WO1988002513A1 - Method and device to execute two instruction sequences in an order determined in advance - Google Patents

Method and device to execute two instruction sequences in an order determined in advance Download PDF

Info

Publication number
WO1988002513A1
WO1988002513A1 PCT/SE1987/000437 SE8700437W WO8802513A1 WO 1988002513 A1 WO1988002513 A1 WO 1988002513A1 SE 8700437 W SE8700437 W SE 8700437W WO 8802513 A1 WO8802513 A1 WO 8802513A1
Authority
WO
WIPO (PCT)
Prior art keywords
sequence
write
address
instruction
execution
Prior art date
Application number
PCT/SE1987/000437
Other languages
French (fr)
Inventor
Sten Edvard Johnson
Lars-Örjan KLING
Original Assignee
Telefonaktiebolaget L M Ericsson
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 Telefonaktiebolaget L M Ericsson filed Critical Telefonaktiebolaget L M Ericsson
Priority to KR1019880700625A priority Critical patent/KR920006769B1/en
Priority to AT87906632T priority patent/ATE73940T1/en
Priority to BR8707473A priority patent/BR8707473A/en
Priority to DE8787906632T priority patent/DE3777632D1/en
Publication of WO1988002513A1 publication Critical patent/WO1988002513A1/en
Priority to FI882468A priority patent/FI93907C/en
Priority to NO882413A priority patent/NO173207C/en
Priority to DK300788A priority patent/DK168135B1/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/38Concurrent instruction execution, e.g. pipeline, look ahead
    • G06F9/3885Concurrent instruction execution, e.g. pipeline, look ahead using a plurality of independent parallel functional units
    • G06F9/3889Concurrent instruction execution, e.g. pipeline, look ahead using a plurality of independent parallel functional units controlled by multiple instructions, e.g. MIMD, decoupled access or execute
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/30Arrangements for executing machine instructions, e.g. instruction decode
    • G06F9/38Concurrent instruction execution, e.g. pipeline, look ahead
    • G06F9/3824Operand accessing
    • G06F9/3834Maintaining memory consistency

Definitions

  • the present invention relates to a method and a device to execute two instruction sequences in an order determined in advance, the executions including selection of read instructions each containing its read address for retrieval of data information stored in one of a plurality of main memory locations accessible by their individual addresses, as well as selection of write instructions each containing its write address and data information for transferring this data information to a main memory location accessible by this write address and the data information used in conjunction with the execution of the sequence which is second due to the order not being guaranteed in advance independent of the data information obtained in conjunction with the execution of the sequence which is first due to the order.
  • a trivial, conventional solution of the above mentioned information handling problem resides in that the execution of that sequence which according to the above and hereinafter is called the second sequence, is not started until the execution of that sequence which according to the above and hereinafter is called the first sequence, is terminated.
  • This trivial solution is obtained as a natural necessity in a data processing system controlled by a single processor such that the sequences are executed one at a time using main memory locations common to both sequences.
  • the present invention relates ' ry data information processing while using a main memory common to both sequences.
  • both sequences are executed in parallel without having to begin by taking notice of the order determined in advance. To ensure the predetermined order it is, however, necessary to prevent that data
  • Addresses obtained due to read instructions selected during execution of the second sequence are intermediate-stored in the auxiliary memory. Every write address is scompared wit each of the read addresses stored in the auxiliary memory. As long as no
  • Write addresses and data information obtained on the basis of write instructions selected during execution of the second sequence are also stored in the auxiliary memory. Every read address selected during execution of the second sequence is compared with each of the write addresses intermediate-stored in the auxiliary memory. Should no address likeness then be determined data information is retrieved from the main memory location which is accessible with the aid of the read address in question, while data information assigned to the address in question is retrieved from the auxiliary memory if address likeness is ascertained.
  • the data information intermediate-stored in the auxiliary memory is transferred to the main memory locations which are accessible by the likewise intermediate-stored associated write addresses.
  • Figure 1 illustrates two data processing units connected via a bus system to a common main memory.
  • Figure 2 illustrates in more detail than Figure 1 an instruction memory device, a circuit for starting and identifying and a gate network, which are included in a data processing unit.
  • Figure 3 illustrates an intermediat - storage unit included in a data processing unit.
  • a data processing system includes a main memory 1 for 5 storing data information, processed by at least two data processing units 2, which via a bus system 3 are each connected to the main memory.
