be illustrative and not restrictive. The scope of the invention is illustrated by the appended claims rather than the foregoing description, and all changes which come within the meaning and range for equivalence thereof are intended to be embraced herein.

Claims

Claims:
1. A reconfigurable chip comprising: a number of data path units able to implement a number of functions; a number of dedicated local memory units operably connected to data path units, the local memory units storing N elements by M bits, the local memory units including at least one global system port, the local memory units adapted to store data that can be operated upon by data path units; and a global memory access means operably connected to the global system ports for interconnection to a system memory.
2. The reconfigurable chip of claim 1, wherein the data path units can initiate a data transfer between local memory units and a system memory.
3. The reconfigurable chip of claim 2, wherein the data path units use instructions and wherein at least one of the instructions can initiate data transfer between local memory units and a system memory.
4. The reconfigurable chip of claim 1, wherein the system memory is external to the reconfigurable chip.
5. The reconfigurable chip of claim 1, wherein the global memory access means includes a bus operably connected to the global system ports.
6. The reconfigurable chip of claim 5, wherein the global memory access means further includes a direct memory access (DMA) controller connected to the bus.
7. The reconfigurable chip of claim 1, wherein the local memory units are arranged with fixed memory addresses with respect to the global memory access means.
8. The reconfigurable chip of claim 1, wherein the data path units are operably connected to a request line to initiate a data transfer between local memory units and a system memory.
9. The reconfigurable chip of claim 8, wherein the request line is connected to a direct memory access (DMA) controller.
10. The reconfigurable chip of claim 1 wherein the local memory units further include a local read port and a local write port.
11. The reconfigurable chip of claim 1 , further comprising reconfigurable connections on local bus lines between local memory units, the reconfigurable connections allowing the selective connection of local memory units.
12. The reconfigurable chip of claim 1, wherein the reconfigurable chip is arranged to implement a memory structure.
13. The reconfigurable chip of claim 12, wherein the reconfigurable chip is arranged to implement a first-in-first-out (FIFO) buffer.
14. The reconfigurable chip of claim 12, wherein the data is swapped between the local memory unit and the system memory unit to implement the memory structure.
15. The reconfigurable chip of claim 1, wherein the local memory units act as cache memory units for the system memory.
16. The reconfigurable chip of claim 15, wherein associated with the local memory units are automatic memory swapping units.
17. The reconfigurable chip of claim 15, wherein the automatic memory swapping units include a tag memory and comparator.
18. The reconfigurable chip of claim 15, wherein the system uses valid bits to indicate whether data in the local memory unit is valid.
19. The reconfigurable chip of claim 1, wherein the local memory units are read and write addressable by the data path units.
20. The reconfigurable chip of claim 1, further comprising reconfigurable connections on local bus lines between data path units and local memory units to operably connect the data path units and local memory units, the reconfigurable connections allowing addresses from one of the data path units to be sent to a single local memory unit or a group of local memory units.
21. The reconfigurable chip of claim 1, wherein the data path units and local memory units are arranged into slices.
