WO1996022571A1 - Bus bridge circuit and method using snoop ahead operations - Google Patents

Bus bridge circuit and method using snoop ahead operations Download PDF

Info

Publication number
WO1996022571A1
WO1996022571A1 PCT/US1996/000506 US9600506W WO9622571A1 WO 1996022571 A1 WO1996022571 A1 WO 1996022571A1 US 9600506 W US9600506 W US 9600506W WO 9622571 A1 WO9622571 A1 WO 9622571A1
Authority
WO
WIPO (PCT)
Prior art keywords
bus
peripheral
over
memory
write data
Prior art date
Application number
PCT/US1996/000506
Other languages
French (fr)
Inventor
George Hayek
Ali S. Oztaskin
Brian Langendorf
Bruce Young
Original Assignee
Intel Corporation
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
Family has litigation
First worldwide family litigation filed litigation Critical https://patents.darts-ip.com/?family=23483143&utm_source=google_patent&utm_medium=platform_link&utm_campaign=public_patent_search&patent=WO1996022571(A1) "Global patent litigation dataset” by Darts-ip is licensed under a Creative Commons Attribution 4.0 International License.
Application filed by Intel Corporation filed Critical Intel Corporation
Priority to AU46993/96A priority Critical patent/AU4699396A/en
Priority to DE69604564T priority patent/DE69604564T2/en
Priority to EP96902679A priority patent/EP0804763B1/en
Publication of WO1996022571A1 publication Critical patent/WO1996022571A1/en
Priority to HK98103810A priority patent/HK1004432A1/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F12/00—Accessing, addressing or allocating within memory systems or architectures
    • G06F12/02—Addressing or allocation; Relocation
    • G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
    • G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
    • G06F12/0806—Multiuser, multiprocessor or multiprocessing cache systems
    • G06F12/0815—Cache consistency protocols
    • G06F12/0831—Cache consistency protocols using a bus scheme, e.g. with bus monitoring or watching means
    • G06F12/0835—Cache consistency protocols using a bus scheme, e.g. with bus monitoring or watching means for main memory peripheral accesses (e.g. I/O or DMA)
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/14—Handling requests for interconnection or transfer
    • G06F13/16—Handling requests for interconnection or transfer for access to memory bus
    • G06F13/1668—Details of memory controller
    • G06F13/1684—Details of memory controller using multiple buses
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00—Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/38—Information transfer, e.g. on bus
    • G06F13/40—Bus structure
    • G06F13/4004—Coupling between buses
    • G06F13/4027—Coupling between buses using bus bridges
    • G06F13/405—Coupling between buses using bus bridges where the bridge performs a synchronising function
    • G06F13/4059—Coupling between buses using bus bridges where the bridge performs a synchronising function where the synchronisation uses buffers, e.g. for speed matching between buses