  • Each of the data processing units executes an instruction sequence for controlling assigned functional units (not illustrated in Figure 1) in carrying out assigned system functions. Depending on whether the data processing units are controlled by a
  • each data processing unit includes a means 5 called start/identifying circuit, for stating whether its own sequence is .the "first independent .or T_he.-sec ⁇ nd, ⁇ ossibly dependent isequence, an Tan intermediate storage unit 6 for intermediate-storing addresses and data
  • Figure 1 illustrates a multibit address bus 8, a line 9 for transferring write signals, and a multibit data bus 10.
  • the data bus is two-way connected to the main memory 1 and the data processing units 2, while the address bus 8 and write signal line 9 are one ⁇ way connected to an addressing circuit 11 and a write activation input 12 of the main memory, on which the reception of an address and a write signal cause the contents of the data bus to be transferred to a main memory location accessible by the address.
  • the start identifying circuits 5 of the data processing units are each provided with a starting signal output 13 which is connected to a starting signal input 14 on the other circuit 5.
  • start identifying circuits 5 are illustrated as being each provided with their identifying signal output 15, which is connected to a first identifying signal input of the intermediate storage unit 6 of the other data processing unit.
  • Figure 1 does not show the conventional buffers and activations which are used and carried out in connection with the data transfers between the data processing units and the bus system, the bus system, as it will be described later, is also used for transferring addresses and write signals between the data processing units.
  • Figure 2 illustrates in more detail than Figure 1 some data processing unit details, knowledge of which is required for understanding the proposed infor- mation handling.
  • the instruction memory device 4 includes an instruction memory 17 for storing an instruction sequence, which is read out with the aid of a selection device 18 from the instruction memory, one instruction at a time, due to activation by one of the selection device-outputs.
  • the sequence includes three instruction types.
  • the first type hereinafter designated “read instruction”, which is identified by a binary ONE set read bit position 19, is used for ordering the retrieval of data from a main memory location, to which access is obtained with the aid of an address, e.g. Al, stored in a number of address bit positions 20, the address being included in the read bit marked instruction.
  • the second instruction type hereinafter designated “write instruction”, which is identified by a ONE set write bit position 21, is used for ordering that data, e.g. D2, is written into a main memory location, this data being included in the write bit marked instruction and stored in a number of data bit positions 22, while using an address., e.g.
  • A2 which is also included In -the write - bit-* marked ' instruction and stored in the address bit positions 20.
  • the third instruction type which is identified by ZERO set bit positions 19 and 21, is used for ordering Information handling without reading from or writing into the main memory.
  • the use of the information, e.g. SI, stored in the bit positions 20 and 22, which are associated with a third type instruction, are not within the scope of the invention, but Figure 2 illustrates a first OR gate 23, which is included in said gate network 7 and has its inputs connected to the bit positions 19 and 21 and its output connected to an inverting activating input of " a H first AND gate device 4, wh ⁇ th i -an-activated state transfers the contents, e.g. SI, of the bit positions 20 and 22 directly to the data processing unit. Instructions of the third type do not load the bus system or the main memory, which may therefore be common to a plurality of data processing units without making too large speed demands on the main memory and bus system.
  • the start/identifying circuits 5 of the data processing units each include second and third OR gates 25 and 26, according to Figure 2.
  • Gate 25 has one of Its inputs connected to the output of the gate 26, the inputs of which are connected to the above-mentioned starting signal input 14 and to a restarting signal output .27 of the intermediate .storage _unit*6 included in the -same-data • processing unit.
  • the data processing units each include a starting signal generator, not illustrated in Figure 2, the output of which is connected via a starting terminal 28 to the other input of gate 25, to the above-mentioned starting signal output 13 and to the setting input of a first flipflop 29, with an output constituting the above-mentioned identifying signal output 15.
  • a second flipflop 30 has its setting input connected to the output of gate 26 and its output connected to a second identifying signal input 31 on the local inter - mediate storage unit 6.
  • the gate 25 selects an instruction of the third type, which is a starting instruction with the contents SI, due to which the data processing unit starts the associated instruction sequence execution with the aid of its selection device 18. It is assumed that the sequence stored in the instruction memory 17 is terminated in conjunction with the activation of a selection device output 32, which is connected to the resetting inputs of the flipflops 29 and 30.
  • Execution of the first sequence is started by the local starting signal generator. Consequently, a binary ONE on the output of the first flipflop 29 indicates that the instruction selections associated with the local sequence are in progress and that the local sequence is the first sequence. A binary ONE on the output of the second flipflop 30 indicates that the instruction selections associated with the local sequence are in progress and that the local sequence is the second sequence.