22. The reconfigurable chip of claim 1, wherein the reconfigurable chip is arranged for reconfigurable computing.
23. A reconfigurable chip comprising: a number of data path units able to implement a number of functions; and a number of dedicated local memory units operably connected to data path units, the local memory units storing N elements by M bits, the local memory units including at least one global system port, wherein the data path unit can initiate a data transfer between local memory units and a system memory.
24. The reconfigurable chip of claim 23, wherein the data path unit can initiate a block data transfer of multiple data elements between local memory units and a system memory.
25. The reconfigurable chip of claim 23, wherein the data path units are operably connected to a request line to initiate a data transfer between local memory units and a system memory.
26. The reconfigurable chip of claim 23, wherein the data path units use instructions and wherein at least one of the instructions can initiate data transfer between local memory units and a system memory.
27. The reconfigurable chip of claim 23, wherein determination of a data transfer is the result of calculation by one of the data path units.
28. The reconfigurable chip of claim 23, wherein the local address, system address, data transfer length indications are sent to a data memory access (DMA) unit from at least one data path unit to implement the data transfer.
29. The reconfigurable chip of claim 23, wherein the data transfer is from the system memory to a local memory unit.
30. The reconfigurable chip of claim 23, further comprising a global memory access means operably connected to the global system ports for interconnection to a system memory.
31. A method of implementing a memory structure on a reconfigurable chip, the reconfigurable chip including a number of dedicated local memory units, the method comprising: implementing a memory structure on the reconfigurable chip, the memory structure including at least one of the local memory units; operating the memory structure; in response to the operating step, swapping data between the at least one of the local memory unit and a system memory.
32. The method of claim 31, wherein the swapping step is automatic.
33. The method of claim 31 , wherein the swapping step uses a tag memory and comparator.
34. The method of claim 31, wherein the swapping step is initiated by a data path unit on the reconfigurable chip.
35. The method of claim 31 , wherein some of the data of the implemented memory structure is stored in the local memory units and other data of the data of the implemented memory structure is stored in the system memory.
36. The method of claim 35, wherein the swapping step comprises moving data out of a local memory unit to the system memory when the local memory unit is full of data for the implemented memory structure.
37. The method of claim 35, wherein the swapping step comprises moving data from the system memory unit to a local memory unit the local memory unit is emptied of data for the implemented memory structure.
38. The method of claim 31, wherein the implemented memory structure is a first-in-first-out (FIFO) buffer.
39. The method of claim 31, wherein the implemented memory structure comprises tree data structure.
40. The method of claim 31 , wherein the implemented memory structure comprises a table data structure.
PCT/US2000/017892 1999-06-30 2000-06-28 System memory access system and method for reconfigurable chip WO2001001575A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
AU58991/00A AU5899100A (en) 1999-06-30 2000-06-28 System memory access system and method for reconfigurable chip
EP00944986A EP1198888A4 (en) 1999-06-30 2000-06-28 System memory access system and method for reconfigurable chip