Definitions

  • the present invention pertains to the field of computer systems. More particularly, this invention relates to an integrated bus bridge and memory controller that enables data streaming to a shared memory resource while maintaining cache coherency.
  • Prior computer systems commonly include a main processor or host processor that communicates with other elements of the computer system via a processor or host bus.
  • Other elements coupled to such a processor or host bus typically include one or more cache memories as well as a main memory for the computer system.
  • Such a main memory is typically comprised of dynamic random access memories (DRAM).
  • DRAM dynamic random access memories
  • Prior computer systems may also include one or more peripheral buses that enable communication among a variety of peripheral components.
  • Such a computer system may also include bridge circuits that enable communication between the processor or host bus and the peripheral buses.
  • Such a bridge circuit typically enables communication between the main or host processor and the various bus agents coupled to the peripheral bus.
  • bus agents coupled to such a peripheral bus may require access to the main memory of the computer system.
  • the data stored in the main memory may be modified either by the main processor or host processor or other elements coupled to the processor bus as well as various bus agents coupled to the peripheral buses.
  • peripheral buses usually enable extremely high speed or high bandwidth communication.
  • PCI peripheral component interconnect
  • the peripheral component interconnect (PCI) published bus standard allows data communication bandwidth of up to 133 megabytes per second.
  • PCI peripheral component interconnect
  • prior computer systems typically limit the available bandwidth on such a peripheral bus well below the maximum allowable provided by the published standard.
  • the bandwidth on such a peripheral bus is limited by the cache coherency transactions that occur on the host or processor bus during memory access transactions that initiate from the peripheral bus.
  • one prior computer system employs an integrated bridge and memory controller circuit that functions as a bus bridge between the host or processor bus and a peripheral bus while at the same time functioning as a memory controller for the main memory of the computer system.
  • Such a system typically limits each atomic data transfer transaction on the peripheral bus targeted for the main memory to a single data line or cache line.
  • a prior integrated bridge and memory controller usually receives a data line over the peripheral bus and generates a corresponding snoop request over the processor bus. After the appropriate invalidate line or write back transaction on the host or processor bus, the integrated bridge and memory controller then transfers the data line to the main memory.
  • a bridge and memory controller usually cannot immediately accept a subsequent data line over the peripheral bus during a data streaming transaction. Instead, such an integrated bridge and memory controller usually signals an end to the write transaction on the peripheral bus. Thereafter, the initiating bus agent on the peripheral bus typically initiates another write transaction to transfer the next data line to the main memory.
  • Such a system that limits atomic transfers to a single data line ensures that the integrated bridge and memory controller obtains a data line address from the originating bus agent on the peripheral bus for each data line transfer to the main memory.
  • the integrated bridge and memory controller typically uses the address for each data line to initiate a snoop request over the processor or host bus.
  • a memory controller could generate addresses for each data line transferred over the peripheral bus in order to provide the proper snoop request transactions on the processor bus.
  • address generation logic typically increases the complexity and cost of such a memory controller and drives up the overall system cost.
  • An integrated bus bridge and memory controller circuit that enables access to a shared memory with high bandwidth data streaming.
  • the circuit enables access to the shared memory over a processor bus and over a high bandwidth peripheral bus.
  • the circuit performs a series of snoop ahead transactions over the processor bus during data streaming transactions to the shared memory that originate over the peripheral bus.
  • Figure 1 illustrates a computer system for one embodiment which comprises a processor, a cache memory, a main memory, and a bridge and memory controller circuit;
  • Figure 2 illustrates the bridge and memory controller for one embodiment that includes a write buffer, a peripheral read prefetch buffer, and a peripheral write buffer along with a state machine that controls snoop ahead transactions;
  • Figure 3 illustrates a write transaction targeted for the main memory that originates over the peripheral bus wherein the write transaction comprises an address phase followed by a series of data phases of line 0 data - line n data;
  • Figure 4 illustrates the format of data stored in the peripheral write buffer for one embodiment including a set of double word data values each with a corresponding snoop status flag;
  • Figure 5 illustrates a read transaction from the main memory over the peripheral bus wherein the read transaction comprises an address phase, and a set of wait states, followed by a set of data phases comprising line 0 data - line n data.
  • FIG. 1 illustrates a computer system 10 for one embodiment.
  • the computer system 10 comprises a processor 12, a cache memory 14, a main memory 18, and a bridge and memory controller 16.
  • the processor 12 contains an internal level one (L1) cache memory (not shown).
  • the processor 12 coupled for communication over a processor bus 24.
  • the cache memory 14 is also coupled to the processor bus 24 and functions as a level two (L2) cache memory for the processor 12.
  • the main memory 18 is comprised of dynamic random access memory (DRAM) and functions as the main memory storage for the computer system 10.
  • DRAM dynamic random access memory
  • the bridge and memory controller 16 functions as a memory controller for the main memory 18. In addition, the bridge and memory controller 16 functions as a cache controller for the cache memory 14. The bridge and memory controller 16 enables read and write access to the main memory 18 from the processor bus 24 as well as a peripheral bus 26.
  • the bridge and memory controller 16 also functions as a bus bridge between the processor bus 24 and the peripheral bus 26.
  • the bridge and memory controller 16 enables transactions originating on the processor bus 24 to propagate to the peripheral bus 26.
  • the bridge and memory controller 16 enables transactions originating on the peripheral bus 26 from one of a set of bus agents 20 and 22 to propagate to the processor bus 24.
  • the bridge and memory controller 16 also enables the bus agents 20 and 22 to access data stored in the main memory 18.
  • the bridge and memory controller 16 coordinates accesses to the main memory 18 that originate on the peripheral bus 26 with accesses to the main memory 18 that originate on the processor bus 24.
  • the bridge and memory controller 16 initiates snoop request transactions on the processor bus 24 during memory accesses to the main memory 18 from the peripheral bus 26 in order to maintain cache coherency among the caches coupled to the processor bus 24.
  • the caches coupled to the processor bus 24 include the internal L1 cache in the processor 12 along with the L2 cache memory 14.