  • the instructions of both sequences are selected in parallel, which is achieved e.g. by means of a starting signal transfer from the starting signal output 13 associated with the first sequence to the starting signal input 14 associated with the second sequence.
  • Each gate network 7 includes a second AND gate device 37 for transferring, when in an active state, via a first OR gate device 38, the contents of the bit positions 20-22 to the address bus 8, the write signal line 9 and the data bus 10 of the bus system.
  • the gate device 37 has its activation input connected to the output of a fourth OR gate 39, the first and second inputs of which are respectively connected to the outputs of a first 40 and second control gate 41.
  • the control gate 40 has its first input connected to the write bit position 21 of the instruction memory and its second input connected to the output of the first flipflop 29.
  • the control gate 41 has its first input connected to the read bit position 19 of the instruction memory and its second inverting input connected to a category signal output 42 of the intermediate storage unit 6.
  • the gate device 37 is activated due to all read instructions and write instructions which are selected during the execution of the first sequence, as well as due to a read instruction of the first 10 category. However, the gate device 37 is not activated due to a third tyoe of instruction or a write instruction which is selected during the execution of the second sequence or due to a read instruction of the second category.
  • Every gate network further includes a third AND gate device 43 for trans ⁇ ferring in an active state via a second OR gate device 44 data information
  • the gate device 43 has its activating input connected to the category signal output 42.
  • the OR gate device 44 has its second input connected to the data bus 10 of the bus system.
  • every gate network 7 includes a fourth AND gate device 46 for transferring in an activated state data and address information as well as write signals from output terminals 45, 47 and 48 of the intermediate storage unit 6 associated with the same date processing unit to the bus system 3.
  • Activation 25 jof ⁇ ie-ga e device ⁇ 6 ⁇ wouid -ij ⁇ -oescriberi in oonjimction wit vthe description of the intermediate storage unit.
  • Figure 3 illustrates an intermediate storage unit which includes an auxiliary memory 49.
  • the latter has columns for intermediate-storing of data information as well as read addresses. and write addresses, transferred via the -30 terminals 33-36, the columns being- selected for writing, reading and erasure with the aid of a scanning device 50. It will be later described how the auxiliary memory is erased when the associated instruction sequence is finally executed.
  • the front edge of a scanning pulse sent from a fifth OR gate 51 sets the scanning device to zero.
  • Access to the auxiliary memory columns is obtained by the front edges of stepping pulses which are generated by a stepping generator 52 and transferred via an activated stepping AND gate 53 to a stepping terminal 54 of the scanning device.
  • the stepping pulses are also sent to a read activating terminal 55, resulting in that the column contents are read one at a time.
  • the intermediate storage unit includes a first NOR gate 56, with its inputs adapted for receiving the write markings and read markings intermediate-stored in write bit positions 57 and read bit positions 58 of the auxiliary memory, and with its inverting output connected via a stop OR gate 59 to an inverting input of the stepping AND gate 53 and to first inputs of a third and a fourth AND control gate 60 and 61, the outputs of which are connected to the OR gate 51.
  • the control gate 60 has its second input connected to a fifth AND control gate 62, the first input of which is connected to the above-mentioned identifying signal input 31, and the second input of which is connected to the output of the above-mentioned first OR gate 23 via a first control terminal 63.
  • a scanning operation is ordered upon selection of a read instruction or a write instruction.
  • a stop in stepping is obtained when an unoccupied column is read; i.e. a column which is neither write marked or read marked.
  • a condition for starting scanning is that the preceding stepping has been stopped.
  • the first NOR gate 56 is connected via a write OR gate 64 to a write activating input 65 of the scanning device. It is assumed that a write instruction with associated address A3 and data D3/1 as well as a read instruction with associated address A4 constitute the first and second instructions of the second sequence directed towards the main memory, these instructions being inter- mediate-stored in the described way in the first and second columns of the auxiliary memory, these columns having bit positions 57 and 58, which were set to zero prior to respective storage.
  • the control gate 61 has its second input connected to a sixth AND control gate 66 t with the first inpuL.connected to the above-mentioned first identifying input 16, and with the second input connected to the bus system write signal line 9 via a first bus terminal 67.
  • a scanning operation is thus also ordered for the 5 second sequence intermediate storage unit upon selecting a write instruction associated with the first sequence.