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US09/343,477 1999-06-30
US09/343,477 US6347346B1 (en) 1999-06-30 1999-06-30 Local memory unit system with global access for use on reconfigurable chips

Publications (2)

Publication Number Publication Date
WO2001001575A1 true WO2001001575A1 (en) 2001-01-04
WO2001001575A9 WO2001001575A9 (en) 2002-07-25

Family

ID=23346270

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2000/017892 WO2001001575A1 (en) 1999-06-30 2000-06-28 System memory access system and method for reconfigurable chip

Country Status (4)

Country Link
US (2) US6347346B1 (en)
EP (1) EP1198888A4 (en)
AU (1) AU5899100A (en)
WO (1) WO2001001575A1 (en)

Families Citing this family (107)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7266725B2 (en) * 2001-09-03 2007-09-04 Pact Xpp Technologies Ag Method for debugging reconfigurable architectures
DE19651075A1 (en) * 1996-12-09 1998-06-10 Pact Inf Tech Gmbh Unit for processing numerical and logical operations, for use in processors (CPU's), multi-computer systems, data flow processors (DFP's), digital signal processors (DSP's) or the like
DE19654595A1 (en) * 1996-12-20 1998-07-02 Pact Inf Tech Gmbh I0 and memory bus system for DFPs as well as building blocks with two- or multi-dimensional programmable cell structures
US6338106B1 (en) * 1996-12-20 2002-01-08 Pact Gmbh I/O and memory bus system for DFPS and units with two or multi-dimensional programmable cell architectures
DE19654846A1 (en) * 1996-12-27 1998-07-09 Pact Inf Tech Gmbh Process for the independent dynamic reloading of data flow processors (DFPs) as well as modules with two- or multi-dimensional programmable cell structures (FPGAs, DPGAs, etc.)
DE59710317D1 (en) * 1996-12-27 2003-07-24 Pact Inf Tech Gmbh METHOD FOR THE INDEPENDENT DYNAMIC RE-LOADING OF DATA FLOW PROCESSORS (DFPs) AND MODULES WITH TWO OR MORE-DIMENSIONAL PROGRAMMABLE CELL STRUCTURES (FPGAs, DPGAs, or the like)
DE19704728A1 (en) 1997-02-08 1998-08-13 Pact Inf Tech Gmbh Method for self-synchronization of configurable elements of a programmable module
US6542998B1 (en) 1997-02-08 2003-04-01 Pact Gmbh Method of self-synchronization of configurable elements of a programmable module
DE19704742A1 (en) * 1997-02-11 1998-09-24 Pact Inf Tech Gmbh Internal bus system for DFPs, as well as modules with two- or multi-dimensional programmable cell structures, for coping with large amounts of data with high networking effort
US8686549B2 (en) 2001-09-03 2014-04-01 Martin Vorbach Reconfigurable elements
DE19861088A1 (en) * 1997-12-22 2000-02-10 Pact Inf Tech Gmbh Repairing integrated circuits by replacing subassemblies with substitutes
DE19807872A1 (en) 1998-02-25 1999-08-26 Pact Inf Tech Gmbh Method of managing configuration data in data flow processors
DE10081643D2 (en) 1999-06-10 2002-05-29 Pact Inf Tech Gmbh Sequence partitioning on cell structures
US6851047B1 (en) * 1999-10-15 2005-02-01 Xilinx, Inc. Configuration in a configurable system on a chip
JP2004506261A (en) * 2000-06-13 2004-02-26 ペーアーツェーテー　イクスペーペー　テクノロジーズ　アクチエンゲゼルシャフト Pipeline CT protocol and CT communication
US8058899B2 (en) 2000-10-06 2011-11-15 Martin Vorbach Logic cell array and bus system
US20040015899A1 (en) * 2000-10-06 2004-01-22 Frank May Method for processing data
US7581076B2 (en) * 2001-03-05 2009-08-25 Pact Xpp Technologies Ag Methods and devices for treating and/or processing data
US9250908B2 (en) 2001-03-05 2016-02-02 Pact Xpp Technologies Ag Multi-processor bus and cache interconnection system
US20090300262A1 (en) * 2001-03-05 2009-12-03 Martin Vorbach Methods and devices for treating and/or processing data
US7844796B2 (en) * 2001-03-05 2010-11-30 Martin Vorbach Data processing device and method