  • FIG. 2 illustrates the bridge and memory controller 16 for one embodiment.
  • the bridge and memory controller 16 includes a write buffer 42, a peripheral read prefetch buffer 44, and a peripheral write buffer 46.
  • the bridge and memory controller 16 includes a state machine 40 that coordinates data transfers between the main memory 18 and the bus agents 20 and 22 via the peripheral bus 26, and that initiates snoop request transactions to the cache memories on the processor bus 24.
  • the write buffer 42 buffers write data targeted for the main memory 18 that originates from agents coupled to the processor bus 24 and agents coupled to the peripheral bus 26.
  • a multiplexer 48 controlled by the state machine 40 selects either a data path from the processor bus 24 or a data path from the peripheral bus 26.
  • the write buffer 42 compensates for the latency of DRAM accesses via a DRAM write path 52 in comparison to the speed of data transfer via the processor bus 24 or the peripheral bus 26.
  • the peripheral write buffer 46 buffers write data targeted for the main memory 18 that originates from bus agents on the peripheral bus 26.
  • the peripheral write buffer 46 stores write data along with flags that are maintained by the state machine 40.
  • the flags maintained by the state machine 40 indicate whether the data stored in the peripheral write buffer 46 corresponds to data lines having a completed snoop request transaction over the processor bus 24.
  • the state machine 40 merges the data from the peripheral write buffer 46 into the write buffer 42 through the multiplexer 48 only after the corresponding snoop request transaction for the data line is complete over the processor bus 24.
  • the peripheral read prefetch buffer 44 prefetches sequential data lines from the main memory 18 in preparation for read transactions on the peripheral bus 26.
  • the state machine 40 controls prefetches into the peripheral read prefetch buffer 44 to ensure that a read to the main memory 18 via a read path 54 does not occur until after the snoop transaction for the corresponding data line as completed via the processor bus 24.
  • Figure 3 illustrates a write transaction targeted for the main memory 18 that originates over the peripheral bus 26. As shown, the write transaction comprises an address phase followed by a series of data phases comprising line 0 data - line n data. During the address phase, the peripheral bus 26 carries the address of data line 0 for the write transaction. Subsequently, the phases for the line 0 data - line n data each transfer an amount of data via the peripheral bus 26 equal to a cache line in the cache memory 14 or the internal cache of the processor 12.
  • the state machine 40 performs a snoop ahead transaction for each data line in the write transaction. For example, during the address phase of the write transaction on the peripheral bus 26, the state machine 40 issues a line 0 snoop request over the processor bus 24. Thereafter, the cache memory 14 and the internal cache of the processor 12 each perform either a writeback operation, also known as a copyback transaction, or an internal invalidate cache line operation depending upon the clean or dirty status of the cache line specified by the line 0 snoop request.
  • a writeback operation also known as a copyback transaction
  • an internal invalidate cache line operation depending upon the clean or dirty status of the cache line specified by the line 0 snoop request.
  • the line 0 snoop transaction on the processor bus 24 completes.
  • the line 0 data buffered in the peripheral write buffer 46 is marked by the state machine 40 with a snoop done (SD) flag.
  • SD snoop done
  • the state machine 40 stores a snoop not done (SND) flag for the line 0 data contained in the peripheral write buffer 46 prior to the completion of the line 0 snoop request on the processor bus 24.
  • the state machine 40 Later during the line 0 data phase, the state machine 40 generates a next sequential address corresponding to the line 1 data. Thereafter, the state machine 40 issues a line 1 snoop request over the processor bus 24. As shown, the state machine 40 continuously issues snoop request transactions over the processor bus 24 prior to the reception of the corresponding data line over the peripheral bus 26 during the write transaction. The snoop ahead transactions issues by the state machine 40 prevent the occurrence of wait cycles or delay cycles on the peripheral bus 26 during write transactions targeted for the main memory 18 through the peripheral write buffer 46.
  • FIG. 4 illustrates the format of data stored in the peripheral write buffer 46 for one embodiment.
  • the peripheral write buffer 46 stores a set of double word data values each with a corresponding snoop status flag.
  • the snoop status flag indicates either snoop done (SD) or snoop not done (SND) depending upon the status of the corresponding snoop ahead request issued by the state machine 40 over the processor bus 24.
  • SD snoop done
  • SND snoop not done
  • the boundary in the peripheral write buffer 46 between data having an SD indication and data having an SND indication provides a snoop boundary between line n-1 and line n.
  • the state machine 40 merges the data from the peripheral write buffer 46 into the write buffer 42 only if the corresponding snoop status flag indicates snoop done.
  • Figure 5 illustrates a read transaction from the main memory 18 over the peripheral bus 26.
  • the read transaction on the peripheral bus 26 comprises an address phase, and a set of wait states, followed by a set of data phases comprising line 0 data - line n data.
  • the originating bus agent on the peripheral bus 26 transfers the address of the initial data line being read from the main memory 18.
  • the state machine 40 issues a line 0 snoop request over the processor bus 24 that specifies the data line address provided during the address phase.
  • the line 0 snoop transaction on the processor bus 24 completes. Thereafter, a DRAM latency occurs while the peripheral read prefetch buffer 44 receives the corresponding data line from the main memory 18 via the read path 54. The line 0 prefetch to the peripheral read prefetch buffer 44 completes before the start of the line 0 data phase on the peripheral bus 26.
  • the peripheral read prefetch buffer 44 transfers the buffered read data over the peripheral bus 26 while the state machine 40 issues a line 1 snoop request over the processor bus 24 using an address generated as a next sequential cache line address from the address specified during the preceding address phase. Later in the line 0 data phase of the read transaction on the peripheral bus 26, the line 1 snoop transaction on the processor bus 24 completes. The completion of the line 1 snoop request enables a prefetch of the corresponding data line from the main memory 18 into the peripheral read prefetch buffer 44. Thereafter, the line 1 data from the peripheral read prefetch buffer 44 is transferred over the peripheral bus 26.
  • the state machine 40 monitors the occurrence of wait states on the peripheral bus 26 during read and write transactions.
  • the state machine 40 counts the number of wait states that occur on the peripheral bus 26.
  • the state machine 40 implements a disconnect time ⁇ out mechanism such that a disconnect is signaled over the peripheral bus 26 if the number of consecutive wait states on the peripheral bus 26 exceeds eight.
  • the eight wait states conforms to the peripheral component interface (PCI) specification for maximum latency during data transfers.
  • PCI peripheral component interface