  • the timing frequency of the stepping generator 52 is assumed to be sufficiently high in relation to the instruction selection speed that all the scanning operations ordered via the control gates 62 and 66 have time to be carried out.
  • the intermediate storage unit includes a first comparison circuit 68, the output of which is activated on there being likeness between the address obtained via terminal 36 from the instruction memory 17 and one of the addresses read from the auxiliary memory address .bit positions -69 -during a scanning operation.
  • a seventh AND control gate 70 has its output connected to stop OR gate 59, its 5 first input connected to the output of the first comparison circuit 68 and its second input disposed for receiving the write markings intermediate-stored in the auxiliary memory write bit positions 57. There is thus obtained that a scanning operation is stopped if a read address or write address selected from the instruction memory agrees with an intermediate-stored write address.
  • the write 43 ft gate _£4 has its second input connected to the joutput of an 'eight AND control gate 71, the first input of which is connected to the first comparison circuit, its second and third inputs respectively receiving the write markings intermediate stored in the auxiliary memory write bit positions 57 and the write marking selected via the terminal 34 from the instruction memory 17. 5 If a write address e g. A3, selected from the instruction memory, agrees with an intermediate-stored write address, ' he coJ imn in .the auxiliary- memory where the scan has been stopped due to the selected write address is updated so that, for example, the auxiliary memory first column intermediate-stores data information D3/2 in the auxiliary memory data bit positions 72 after updating.
  • the control gate 70 has. its output connected to one input of a ninth AND control gate 73, which" has its second input disposed for -receiving the read marking selected via terminal 33 from the instruction memory 17, the output of the gate 73 constituting the above-mentioned category signal output 42, which consequently has transferred a binary ZERO to the above-mentioned control gate 41 and gate device 42 in conjunction with intermediate-storage of the address A4 in the second column of the auxiliary memory. It is assumed that the selection of the read instruction with associated address A4 is repeated a first time after selection of a number of instructions (not illustrated in Figure 3) with addresses other than A3 and A4.
  • the selection repeated for the first time does not result in a binary ONE on the category signal output 42 and a scanning stop due to address likeness determined by the first comparison circuit 68, so that the read instruction repeated a first time is intermediate stored a second time in a column where bit positions 57 and 58 were previously set to ZERO.
  • the second sequence read instructions with associated address Ax are called "first category instructions" if they are selected before the same address Ax is present in the second sequence in conjunction with the selection of a write instruction.
  • An output 42 set to ZERO identifies a first category instruction.
  • the write marking in the bit position 57 is retained and no read marking is carried out in the bit position 58, and a binary ONE is obtained on the category signal output 42, whereby the read instruction with associated address A4 repeated a second time is identified as a second category instruction. Consequently, the auxiliary memory contains no second category instructions, the selection of which results, however, in that the control gate 41 illustrated in Figure 2 stops the instruction transfer to the bus system 3 and that the data information, e.g. D4, which is read in conjunction with the associated scanning stop from the auxiliary memory data bit positions 72 is transferred to the data processing unit via the gate devices 43 and 44 illustrated in Figure 2.
  • the intermediate storage unit includes a second comparison circuit 74, the output of which is activated on likeness between the address obtained from the
  • the second comparison circuit is connected to the first input of a tenth AND control gate 76, the second input of which is connected to the above-mentioned control gate 66, its third input receiving the read markings intermediate-stored in the auxiliary memory bit positions 58 and its output, constituting the above mentioned restarting signal output 27, being connected to an erase OR gate 77.
  • the OR gate 26 illustrated in Figure 2 receives a restarting signal from an activated control gate 76.
  • the scanning device 50 receives from an activated erase OR gate 77 an erasure order pulse, due to which all address information intermediate-stored in the auxiliary memory is immediately cancelled.
  • the intermediate storage unit includes a second NOR gate 78 with inputs connected to the above-mentioned identifying signal inputs 16 and 31 and which has its inverting output connected -to ihe-OR gate 51.
  • AfiJ en the selections of the first and the second sequences are terminated, a scanning operation is con ⁇ sequently ordered, although the scanning is without result in the intermediate storage unit included in the data processing unit which has executed the first sequence.
  • the output of the NOR gate 78 is connected to the first input of an eleventh AND control ⁇ ,ate 79, the second .input of which .receives the write markings intermediate stored in the auxiliary memory bit positions 57, ' the output of gate 79 being connected to the above mentioned gate device 46 via a second control terminal 80.