US9436631B2 (en) * 2001-03-05 2016-09-06 Pact Xpp Technologies Ag Chip including memory element storing higher level memory data on a page by page basis
US9552047B2 (en) 2001-03-05 2017-01-24 Pact Xpp Technologies Ag Multiprocessor having runtime adjustable clock and clock dependent power supply
US9037807B2 (en) * 2001-03-05 2015-05-19 Pact Xpp Technologies Ag Processor arrangement on a chip including data processing, memory, and interface elements
US7210129B2 (en) * 2001-08-16 2007-04-24 Pact Xpp Technologies Ag Method for translating programs for reconfigurable architectures
US20090210653A1 (en) * 2001-03-05 2009-08-20 Pact Xpp Technologies Ag Method and device for treating and processing data
US9411532B2 (en) 2001-09-07 2016-08-09 Pact Xpp Technologies Ag Methods and systems for transferring data between a processing device and external devices
US7444531B2 (en) 2001-03-05 2008-10-28 Pact Xpp Technologies Ag Methods and devices for treating and processing data
US20040133745A1 (en) 2002-10-28 2004-07-08 Quicksilver Technology, Inc. Adaptable datapath for a digital processing system
US7962716B2 (en) * 2001-03-22 2011-06-14 Qst Holdings, Inc. Adaptive integrated circuitry with heterogeneous and reconfigurable matrices of diverse and adaptive computational units having fixed, application specific computational elements
US7752419B1 (en) * 2001-03-22 2010-07-06 Qst Holdings, Llc Method and system for managing hardware resources to implement system functions using an adaptive computing architecture
US7489779B2 (en) * 2001-03-22 2009-02-10 Qstholdings, Llc Hardware implementation of the secure hash standard
US7653710B2 (en) * 2002-06-25 2010-01-26 Qst Holdings, Llc. Hardware task manager
US6836839B2 (en) 2001-03-22 2004-12-28 Quicksilver Technology, Inc. Adaptive integrated circuitry with heterogeneous and reconfigurable matrices of diverse and adaptive computational units having fixed, application specific computational elements
US7400668B2 (en) * 2001-03-22 2008-07-15 Qst Holdings, Llc Method and system for implementing a system acquisition function for use with a communication device
US6577678B2 (en) * 2001-05-08 2003-06-10 Quicksilver Technology Method and system for reconfigurable channel coding
US20020184291A1 (en) * 2001-05-31 2002-12-05 Hogenauer Eugene B. Method and system for scheduling in an adaptable computing engine
US7657877B2 (en) * 2001-06-20 2010-02-02 Pact Xpp Technologies Ag Method for processing data
US10031733B2 (en) 2001-06-20 2018-07-24 Scientia Sol Mentis Ag Method for processing data
US7996827B2 (en) 2001-08-16 2011-08-09 Martin Vorbach Method for the translation of programs for reconfigurable architectures
US7434191B2 (en) * 2001-09-03 2008-10-07 Pact Xpp Technologies Ag Router
US7107374B1 (en) 2001-09-05 2006-09-12 Xilinx, Inc. Method for bus mastering for devices resident in configurable system logic
US8686475B2 (en) 2001-09-19 2014-04-01 Pact Xpp Technologies Ag Reconfigurable elements
US7046635B2 (en) * 2001-11-28 2006-05-16 Quicksilver Technology, Inc. System for authorizing functionality in adaptable hardware devices
US6986021B2 (en) 2001-11-30 2006-01-10 Quick Silver Technology, Inc. Apparatus, method, system and executable module for configuration and operation of adaptive integrated circuitry having fixed, application specific computational elements
US8412915B2 (en) 2001-11-30 2013-04-02 Altera Corporation Apparatus, system and method for configuration of adaptive integrated circuitry having heterogeneous computational elements
US7602740B2 (en) * 2001-12-10 2009-10-13 Qst Holdings, Inc. System for adapting device standards after manufacture