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Memory System Of A Hierarchy Structure (AREA)

Abstract

An integrated bus bridge and memory controller circuit (16) enables access to a shared memory (18) with high bandwidth data streaming. The bus bridge and memory controller circuit (16) performs a series of snoop ahead transactions over the processor bus (24) during access transactions to the shared memory (18) that originate over the peripheral bus (26) and thereby enables high bandwidth data streaming on the peripheral bus (26). The bus bridge and memory controller circuit (16) may be used in a computer system (10) having a processor (12) and a cache memory (14) coupled to the processor bus (24), and a set of bus agents (20, 22) coupled to the peripheral bus (26). The circuit (16) may include a write buffer (42) for merging write data received over the buses, a peripheral write buffer (46) that stores a snoop done flag, and a peripheral read prefetch buffer (44).

Description

BUS BRIDGE CIRCUIT AND METHOD USING SNOOP AHEAD OPERATIONS
FIELD OF THE INVENTION
The present invention pertains to the field of computer systems. More particularly, this invention relates to an integrated bus bridge and memory controller that enables data streaming to a shared memory resource while maintaining cache coherency.
BACKGROUND
Prior computer systems commonly include a main processor or host processor that communicates with other elements of the computer system via a processor or host bus. Other elements coupled to such a processor or host bus typically include one or more cache memories as well as a main memory for the computer system. Such a main memory is typically comprised of dynamic random access memories (DRAM).
Prior computer systems may also include one or more peripheral buses that enable communication among a variety of peripheral components. Such a computer system may also include bridge circuits that enable communication between the processor or host bus and the peripheral buses. Such a bridge circuit typically enables communication between the main or host processor and the various bus agents coupled to the peripheral bus.
In addition, the bus agents coupled to such a peripheral bus may require access to the main memory of the computer system. In such a system, the data stored in the main memory may be modified either by the main processor or host processor or other elements coupled to the processor bus as well as various bus agents coupled to the peripheral buses.
In some prior computer systems, such peripheral buses usually enable extremely high speed or high bandwidth communication. For example, the peripheral component interconnect (PCI) published bus standard allows data communication bandwidth of up to 133 megabytes per second. Unfortunately, prior computer systems typically limit the available bandwidth on such a peripheral bus well below the maximum allowable provided by the published standard.
Typically, the bandwidth on such a peripheral bus is limited by the cache coherency transactions that occur on the host or processor bus during memory access transactions that initiate from the peripheral bus. For example, one prior computer system employs an integrated bridge and memory controller circuit that functions as a bus bridge between the host or processor bus and a peripheral bus while at the same time functioning as a memory controller for the main memory of the computer system. Such a system typically limits each atomic data transfer transaction on the peripheral bus targeted for the main memory to a single data line or cache line.
For example, during a memory write transaction in such a system that originates on the peripheral bus, a prior integrated bridge and memory controller usually receives a data line over the peripheral bus and generates a corresponding snoop request over the processor bus. After the appropriate invalidate line or write back transaction on the host or processor bus, the integrated bridge and memory controller then transfers the data line to the main memory. Unfortunately, such a bridge and memory controller usually cannot immediately accept a subsequent data line over the peripheral bus during a data streaming transaction. Instead, such an integrated bridge and memory controller usually signals an end to the write transaction on the peripheral bus. Thereafter, the initiating bus agent on the peripheral bus typically initiates another write transaction to transfer the next data line to the main memory.
Such a system that limits atomic transfers to a single data line ensures that the integrated bridge and memory controller obtains a data line address from the originating bus agent on the peripheral bus for each data line transfer to the main memory. The integrated bridge and memory controller typically uses the address for each data line to initiate a snoop request over the processor or host bus. A memory controller could generate addresses for each data line transferred over the peripheral bus in order to provide the proper snoop request transactions on the processor bus. However, such address generation logic typically increases the complexity and cost of such a memory controller and drives up the overall system cost.
SUMMARY QF THE INVENTION
An integrated bus bridge and memory controller circuit is disclosed that enables access to a shared memory with high bandwidth data streaming. The circuit enables access to the shared memory over a processor bus and over a high bandwidth peripheral bus. The circuit performs a series of snoop ahead transactions over the processor bus during data streaming transactions to the shared memory that originate over the peripheral bus.
Other features and advantages of the present invention will be apparent from the accompanying drawings, and from the detailed description that follows below. BRIEF DESCRIPTION QF THE PRAWINGS
The present invention is illustrated by way of example and not limitation in the figures of the accompanying drawings in which like references indicate similar elements, and in which: Figure 1 illustrates a computer system for one embodiment which comprises a processor, a cache memory, a main memory, and a bridge and memory controller circuit;
Figure 2 illustrates the bridge and memory controller for one embodiment that includes a write buffer, a peripheral read prefetch buffer, and a peripheral write buffer along with a state machine that controls snoop ahead transactions;
Figure 3 illustrates a write transaction targeted for the main memory that originates over the peripheral bus wherein the write transaction comprises an address phase followed by a series of data phases of line 0 data - line n data;
Figure 4 illustrates the format of data stored in the peripheral write buffer for one embodiment including a set of double word data values each with a corresponding snoop status flag;
Figure 5 illustrates a read transaction from the main memory over the peripheral bus wherein the read transaction comprises an address phase, and a set of wait states, followed by a set of data phases comprising line 0 data - line n data.
DETAILED DESCRIPTION Figure 1 illustrates a computer system 10 for one embodiment. The computer system 10 comprises a processor 12, a cache memory 14, a main memory 18, and a bridge and memory controller 16. In addition, the processor 12 contains an internal level one (L1) cache memory (not shown).
The processor 12 coupled for communication over a processor bus 24. The cache memory 14 is also coupled to the processor bus 24 and functions as a level two (L2) cache memory for the processor 12. The main memory 18 is comprised of dynamic random access memory (DRAM) and functions as the main memory storage for the computer system 10.
The bridge and memory controller 16 functions as a memory controller for the main memory 18. In addition, the bridge and memory controller 16 functions as a cache controller for the cache memory 14. The bridge and memory controller 16 enables read and write access to the main memory 18 from the processor bus 24 as well as a peripheral bus 26.
The bridge and memory controller 16 also functions as a bus bridge between the processor bus 24 and the peripheral bus 26. The bridge and memory controller 16 enables transactions originating on the processor bus 24 to propagate to the peripheral bus 26. In addition, the bridge and memory controller 16 enables transactions originating on the peripheral bus 26 from one of a set of bus agents 20 and 22 to propagate to the processor bus 24.