  • the intermediate storage unit according to Figure 3 has a twelfth AND control gate 81, with its inputs connected to the NOR gates 56 and 78 and its output connected to erase OR gate 77. There is obtained an erase order pulse when the scanning operation started by NOR gate 78 is terminated.

Abstract

A data processing system executes two instruction sequences in an order determined in advance. With the aid of instructions a main memory (1) common to both sequences is activated for data information reading/writing. Increased data handling capacity is achieved in the following manner: both sequences (5) are executed in parallel to start with. During execution of the first sequence the main memory is prevented from being activated for writing due to the second sequence write instructions (37, 39, 41). A write address and data information included in a write instruction (49, 50) associated with the second sequence are intermediate-stored. The intermediate-stored write address is compared with the read addresses (68, 70, 73) of the second sequence, and data information is prevented from being read from the main memory upon likeness in addresses, the intermediate-stored data information (37, 38, 39, 41, 44) being read instead. An address included in a read instruction associated with the second sequence is intermediate-stored if this address has not been previously selected in conjunction with a write instruction associated with the second sequence (49, 50). The intermediate-stored read address is compared with the write addresses of the first sequence and execution of the second sequence (66, 74, 76) is restarted upon likeness in addresses.

Description

METHOD AND DEVICE TO EXECUTE TWO INSTRUCTION SEQUENCES IN AN ORDER DETERMINED IN ADVANCE
TECHNICAL FIELD
The present invention relates to a method and a device to execute two instruction sequences in an order determined in advance, the executions including selection of read instructions each containing its read address for retrieval of data information stored in one of a plurality of main memory locations accessible by their individual adresses, as well as selection of write instructions each containing its write address and data information for transferring this data information to a main memory location accessible by this write address and the data information used in conjunction with the execution of the sequence which is second due to the order not being guaranteed in advance independent of the data information obtained in conjunction with the execution of the sequence which is first due to the order.
BACKGROUND ART
A trivial, conventional solution of the above mentioned information handling problem resides in that the execution of that sequence which according to the above and hereinafter is called the second sequence, is not started until the execution of that sequence which according to the above and hereinafter is called the first sequence, is terminated. This trivial solution is obtained as a natural necessity in a data processing system controlled by a single processor such that the sequences are executed one at a time using main memory locations common to both sequences.
It is known to increase data processing capacity by parallel execution of the instruction sequences. As long as the sequences are guaranteed in advance mutually independent, fault-free parallel operation is achieved with the aid of so-called pre-processing or multi-processing, or also with the aid of a one- processor system which includes at least two data processing units, each of which executes its instruction sequence. It is known to realise information handling both by means of a main memory which is common to a plurality of data processing units and by means of a plurality of separate memories each associated with its data processing unit and mutually updated from time to time.
When there are sensitive instruction sequences which affect each other, and which must therefore be executed in a prescribed order, there is used e.g. 5 according to the journal "Computer Design", August 15, 1985, pp 76-81" or "Balance 8000 System Technical Summary, Sequent Computer Systems, Inc" programming languages, cσmpilators and sequence hardware for parallel proces¬ sing of mutually independent sequences while parallel processing of the sensitive sequences is prevented.
DISCLOSURE OF INVENTION
H) As. already mentioned .In -the introduction, the present invention relates ' ry data information processing while using a main memory common to both sequences. In the proposed Information processing, both sequences are executed in parallel without having to begin by taking notice of the order determined in advance. To ensure the predetermined order it is, however, necessary to prevent that data
15 information obtained due to selected write instructions associated with the second sequence is transferred to the main memory locations during the preceding execution of the first sequence. The dependence of the second
.sequence on the ζftrst sequence as 'monitored rand the prescribed order is achieved with the aid of an intermediate storage unit which includes an
20 auxiliary memory and comparison circuits.
Addresses obtained due to read instructions selected during execution of the second sequence are intermediate-stored in the auxiliary memory. Every write address
Figure imgf000004_0001
is scompared wit each of the read addresses stored in the auxiliary memory. As long as no
25 likeness of address is determined, no data information dependent on data infor¬ mation obtained during the execution of the first sequence is used during the execution of the second sequence. If it occurs during execution of the second sequence that inf ormtion has been retrieved from a main memory-location, this information then "corrected due to a write operation associated with the first
30 sequence, i.e. if the two sequences are no longer mutually independent, the auxiliary memory is erased and the instruction selections of the second sequence are started once again. Restarting execution of the second sequence then takes place at a time when a first part of the first sequence is already executed, and consequently there is less risk of the second sequence being dependent on the remaining part of the first sequence.