US7088825B2 (en) * 2001-12-12 2006-08-08 Quicksilver Technology, Inc. Low I/O bandwidth method and system for implementing detection and identification of scrambling codes
US7215701B2 (en) * 2001-12-12 2007-05-08 Sharad Sambhwani Low I/O bandwidth method and system for implementing detection and identification of scrambling codes
US7231508B2 (en) * 2001-12-13 2007-06-12 Quicksilver Technologies Configurable finite state machine for operation of microinstruction providing execution enable control value
US7577822B2 (en) * 2001-12-14 2009-08-18 Pact Xpp Technologies Ag Parallel task operation in processor and reconfigurable coprocessor configured based on information in link list including termination information for synchronization
US7403981B2 (en) 2002-01-04 2008-07-22 Quicksilver Technology, Inc. Apparatus and method for adaptive multimedia reception and transmission in communication environments
WO2003071418A2 (en) * 2002-01-18 2003-08-28 Pact Xpp Technologies Ag Method and device for partitioning large computer programs
WO2003060747A2 (en) * 2002-01-19 2003-07-24 Pact Xpp Technologies Ag Reconfigurable processor
EP2043000B1 (en) 2002-02-18 2011-12-21 Richter, Thomas Bus systems and reconfiguration method
US8914590B2 (en) * 2002-08-07 2014-12-16 Pact Xpp Technologies Ag Data processing method and device
US20060075211A1 (en) * 2002-03-21 2006-04-06 Martin Vorbach Method and device for data processing
US9170812B2 (en) 2002-03-21 2015-10-27 Pact Xpp Technologies Ag Data processing system having integrated pipelined array data processor
US7660984B1 (en) 2003-05-13 2010-02-09 Quicksilver Technology Method and system for achieving individualized protected space in an operating system
US7328414B1 (en) * 2003-05-13 2008-02-05 Qst Holdings, Llc Method and system for creating and programming an adaptive computing engine
US20110238948A1 (en) * 2002-08-07 2011-09-29 Martin Vorbach Method and device for coupling a data processing unit and a data processing array
AU2003286131A1 (en) * 2002-08-07 2004-03-19 Pact Xpp Technologies Ag Method and device for processing data
US7657861B2 (en) * 2002-08-07 2010-02-02 Pact Xpp Technologies Ag Method and device for processing data
US8108656B2 (en) 2002-08-29 2012-01-31 Qst Holdings, Llc Task definition for specifying resource requirements
AU2003265891A1 (en) * 2002-09-04 2004-03-29 Mentor Graphics (Holdings) Ltd. Polymorphic computational system and method in signals intelligence analysis
US7394284B2 (en) * 2002-09-06 2008-07-01 Pact Xpp Technologies Ag Reconfigurable sequencer structure
US7937591B1 (en) 2002-10-25 2011-05-03 Qst Holdings, Llc Method and system for providing a device which can be adapted on an ongoing basis
US8276135B2 (en) 2002-11-07 2012-09-25 Qst Holdings Llc Profiling of software and circuit designs utilizing data operation analyses
US7225301B2 (en) * 2002-11-22 2007-05-29 Quicksilver Technologies External memory controller node
JP2006524850A (en) * 2003-04-04 2006-11-02 ペーアーツェーテー　イクスペーペー　テクノロジーズ　アクチエンゲゼルシャフト Data processing method and data processing apparatus
US7609297B2 (en) * 2003-06-25 2009-10-27 Qst Holdings, Inc. Configurable hardware based digital imaging apparatus
US7200837B2 (en) * 2003-08-21 2007-04-03 Qst Holdings, Llc System, method and software for static and dynamic programming and configuration of an adaptive computing architecture
EP1676208A2 (en) 2003-08-28 2006-07-05 PACT XPP Technologies AG Data processing device and method
JP4174402B2 (en) * 2003-09-26 2008-10-29 株式会社東芝 Control circuit and reconfigurable logic block
US7882165B2 (en) * 2003-12-29 2011-02-01 Xilinx, Inc. Digital signal processing element having an arithmetic logic unit
US7853634B2 (en) * 2003-12-29 2010-12-14 Xilinx, Inc. Digital signal processing circuit having a SIMD circuit