The bridge and memory controller 16 also enables the bus agents 20 and 22 to access data stored in the main memory 18. The bridge and memory controller 16 coordinates accesses to the main memory 18 that originate on the peripheral bus 26 with accesses to the main memory 18 that originate on the processor bus 24. In addition, the bridge and memory controller 16 initiates snoop request transactions on the processor bus 24 during memory accesses to the main memory 18 from the peripheral bus 26 in order to maintain cache coherency among the caches coupled to the processor bus 24. The caches coupled to the processor bus 24 include the internal L1 cache in the processor 12 along with the L2 cache memory 14.
Figure 2 illustrates the bridge and memory controller 16 for one embodiment. The bridge and memory controller 16 includes a write buffer 42, a peripheral read prefetch buffer 44, and a peripheral write buffer 46. In addition, the bridge and memory controller 16 includes a state machine 40 that coordinates data transfers between the main memory 18 and the bus agents 20 and 22 via the peripheral bus 26, and that initiates snoop request transactions to the cache memories on the processor bus 24.
The write buffer 42 buffers write data targeted for the main memory 18 that originates from agents coupled to the processor bus 24 and agents coupled to the peripheral bus 26. A multiplexer 48 controlled by the state machine 40 selects either a data path from the processor bus 24 or a data path from the peripheral bus 26. The write buffer 42 compensates for the latency of DRAM accesses via a DRAM write path 52 in comparison to the speed of data transfer via the processor bus 24 or the peripheral bus 26.
The peripheral write buffer 46 buffers write data targeted for the main memory 18 that originates from bus agents on the peripheral bus 26. The peripheral write buffer 46 stores write data along with flags that are maintained by the state machine 40. The flags maintained by the state machine 40 indicate whether the data stored in the peripheral write buffer 46 corresponds to data lines having a completed snoop request transaction over the processor bus 24. The state machine 40 merges the data from the peripheral write buffer 46 into the write buffer 42 through the multiplexer 48 only after the corresponding snoop request transaction for the data line is complete over the processor bus 24.
The peripheral read prefetch buffer 44 prefetches sequential data lines from the main memory 18 in preparation for read transactions on the peripheral bus 26. The state machine 40 controls prefetches into the peripheral read prefetch buffer 44 to ensure that a read to the main memory 18 via a read path 54 does not occur until after the snoop transaction for the corresponding data line as completed via the processor bus 24. Figure 3 illustrates a write transaction targeted for the main memory 18 that originates over the peripheral bus 26. As shown, the write transaction comprises an address phase followed by a series of data phases comprising line 0 data - line n data. During the address phase, the peripheral bus 26 carries the address of data line 0 for the write transaction. Subsequently, the phases for the line 0 data - line n data each transfer an amount of data via the peripheral bus 26 equal to a cache line in the cache memory 14 or the internal cache of the processor 12.
To enable a high bandwidth write transactions to the main memory 18 from the peripheral bus 26, the state machine 40 performs a snoop ahead transaction for each data line in the write transaction. For example, during the address phase of the write transaction on the peripheral bus 26, the state machine 40 issues a line 0 snoop request over the processor bus 24. Thereafter, the cache memory 14 and the internal cache of the processor 12 each perform either a writeback operation, also known as a copyback transaction, or an internal invalidate cache line operation depending upon the clean or dirty status of the cache line specified by the line 0 snoop request.
Thereafter, during the line 0 data phase of the write transaction on the peripheral bus 26, the line 0 snoop transaction on the processor bus 24 completes. As a consequence, the line 0 data buffered in the peripheral write buffer 46 is marked by the state machine 40 with a snoop done (SD) flag. Initially, the state machine 40 stores a snoop not done (SND) flag for the line 0 data contained in the peripheral write buffer 46 prior to the completion of the line 0 snoop request on the processor bus 24.
Later during the line 0 data phase, the state machine 40 generates a next sequential address corresponding to the line 1 data. Thereafter, the state machine 40 issues a line 1 snoop request over the processor bus 24. As shown, the state machine 40 continuously issues snoop request transactions over the processor bus 24 prior to the reception of the corresponding data line over the peripheral bus 26 during the write transaction. The snoop ahead transactions issues by the state machine 40 prevent the occurrence of wait cycles or delay cycles on the peripheral bus 26 during write transactions targeted for the main memory 18 through the peripheral write buffer 46.
Figure 4 illustrates the format of data stored in the peripheral write buffer 46 for one embodiment. The peripheral write buffer 46 stores a set of double word data values each with a corresponding snoop status flag. The snoop status flag indicates either snoop done (SD) or snoop not done (SND) depending upon the status of the corresponding snoop ahead request issued by the state machine 40 over the processor bus 24. The boundary in the peripheral write buffer 46 between data having an SD indication and data having an SND indication provides a snoop boundary between line n-1 and line n. The state machine 40 merges the data from the peripheral write buffer 46 into the write buffer 42 only if the corresponding snoop status flag indicates snoop done.
Figure 5 illustrates a read transaction from the main memory 18 over the peripheral bus 26. The read transaction on the peripheral bus 26 comprises an address phase, and a set of wait states, followed by a set of data phases comprising line 0 data - line n data. During the address phases of a read transaction, the originating bus agent on the peripheral bus 26 transfers the address of the initial data line being read from the main memory 18. As shown during the address phase, the state machine 40 issues a line 0 snoop request over the processor bus 24 that specifies the data line address provided during the address phase.
During the wait states on the peripheral bus 26, the line 0 snoop transaction on the processor bus 24 completes. Thereafter, a DRAM latency occurs while the peripheral read prefetch buffer 44 receives the corresponding data line from the main memory 18 via the read path 54. The line 0 prefetch to the peripheral read prefetch buffer 44 completes before the start of the line 0 data phase on the peripheral bus 26.
During the line 0 data phase on the peripheral bus 26, the peripheral read prefetch buffer 44 transfers the buffered read data over the peripheral bus 26 while the state machine 40 issues a line 1 snoop request over the processor bus 24 using an address generated as a next sequential cache line address from the address specified during the preceding address phase. Later in the line 0 data phase of the read transaction on the peripheral bus 26, the line 1 snoop transaction on the processor bus 24 completes. The completion of the line 1 snoop request enables a prefetch of the corresponding data line from the main memory 18 into the peripheral read prefetch buffer 44. Thereafter, the line 1 data from the peripheral read prefetch buffer 44 is transferred over the peripheral bus 26.
In addition, the state machine 40 monitors the occurrence of wait states on the peripheral bus 26 during read and write transactions. The state machine 40 counts the number of wait states that occur on the peripheral bus 26. The state machine 40 implements a disconnect time¬ out mechanism such that a disconnect is signaled over the peripheral bus 26 if the number of consecutive wait states on the peripheral bus 26 exceeds eight. The eight wait states conforms to the peripheral component interface (PCI) specification for maximum latency during data transfers. The disconnect indication on the peripheral bus 26 causes the bus agent that initiated the read transaction or the write transaction to terminate the transaction and try again later.
In the foregoing specification the invention has been described with reference to specific exemplary embodiments thereof. It will, however, be evident that various modifications and changes may be made thereto without departing from the broader spirit and scope of the invention as set forth in the appended claims. The specification and drawings are accordingly to be regarded as illustrative rather than a restrictive sense.