Write addresses and data information obtained on the basis of write instructions selected during execution of the second sequence are also stored in the auxiliary memory. Every read address selected during execution of the second sequence is compared with each of the write addresses intermediate-stored in the auxiliary memory. Should no address likeness then be determined data information is retrieved from the main memory location which is accessible with the aid of the read address in question, while data information assigned to the address in question is retrieved from the auxiliary memory if address likeness is ascertained.
When the execution of the first sequence is terminated, the data information intermediate-stored in the auxiliary memory is transferred to the main memory locations which are accessible by the likewise intermediate-stored associated write addresses.
In using the proposed information handling there is obtained an increase in the data processing capacity, if the execution of the second sequence is indepen- dent of at least the instructions selected at the excution termination of the first sequence. The increased capacity is further improved if a selected address is intermediate-stored as a read address in the auxiliary memory solely if this address has not been present as a write address earlier during execution of the second sequence.
The characterizing features of the invention are apparent from the claims.
BRIEF DESCRIPTION OF DRAWINGS
The invention will now be described in detail below and with reference to the accompanying drawing, on which Figure 1 illustrates two data processing units connected via a bus system to a common main memory. Figure 2 illustrates in more detail than Figure 1 an instruction memory device, a circuit for starting and identifying and a gate network, which are included in a data processing unit. Figure 3 illustrates an intermediat - storage unit included in a data processing unit.
DETAILED DESCRIPTION OF A MODE FOR CARRYING OUT THE INVEN¬ TION
A data processing system according to Figure 1 includes a main memory 1 for 5 storing data information, processed by at least two data processing units 2, which via a bus system 3 are each connected to the main memory. Each of the data processing units executes an instruction sequence for controlling assigned functional units (not illustrated in Figure 1) in carrying out assigned system functions. Depending on whether the data processing units are controlled by a
^TO "common processor or "whether each' unit includes a> plurality of processors, there is conventionally obtained in principle a one- or multi-processor system, including a common bus and a common main memory. Such a data processing system including a plurality of data processing units and a bus is described, for example, in Intel's "APX" 286 Hardware Reference Manual" and the article
15 "Backup Support gives VME bus powerful milti-processing architecture" in "Electronics" March 22, 1984.
'Figure liliustoates in .a greatly -simplified way 3a*άng intα account the-preserit invention, two data processing units 2 for executing the above mentioned first and second instruction sequences in an order determined in advance. Apart from
20 a conventional instruction memory device 4 for selecting, one at a time, instructions stored in an instruction memory, each data processing unit includes a means 5 called start/identifying circuit, for stating whether its own sequence is .the "first independent .or T_he.-secσnd,φossibly dependent isequence, an Tan intermediate storage unit 6 for intermediate-storing addresses and data
25 information included in the instructions, such data being however solely used in the data processing unit executing the second sequence, and finally a gate network 7 for controlling connection of the data processing unit to the bus -system .3.
It will be understood from the later description that the bus system load is 30 increased somewhat due to the required instruction intermediate-storages, but it is assumed that the accessibility capacity of the bus system towards the data processing units and towards the main memory is such that the total data processing can be carried out without disturbance in the order determined in advance. Of the bus system, Figure 1 illustrates a multibit address bus 8, a line 9 for transferring write signals, and a multibit data bus 10. It is further indicated that the data bus is two-way connected to the main memory 1 and the data processing units 2, while the address bus 8 and write signal line 9 are one¬ way connected to an addressing circuit 11 and a write activation input 12 of the main memory, on which the reception of an address and a write signal cause the contents of the data bus to be transferred to a main memory location accessible by the address. On solely receiving an address, data is transferred from the main memory, this data being stored in a main memory location accessible by the address, via the data bus 10 to the data processing unit sending the address. The start identifying circuits 5 of the data processing units are each provided with a starting signal output 13 which is connected to a starting signal input 14 on the other circuit 5. There is thus indicated the possibility of executing both sequences in parallel that is described hereinafter and used in the proposed information handling.
Finally, in Figure 1 the start identifying circuits 5 are illustrated as being each provided with their identifying signal output 15, which is connected to a first identifying signal input of the intermediate storage unit 6 of the other data processing unit.
Figure 1 does not show the conventional buffers and activations which are used and carried out in connection with the data transfers between the data processing units and the bus system, the bus system, as it will be described later, is also used for transferring addresses and write signals between the data processing units.