US8495122B2 (en) * 2003-12-29 2013-07-23 Xilinx, Inc. Programmable device with dynamic DSP architecture
US7860915B2 (en) * 2003-12-29 2010-12-28 Xilinx, Inc. Digital signal processing circuit having a pattern circuit for determining termination conditions
US7853632B2 (en) * 2003-12-29 2010-12-14 Xilinx, Inc. Architectural floorplan for a digital signal processing circuit
US7567997B2 (en) * 2003-12-29 2009-07-28 Xilinx, Inc. Applications of cascading DSP slices
US7870182B2 (en) * 2003-12-29 2011-01-11 Xilinx Inc. Digital signal processing circuit having an adder circuit with carry-outs
US7853636B2 (en) 2003-12-29 2010-12-14 Xilinx, Inc. Digital signal processing circuit having a pattern detector circuit for convergent rounding
US7480690B2 (en) * 2003-12-29 2009-01-20 Xilinx, Inc. Arithmetic circuit with multiplexed addend inputs
US7472155B2 (en) * 2003-12-29 2008-12-30 Xilinx, Inc. Programmable logic device with cascading DSP slices
US7840630B2 (en) * 2003-12-29 2010-11-23 Xilinx, Inc. Arithmetic logic unit circuit
US7840627B2 (en) 2003-12-29 2010-11-23 Xilinx, Inc. Digital signal processing circuit having input register blocks
US7467175B2 (en) * 2003-12-29 2008-12-16 Xilinx, Inc. Programmable logic device with pipelined DSP slices
US7844653B2 (en) * 2003-12-29 2010-11-30 Xilinx, Inc. Digital signal processing circuit having a pre-adder circuit
US7849119B2 (en) * 2003-12-29 2010-12-07 Xilinx, Inc. Digital signal processing circuit having a pattern detector circuit
US7865542B2 (en) * 2003-12-29 2011-01-04 Xilinx, Inc. Digital signal processing block having a wide multiplexer
JP2008530642A (en) * 2005-02-07 2008-08-07 ペーアーツェーテー イクスペーペー テクノロジーズ アクチエンゲゼルシャフト Low latency mass parallel data processor
WO2007062327A2 (en) * 2005-11-18 2007-05-31 Ideal Industries, Inc. Releasable wire connector
US8250503B2 (en) 2006-01-18 2012-08-21 Martin Vorbach Hardware definition method including determining whether to implement a function as hardware or software
EP2159799A1 (en) * 2008-08-27 2010-03-03 Panasonic Corporation Semiconductor memory with shared global busses for reconfigurable logic device
US8479133B2 (en) * 2009-01-27 2013-07-02 Xilinx, Inc. Method of and circuit for implementing a filter in an integrated circuit
US8543635B2 (en) * 2009-01-27 2013-09-24 Xilinx, Inc. Digital signal processing block with preadder stage
KR101814221B1 (en) 2010-01-21 2018-01-02 스비랄 인크 A method and apparatus for a general-purpose, multiple-core system for implementing stream-based computations
CN103246542B (en) * 2012-02-01 2017-11-14 中兴通讯股份有限公司 Intelligent buffer and intelligent terminal
CN105302577B (en) * 2015-11-26 2019-05-07 上海兆芯集成电路有限公司 Drive the machine code generation method and device of execution unit
CN114003547B (en) 2017-03-14 2023-12-19 珠海市芯动力科技有限公司 Reconfigurable parallel processing
JP6911600B2 (en) * 2017-07-18 2021-07-28 富士通株式会社 Information processing equipment, information processing methods and information processing programs
WO2019028327A1 (en) 2017-08-03 2019-02-07 Next Silicon, Ltd. Reconfigurable cache architecture and methods for cache coherency
EP3662384A4 (en) 2017-08-03 2021-05-05 Next Silicon Ltd Runtime optimization of configurable hardware
US20190197018A1 (en) * 2017-08-19 2019-06-27 Wave Computing, Inc. Dynamic reconfiguration using data transfer control
WO2019055675A1 (en) 2017-09-13 2019-03-21 Next Silicon, Ltd. Directed and interconnected grid dataflow architecture
JP7419764B2 (en) * 2019-11-20 2024-01-23 株式会社リコー Information processing device and configuration method
US11269526B2 (en) 2020-04-23 2022-03-08 Next Silicon Ltd Interconnected memory grid with bypassable units