Claims

ς____What is claimed is:
1. A circuit that enables access to a memory over a processor bus and over a peripheral bus and that performs a series of snoop ahead transactions over the processor bus during access transactions to the memory that originate over the peripheral bus.
2. The circuit of claim 1 , further comprising a write buffer for buffering write data targeted for the memory and circuitry that merges write data received over the processor bus with write data received over the peripheral bus into the write buffer.
3. The circuit of claim 2, further comprising a peripheral write buffer that buffers write data received over the peripheral bus and that stores a snoop done flag for the write data that indicates whether a corresponding snoop ahead transaction for the write data is complete.
4. The circuit of claim 2, further comprising a peripheral read prefetch buffer that prefetches read data from the memory during read transactions over the peripheral bus only after a corresponding snoop ahead transaction for the read data is complete.
5. The circuit of claim 2, further comprising circuitry for controlling a cache memory coupled to the processor bus.
6. A computer system, comprising: processor and a cache memory each coupled to a processor bus; a set of bus agents coupled to a peripheral bus; circuit that enables access to a memory via the processor bus and via the peripheral bus and that performs a series of snoop ahead transactions over the processor bus during access transactions to the memory that originate over the peripheral bus.
7. The computer system of claim 6, wherein the circuit includes a write buffer for buffering write data targeted for the memory and circuitry that merges write data received over the processor bus with write data received over the peripheral bus into the write buffer.
8. The computer system of claim 7, wherein the circuit includes a peripheral write buffer that buffers write data received over the peripheral bus and that stores a snoop done flag for the write data that indicates whether a corresponding snoop ahead transaction for the write data is complete.
9. The computer system of claim 7, wherein the circuit includes a peripheral read prefetch buffer that prefetches read data from the memory during read transactions over the peripheral bus only after a corresponding snoop ahead transaction for the read data is complete.
10. The computer system of claim 7, wherein the circuit includes for controlling the cache memory coupled to the processor bus.
11. A method for accessing a shared memory comprising the step of performing a series of snoop ahead transactions over a processor bus during access transactions to the shared memory that originate over a peripheral bus.
12. The method of claim 11 , further comprising the steps of buffering write data targeted for the shared memory and merging write data received over the processor bus with write data received over the peripheral bus.
13. The method of claim 12, further comprising the steps of buffering write data received over the peripheral bus and storing a snoop done flag for the write data that indicates whether a corresponding snoop ahead transaction for the write data is complete.
14. The method of claim 12, further comprising the step of prefetching read data from the shared memory during read transactions over the peripheral bus only after a corresponding snoop ahead transaction for the read data is complete.
PCT/US1996/000506 1995-01-20 1996-01-16 Bus bridge circuit and method using snoop ahead operations WO1996022571A1 (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
AU46993/96A AU4699396A (en) 1995-01-20 1996-01-16 Bus bridge circuit and method using snoop ahead operations
DE69604564T DE69604564T2 (en) 1995-01-20 1996-01-16 BUS BRIDGE CIRCUIT AND PROCESS WITH PREDICTION SNOOPOPERATIONS
EP96902679A EP0804763B1 (en) 1995-01-20 1996-01-16 Bus bridge circuit and method using snoop ahead operations
HK98103810A HK1004432A1 (en) 1995-01-20 1998-05-04 Bus bridge circuit and method using snoop ahead operations

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US08/375,972 US5630094A (en) 1995-01-20 1995-01-20 Integrated bus bridge and memory controller that enables data streaming to a shared memory of a computer system using snoop ahead transactions
US08/375,972 1995-01-20

Publications (1)

Publication Number Publication Date
WO1996022571A1 true WO1996022571A1 (en) 1996-07-25

Family

ID=23483143

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1996/000506 WO1996022571A1 (en) 1995-01-20 1996-01-16 Bus bridge circuit and method using snoop ahead operations

Country Status (6)

Country Link
US (2) US5630094A (en)
EP (1) EP0804763B1 (en)
AU (1) AU4699396A (en)
DE (1) DE69604564T2 (en)
HK (1) HK1004432A1 (en)
WO (1) WO1996022571A1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1998025208A1 (en) * 1996-12-03 1998-06-11 Intel Corporation Computer system including multiple snooped, multiple mastered system buses and method for interconnecting said buses
EP0924620A2 (en) * 1997-12-22 1999-06-23 Compaq Computer Corporation Computer system including a bus bridge implementing adaptive speculative read operations
WO2001084330A2 (en) * 2000-05-03 2001-11-08 Broadcom Corporation A method and system for multi-channel transfer of data and control information