Figure 2 illustrates in more detail than Figure 1 some data processing unit details, knowledge of which is required for understanding the proposed infor- mation handling.
The instruction memory device 4 includes an instruction memory 17 for storing an instruction sequence, which is read out with the aid of a selection device 18 from the instruction memory, one instruction at a time, due to activation by one of the selection device-outputs.
The sequence includes three instruction types. The first type, hereinafter designated "read instruction", which is identified by a binary ONE set read bit position 19, is used for ordering the retrieval of data from a main memory location, to which access is obtained with the aid of an address, e.g. Al, stored in a number of address bit positions 20, the address being included in the read bit marked instruction. The second instruction type, hereinafter designated "write instruction", which is identified by a ONE set write bit position 21, is used for ordering that data, e.g. D2, is written into a main memory location, this data being included in the write bit marked instruction and stored in a number of data bit positions 22, while using an address., e.g. A2, which is also included In -the write - bit-* marked' instruction and stored in the address bit positions 20. The third instruction type, which is identified by ZERO set bit positions 19 and 21, is used for ordering Information handling without reading from or writing into the main memory. The use of the information, e.g. SI, stored in the bit positions 20 and 22, which are associated with a third type instruction, are not within the scope of the invention, but Figure 2 illustrates a first OR gate 23, which is included in said gate network 7 and has its inputs connected to the bit positions 19 and 21 and its output connected to an inverting activating input of "aH first AND gate device 4, whϊth i -an-activated state transfers the contents, e.g. SI, of the bit positions 20 and 22 directly to the data processing unit. Instructions of the third type do not load the bus system or the main memory, which may therefore be common to a plurality of data processing units without making too large speed demands on the main memory and bus system.
The start/identifying circuits 5 of the data processing units each include second and third OR gates 25 and 26, according to Figure 2. Gate 25 has one of Its inputs connected to the output of the gate 26, the inputs of which are connected to the above-mentioned starting signal input 14 and to a restarting signal output .27 of the intermediate .storage _unit*6 included in the -same-data processing unit. The data processing units each include a starting signal generator, not illustrated in Figure 2, the output of which is connected via a starting terminal 28 to the other input of gate 25, to the above-mentioned starting signal output 13 and to the setting input of a first flipflop 29, with an output constituting the above-mentioned identifying signal output 15. A second flipflop 30 has its setting input connected to the output of gate 26 and its output connected to a second identifying signal input 31 on the local inter - mediate storage unit 6. In its activated state the gate 25 selects an instruction of the third type, which is a starting instruction with the contents SI, due to which the data processing unit starts the associated instruction sequence execution with the aid of its selection device 18. It is assumed that the sequence stored in the instruction memory 17 is terminated in conjunction with the activation of a selection device output 32, which is connected to the resetting inputs of the flipflops 29 and 30.
Execution of the first sequence is started by the local starting signal generator. Consequently, a binary ONE on the output of the first flipflop 29 indicates that the instruction selections associated with the local sequence are in progress and that the local sequence is the first sequence. A binary ONE on the output of the second flipflop 30 indicates that the instruction selections associated with the local sequence are in progress and that the local sequence is the second sequence. In accordance with the invention, the instructions of both sequences are selected in parallel, which is achieved e.g. by means of a starting signal transfer from the starting signal output 13 associated with the first sequence to the starting signal input 14 associated with the second sequence.
The instructions selected from the instruction memory 17 are transferred to the gate network 7 of the data processing unit, and via terminals 33-36 to the intermediate storage unit 6 of the data processsing unit. Each gate network 7 includes a second AND gate device 37 for transferring, when in an active state, via a first OR gate device 38, the contents of the bit positions 20-22 to the address bus 8, the write signal line 9 and the data bus 10 of the bus system. The gate device 37 has its activation input connected to the output of a fourth OR gate 39, the first and second inputs of which are respectively connected to the outputs of a first 40 and second control gate 41. The control gate 40 has its first input connected to the write bit position 21 of the instruction memory and its second input connected to the output of the first flipflop 29. The control gate 41 has its first input connected to the read bit position 19 of the instruction memory and its second inverting input connected to a category signal output 42 of the intermediate storage unit 6.