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5760607A (en) * 1995-07-10 1998-06-02 Xilinx, Inc. System comprising field programmable gate array and intelligent memory

Family Cites Families (37)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3631404A (en) * 1969-07-28 1971-12-28 Gen Electric Data communication system including address-generating means and method
JPS5279630A (en) * 1975-12-25 1977-07-04 Toshiba Corp Data processing unit
DE3121444A1 (en) * 1981-05-29 1982-12-16 Siemens AG, 1000 Berlin und 8000 München METHOD AND ARRANGEMENT FOR DEMODULATING FSK SIGNALS
US4507730A (en) * 1981-10-01 1985-03-26 Honeywell Information Systems Inc. Memory system with automatic memory configuration
US4823286A (en) * 1987-02-12 1989-04-18 International Business Machines Corporation Pixel data path for high performance raster displays with all-point-addressable frame buffers
KR910009555B1 (en) * 1989-01-09 1991-11-21 조경연 Semiconductor memory device
US5210749A (en) 1990-05-29 1993-05-11 Advanced Micro Devices, Inc. Configuration of srams as logical fifos for transmit and receive of packet data
US5261077A (en) * 1990-06-29 1993-11-09 Digital Equipment Corporation Configurable data path arrangement for resolving data type incompatibility
JP3114237B2 (en) 1991-04-30 2000-12-04 日本電気株式会社 Semiconductor storage device
US5392292A (en) * 1991-06-27 1995-02-21 Cray Research, Inc. Configurable spare memory chips
US5298805A (en) * 1991-08-29 1994-03-29 National Semiconductor Corporation Versatile and efficient cell-to-local bus interface in a configurable logic array
JP3594626B2 (en) 1993-03-04 2004-12-02 株式会社ルネサステクノロジ Non-volatile memory device
US5506814A (en) 1993-05-28 1996-04-09 Micron Technology, Inc. Video random access memory device and method implementing independent two WE nibble control
US5623620A (en) 1993-06-30 1997-04-22 Intel Corporation Special test modes for a page buffer shared resource in a memory device
GB9403030D0 (en) * 1994-02-17 1994-04-06 Austin Kenneth Re-configurable application specific device
US5560030A (en) 1994-03-08 1996-09-24 Texas Instruments Incorporated Transfer processor with transparency
US6351780B1 (en) 1994-11-21 2002-02-26 Cirrus Logic, Inc. Network controller using held data frame monitor and decision logic for automatically engaging DMA data transfer when buffer overflow is anticipated
JPH08314794A (en) * 1995-02-28 1996-11-29 Matsushita Electric Ind Co Ltd Method and system for shortening wait time of access to stable storage device
US5646544A (en) * 1995-06-05 1997-07-08 International Business Machines Corporation System and method for dynamically reconfiguring a programmable gate array
US5692147A (en) * 1995-06-07 1997-11-25 International Business Machines Corporation Memory mapping method and apparatus to fold sparsely populated structures into densely populated memory columns or rows by selectively transposing X and Y address portions, and programmable gate array applications thereof
US5778439A (en) * 1995-08-18 1998-07-07 Xilinx, Inc. Programmable logic device with hierarchical confiquration and state storage
US5914906A (en) * 1995-12-20 1999-06-22 International Business Machines Corporation Field programmable memory array
US5737766A (en) * 1996-02-14 1998-04-07 Hewlett Packard Company Programmable gate array configuration memory which allows sharing with user memory
US6085302A (en) * 1996-04-17 2000-07-04 Advanced Micro Devices, Inc. Microprocessor having address generation units for efficient generation of memory operation addresses
US5784636A (en) * 1996-05-28 1998-07-21 National Semiconductor Corporation Reconfigurable computer architecture for use in signal processing applications
US5793992A (en) * 1996-06-13 1998-08-11 Vlsi Technology, Inc. Method and apparatus for arbitrating access to main memory of a computer system
US5838165A (en) * 1996-08-21 1998-11-17 Chatter; Mukesh High performance self modifying on-the-fly alterable logic FPGA, architecture and method
US5933023A (en) * 1996-09-03 1999-08-03 Xilinx, Inc. FPGA architecture having RAM blocks with programmable word length and width and dedicated address and data lines
US6034547A (en) * 1996-09-04 2000-03-07 Advantage Logic, Inc. Method and apparatus for universal program controlled bus
DE19654846A1 (en) * 1996-12-27 1998-07-09 Pact Inf Tech Gmbh Process for the independent dynamic reloading of data flow processors (DFPs) as well as modules with two- or multi-dimensional programmable cell structures (FPGAs, DPGAs, etc.)
US5933855A (en) * 1997-03-21 1999-08-03 Rubinstein; Richard Shared, reconfigurable memory architectures for digital signal processing
US6011407A (en) * 1997-06-13 2000-01-04 Xilinx, Inc. Field programmable gate array with dedicated computer bus interface and method for configuring both
US6044416A (en) * 1997-09-19 2000-03-28 Samsung Electronics Co., Ltd. Configurable first-in first-out memory interface
KR100269313B1 (en) 1997-11-07 2000-12-01 윤종용 Semiconductor memory device for consuming small current at stand-by state
US6034538A (en) * 1998-01-21 2000-03-07 Lucent Technologies Inc. Virtual logic system for reconfigurable hardware
US6049487A (en) * 1998-03-16 2000-04-11 Actel Corporation Embedded static random access memory for field programmable gate array
US6385751B1 (en) * 1998-12-30 2002-05-07 Texas Instruments Incorporated Programmable, reconfigurable DSP implementation of a Reed-Solomon encoder/decoder