Families Citing this family (46)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6002720A (en) * 1991-01-07 1999-12-14 H. Lee Browne, D/B/A Greenwich Information Technologies Llc Audio and video transmission and receiving system
US5253275A (en) 1991-01-07 1993-10-12 H. Lee Browne Audio and video transmission and receiving system
US5832241A (en) * 1995-02-23 1998-11-03 Intel Corporation Data consistency across a bus transactions that impose ordering constraints
US5710906A (en) * 1995-07-07 1998-01-20 Opti Inc. Predictive snooping of cache memory for master-initiated accesses
US5859988A (en) * 1995-09-29 1999-01-12 Intel Corporation Triple-port bus bridge
US5712986A (en) * 1995-12-19 1998-01-27 Ncr Corporation Asynchronous PCI-to-PCI Bridge
US5717876A (en) * 1996-02-26 1998-02-10 International Business Machines Corporation Method for avoiding livelock on bus bridge receiving multiple requests
US6748463B1 (en) * 1996-03-13 2004-06-08 Hitachi, Ltd. Information processor with snoop suppressing function, memory controller, and direct memory access processing method
US5724613A (en) * 1996-05-06 1998-03-03 Vlsi Technology, Inc. System and method for automatically enabling and disabling a prefetching capability
US5920891A (en) * 1996-05-20 1999-07-06 Advanced Micro Devices, Inc. Architecture and method for controlling a cache memory
US5903906A (en) * 1996-06-05 1999-05-11 Compaq Computer Corporation Receiving a write request that allows less than one cache line of data to be written and issuing a subsequent write request that requires at least one cache line of data to be written
US6108741A (en) * 1996-06-05 2000-08-22 Maclaren; John M. Ordering transactions
US5872941A (en) * 1996-06-05 1999-02-16 Compaq Computer Corp. Providing data from a bridge to a requesting device while the bridge is receiving the data
US6021480A (en) * 1996-06-05 2000-02-01 Compaq Computer Corporation Aligning a memory read request with a cache line boundary when the request is for data beginning at a location in the middle of the cache line
US6055590A (en) * 1996-06-05 2000-04-25 Compaq Computer Corporation Bridge circuit comprising independent transaction buffers with control logic adapted to store overflow data in second buffer when transaction size exceeds the first buffer size
US5872939A (en) * 1996-06-05 1999-02-16 Compaq Computer Corporation Bus arbitration
US5987539A (en) * 1996-06-05 1999-11-16 Compaq Computer Corporation Method and apparatus for flushing a bridge device read buffer
US6035362A (en) * 1996-06-05 2000-03-07 Goodrum; Alan L. Storing data associated with one request while continuing to store data associated with a previous request from the same device
US6052513A (en) * 1996-06-05 2000-04-18 Compaq Computer Corporation Multi-threaded bus master
US6075929A (en) * 1996-06-05 2000-06-13 Compaq Computer Corporation Prefetching data in response to a read transaction for which the requesting device relinquishes control of the data bus while awaiting data requested in the transaction
US5768545A (en) * 1996-06-11 1998-06-16 Intel Corporation Collect all transfers buffering mechanism utilizing passive release for a multiple bus environment
US5815675A (en) * 1996-06-13 1998-09-29 Vlsi Technology, Inc. Method and apparatus for direct access to main memory by an I/O bus
US6021473A (en) * 1996-08-27 2000-02-01 Vlsi Technology, Inc. Method and apparatus for maintaining coherency for data transaction of CPU and bus device utilizing selective flushing mechanism
KR100200968B1 (en) * 1996-10-17 1999-06-15 윤종용 Host interface circuit of image making apparatus
US5832245A (en) * 1996-10-21 1998-11-03 Advanced Micro Devices, Inc. Method for isochronous flow control across an inter-chip bus
US6138192A (en) * 1996-12-31 2000-10-24 Compaq Computer Corporation Delivering a request to write or read data before delivering an earlier write request
US5835741A (en) * 1996-12-31 1998-11-10 Compaq Computer Corporation Bus-to-bus bridge in computer system, with fast burst memory range
US5974511A (en) * 1997-03-31 1999-10-26 Sun Microsystems, Inc. Cache subsystem with pseudo-packet switch
US5857082A (en) * 1997-04-25 1999-01-05 Intel Corporation Method and apparatus for quickly transferring data from a first bus to a second bus
US5935233A (en) * 1997-05-21 1999-08-10 Micron Electronics, Inc. Computer system with a switch interconnector for computer devices
US5931933A (en) * 1997-06-16 1999-08-03 International Business Machines Corporation Apparatus and method for communication and translation for selected one of a variety of data bus formats
US6157398A (en) * 1997-12-30 2000-12-05 Micron Technology, Inc. Method of implementing an accelerated graphics port for a multiple memory controller computer system
US7071946B2 (en) * 1997-12-30 2006-07-04 Micron Technology, Inc. Accelerated graphics port for a multiple memory controller computer system
US6321307B1 (en) * 1997-12-31 2001-11-20 Compaq Computer Corporation Computer system and method employing speculative snooping for optimizing performance
US6003106A (en) * 1998-05-27 1999-12-14 International Business Machines Corporation DMA cache control logic
US6275885B1 (en) * 1998-09-30 2001-08-14 Compaq Computer Corp. System and method for maintaining ownership of a processor bus while sending a programmed number of snoop cycles to the processor cache
US6434649B1 (en) * 1998-10-14 2002-08-13 Hitachi, Ltd. Data streamer
US6385703B1 (en) * 1998-12-03 2002-05-07 Intel Corporation Speculative request pointer advance for fast back-to-back reads
US6604162B1 (en) 2000-06-28 2003-08-05 Intel Corporation Snoop stall reduction on a microprocessor external bus
US6523093B1 (en) * 2000-09-29 2003-02-18 Intel Corporation Prefetch buffer allocation and filtering system
US6829665B2 (en) * 2001-09-28 2004-12-07 Hewlett-Packard Development Company, L.P. Next snoop predictor in a host controller
US7055005B2 (en) * 2003-04-07 2006-05-30 Hewlett-Packard Development Company, L.P. Methods and apparatus used to retrieve data from memory into a RAM controller before such data is requested
US6934770B1 (en) * 2003-05-01 2005-08-23 Adaptec, Inc. Method for aborting data transfer commands
US7430641B2 (en) * 2004-08-09 2008-09-30 Xiv Ltd. System method and circuit for retrieving into cache data from one or more mass data storage devices
US20080151765A1 (en) * 2006-12-20 2008-06-26 Sanal Chandran Cheruvathery Enhanced Jitter Buffer
US20090106498A1 (en) * 2007-10-23 2009-04-23 Kevin Michael Lepak Coherent dram prefetcher

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5072369A (en) * 1989-04-07 1991-12-10 Tektronix, Inc. Interface between buses attached with cached modules providing address space mapped cache coherent memory access with SNOOP hit memory updates
US5195089A (en) * 1990-12-31 1993-03-16 Sun Microsystems, Inc. Apparatus and method for a synchronous, high speed, packet-switched bus
JPH0637768A (en) * 1992-07-17 1994-02-10 Hitachi Ltd Bus bridge for information processing unit
US5325503A (en) * 1992-02-21 1994-06-28 Compaq Computer Corporation Cache memory system which snoops an operation to a first location in a cache line and does not snoop further operations to locations in the same line
US5335335A (en) * 1991-08-30 1994-08-02 Compaq Computer Corporation Multiprocessor cache snoop access protocol wherein snoop means performs snooping operations after host bus cycle completion and delays subsequent host bus cycles until snooping operations are completed
US5341487A (en) * 1991-12-20 1994-08-23 International Business Machines Corp. Personal computer having memory system with write-through cache and pipelined snoop cycles
US5353415A (en) * 1992-10-02 1994-10-04 Compaq Computer Corporation Method and apparatus for concurrency of bus operations
US5355467A (en) * 1991-06-04 1994-10-11 Intel Corporation Second level cache controller unit and system
US5379384A (en) * 1992-06-05 1995-01-03 Intel Corporation Configuration data loopback in a bus bridge circuit
US5420991A (en) * 1994-01-04 1995-05-30 Intel Corporation Apparatus and method for maintaining processing consistency in a computer system having multiple processors

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4814981A (en) * 1986-09-18 1989-03-21 Digital Equipment Corporation Cache invalidate protocol for digital data processing system
DE3782335T2 (en) * 1987-04-22 1993-05-06 Ibm MEMORY CONTROL SYSTEM.
US5317718A (en) * 1990-03-27 1994-05-31 Digital Equipment Corporation Data processing system and method with prefetch buffers
US5359723A (en) * 1991-12-16 1994-10-25 Intel Corporation Cache memory hierarchy having a large write through first level that allocates for CPU read misses only and a small write back second level that allocates for CPU write misses only
DE4391002T1 (en) * 1992-03-06 1995-02-23 Rambus Inc Pre-fetch into a cache to minimize memory access time and cache size in a computer system
US5511226A (en) * 1992-08-25 1996-04-23 Intel Corporation System for generating snoop addresses and conditionally generating source addresses whenever there is no snoop hit, the source addresses lagging behind the corresponding snoop addresses
US5572702A (en) * 1994-02-28 1996-11-05 Intel Corporation Method and apparatus for supporting read, write, and invalidation operations to memory which maintain cache consistency
US5572703A (en) * 1994-03-01 1996-11-05 Intel Corporation Method and apparatus for snoop stretching using signals that convey snoop results