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5760607A (en) * 1995-07-10 1998-06-02 Xilinx, Inc. System comprising field programmable gate array and intelligent memory

Also Published As

Publication number Publication date
WO2001001575A9 (en) 2002-07-25
US6347346B1 (en) 2002-02-12
AU5899100A (en) 2001-01-31
EP1198888A4 (en) 2002-10-02
EP1198888A1 (en) 2002-04-24
US20020038414A1 (en) 2002-03-28

Similar Documents

Publication Publication Date Title
US6347346B1 (en) Local memory unit system with global access for use on reconfigurable chips
US5740402A (en) Conflict resolution in interleaved memory systems with multiple parallel accesses
EP1222559B1 (en) Multiprocessor node controller circuit and method
US3537074A (en) Parallel operating array computer
US10210914B2 (en) Programmable logic accelerator in system on chip
US5101498A (en) Pin selectable multi-mode processor
JP3853736B2 (en) User configurable on-chip memory system
US6653859B2 (en) Heterogeneous integrated circuit with reconfigurable logic cores
US7930462B2 (en) Interface controller that has flexible configurability and low cost
JPH1131122A (en) Bus mutual connection system
JPS6137662B1 (en)
JPS6327739B2 (en)
JPS6035696B2 (en) Bus control device in data processing equipment
JPH05181816A (en) Parallel data processor and microprocessor
AU2009209933B2 (en) Reconfigurable device
US8284792B2 (en) Buffer minimization in interface controller
US8352695B2 (en) Selectable access rates in a memory and memory communication system
CN114254577A (en) Logic structure based on micro-sector infrastructure with data registers having scan registers
JPH04260959A (en) Microprocessor
US6571301B1 (en) Multi processor system and FIFO circuit
CN114662432A (en) Micro-network on chip and micro-sector infrastructure
EP0953916A1 (en) Microcomputer chips with interconnected address and data paths
EP0953915A1 (en) Microcomputer with packet translation for event packets and memory access packets
US5555433A (en) Circuit for interfacing data busses
US5862408A (en) Microprocessor system having multiplexor disposed in first and second read paths between memory CPU and DMA for selecting data from either read path

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
WWE Wipo information: entry into national phase

Ref document number: 2000944986

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 2000944986

Country of ref document: EP

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

AK Designated states

Kind code of ref document: C2

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: C2

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

COP Corrected version of pamphlet

Free format text: PAGES 1/15-15/15, DRAWINGS, REPLACED BY NEW PAGES 1/13-13/13; DUE TO LATE TRANSMITTAL BY THE RECEIVING OFFICE

NENP Non-entry into the national phase

Ref country code: JP

WWW Wipo information: withdrawn in national office

Ref document number: 2000944986

Country of ref document: EP