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5072369A (en) * 1989-04-07 1991-12-10 Tektronix, Inc. Interface between buses attached with cached modules providing address space mapped cache coherent memory access with SNOOP hit memory updates
US5195089A (en) * 1990-12-31 1993-03-16 Sun Microsystems, Inc. Apparatus and method for a synchronous, high speed, packet-switched bus
US5355467A (en) * 1991-06-04 1994-10-11 Intel Corporation Second level cache controller unit and system
US5335335A (en) * 1991-08-30 1994-08-02 Compaq Computer Corporation Multiprocessor cache snoop access protocol wherein snoop means performs snooping operations after host bus cycle completion and delays subsequent host bus cycles until snooping operations are completed
US5341487A (en) * 1991-12-20 1994-08-23 International Business Machines Corp. Personal computer having memory system with write-through cache and pipelined snoop cycles
US5325503A (en) * 1992-02-21 1994-06-28 Compaq Computer Corporation Cache memory system which snoops an operation to a first location in a cache line and does not snoop further operations to locations in the same line
US5379384A (en) * 1992-06-05 1995-01-03 Intel Corporation Configuration data loopback in a bus bridge circuit
JPH0637768A (en) * 1992-07-17 1994-02-10 Hitachi Ltd Bus bridge for information processing unit
US5353415A (en) * 1992-10-02 1994-10-04 Compaq Computer Corporation Method and apparatus for concurrency of bus operations
US5420991A (en) * 1994-01-04 1995-05-30 Intel Corporation Apparatus and method for maintaining processing consistency in a computer system having multiple processors

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP0804763A4 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1998025208A1 (en) * 1996-12-03 1998-06-11 Intel Corporation Computer system including multiple snooped, multiple mastered system buses and method for interconnecting said buses
EP0924620A2 (en) * 1997-12-22 1999-06-23 Compaq Computer Corporation Computer system including a bus bridge implementing adaptive speculative read operations
EP0924620A3 (en) * 1997-12-22 2000-06-28 Compaq Computer Corporation Computer system including a bus bridge implementing adaptive speculative read operations
WO2001084330A2 (en) * 2000-05-03 2001-11-08 Broadcom Corporation A method and system for multi-channel transfer of data and control information
WO2001084330A3 (en) * 2000-05-03 2002-06-13 Broadcom Corp A method and system for multi-channel transfer of data and control information
US7107383B1 (en) 2000-05-03 2006-09-12 Broadcom Corporation Method and system for multi-channel transfer of data and control information

Also Published As

Publication number Publication date
EP0804763A4 (en) 1998-04-01
HK1004432A1 (en) 1998-11-27
EP0804763B1 (en) 1999-10-06
US5630094A (en) 1997-05-13
AU4699396A (en) 1996-08-07
DE69604564T2 (en) 2000-05-31
DE69604564D1 (en) 1999-11-11
EP0804763A1 (en) 1997-11-05
US6115796A (en) 2000-09-05

Similar Documents

Publication Publication Date Title
US5630094A (en) Integrated bus bridge and memory controller that enables data streaming to a shared memory of a computer system using snoop ahead transactions
US5572703A (en) Method and apparatus for snoop stretching using signals that convey snoop results
US11500797B2 (en) Computer memory expansion device and method of operation
US5463753A (en) Method and apparatus for reducing non-snoop window of a cache controller by delaying host bus grant signal to the cache controller
US5664150A (en) Computer system with a device for selectively blocking writebacks of data from a writeback cache to memory
US6405271B1 (en) Data flow control mechanism for a bus supporting two-and three-agent transactions
US5524235A (en) System for arbitrating access to memory with dynamic priority assignment
US5353415A (en) Method and apparatus for concurrency of bus operations
US6321296B1 (en) SDRAM L3 cache using speculative loads with command aborts to lower latency
US6718441B2 (en) Method to prefetch data from system memory using a bus interface unit
EP0617364B1 (en) Computer system which overrides write protection status during execution in system management mode
US5479636A (en) Concurrent cache line replacement method and apparatus in microprocessor system with write-back cache memory
US6321307B1 (en) Computer system and method employing speculative snooping for optimizing performance
US6321302B1 (en) Stream read buffer for efficient interface with block oriented devices
US5353429A (en) Cache memory systems that accesses main memory without wait states during cache misses, using a state machine and address latch in the memory controller
JP2001147854A (en) Processing system and method for optimizing storage in writing buffer unit and method for storing and distributing data
US5987570A (en) Performing overlapping burst memory accesses and interleaved memory accesses on cache misses
US6085261A (en) Method and apparatus for burst protocol in a data processing system
WO1997034228A1 (en) Information processor with snoop suppressing function, memory controller, and direct memory access processing method
US5761709A (en) Write cache for servicing write requests within a predetermined address range
US5893917A (en) Memory controller and method of closing a page of system memory
US5862358A (en) Method and apparatus for reducing the apparent read latency when connecting busses with fixed read reply timeouts to CPUs with write-back caches
US7406571B2 (en) Memory system and method for controlling the same, and method for maintaining data coherency
US5974511A (en) Cache subsystem with pseudo-packet switch
US5590286A (en) Method and apparatus for the pipelining of data during direct memory accesses

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AL AM AT AT AU AZ BB BG BR BY CA CH CN CZ CZ DE DE DK DK EE EE ES FI FI GB GE HU IS JP KE KG KP KR KZ LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SK TJ TM TR TT UA UG UZ VN AZ BY KG KZ RU TJ TM

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): KE LS MW SD SZ UG AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI

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: 1996902679

Country of ref document: EP

WWP Wipo information: published in national office

Ref document number: 1996902679

Country of ref document: EP

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

WWG Wipo information: grant in national office

Ref document number: 1996902679

Country of ref document: EP