It is described later on that there "is a irst and a second category among the read instructions which are selected during the execution of the second sequence, that the first and second categories are respectively identified by a 5 binary ZERO and ONE transferred via said category output 42, and that said category output 42 sends a binary ZERO if the associated data processing unit executes the first sequence. Consequently-, the gate device 37 is activated due to all read instructions and write instructions which are selected during the execution of the first sequence, as well as due to a read instruction of the first 10 category. However, the gate device 37 is not activated due to a third tyoe of instruction or a write instruction which is selected during the execution of the second sequence or due to a read instruction of the second category.
Every gate network further includes a third AND gate device 43 for trans¬ ferring in an active state via a second OR gate device 44 data information
15 obtained from a terminal 45 in the intermediate storage unit 6, and which are processed by the associated data processing unit. The gate device 43 has its activating input connected to the category signal output 42. The OR gate device 44 has its second input connected to the data bus 10 of the bus system. When selecting a second category instruction, data information is thus not
20 .retrie vedfrom the main .memory A but faom tne intermediate .storage .unit 6.
Finally, every gate network 7 includes a fourth AND gate device 46 for transferring in an activated state data and address information as well as write signals from output terminals 45, 47 and 48 of the intermediate storage unit 6 associated with the same date processing unit to the bus system 3. Activation 25 jof ϋie-ga e device τ 6 ^wouid -ijø-oescriberi in oonjimction wit vthe description of the intermediate storage unit.
Figure 3 illustrates an intermediate storage unit which includes an auxiliary memory 49. The latter has columns for intermediate-storing of data information as well as read addresses. and write addresses, transferred via the -30 terminals 33-36, the columns being- selected for writing, reading and erasure with the aid of a scanning device 50. It will be later described how the auxiliary memory is erased when the associated instruction sequence is finally executed. The front edge of a scanning pulse sent from a fifth OR gate 51 sets the scanning device to zero. Access to the auxiliary memory columns is obtained by the front edges of stepping pulses which are generated by a stepping generator 52 and transferred via an activated stepping AND gate 53 to a stepping terminal 54 of the scanning device. The stepping pulses are also sent to a read activating terminal 55, resulting in that the column contents are read one at a time.
The intermediate storage unit according to Figure 3 includes a first NOR gate 56, with its inputs adapted for receiving the write markings and read markings intermediate-stored in write bit positions 57 and read bit positions 58 of the auxiliary memory, and with its inverting output connected via a stop OR gate 59 to an inverting input of the stepping AND gate 53 and to first inputs of a third and a fourth AND control gate 60 and 61, the outputs of which are connected to the OR gate 51. The control gate 60 has its second input connected to a fifth AND control gate 62, the first input of which is connected to the above-mentioned identifying signal input 31, and the second input of which is connected to the output of the above-mentioned first OR gate 23 via a first control terminal 63.
Consequently, scanning pulses are only generated in the intermediate storage unit included in the data processing unit executing the second sequence. A scanning operation is ordered upon selection of a read instruction or a write instruction. A stop in stepping is obtained when an unoccupied column is read; i.e. a column which is neither write marked or read marked. A condition for starting scanning is that the preceding stepping has been stopped.
The first NOR gate 56 is connected via a write OR gate 64 to a write activating input 65 of the scanning device. It is assumed that a write instruction with associated address A3 and data D3/1 as well as a read instruction with associated address A4 constitute the first and second instructions of the second sequence directed towards the main memory, these instructions being inter- mediate-stored in the described way in the first and second columns of the auxiliary memory, these columns having bit positions 57 and 58, which were set to zero prior to respective storage. The control gate 61 has its second input connected to a sixth AND control gate 66t with the first inpuL.connected to the above-mentioned first identifying input 16, and with the second input connected to the bus system write signal line 9 via a first bus terminal 67. A scanning operation is thus also ordered for the 5 second sequence intermediate storage unit upon selecting a write instruction associated with the first sequence. The timing frequency of the stepping generator 52 is assumed to be sufficiently high in relation to the instruction selection speed that all the scanning operations ordered via the control gates 62 and 66 have time to be carried out.
o The intermediate storage unit includes a first comparison circuit 68, the output of which is activated on there being likeness between the address obtained via terminal 36 from the instruction memory 17 and one of the addresses read from the auxiliary memory address .bit positions -69 -during a scanning operation. A seventh AND control gate 70 has its output connected to stop OR gate 59, its 5 first input connected to the output of the first comparison circuit 68 and its second input disposed for receiving the write markings intermediate-stored in the auxiliary memory write bit positions 57. There is thus obtained that a scanning operation is stopped if a read address or write address selected from the instruction memory agrees with an intermediate-stored write address.
D The write 43 ft gate _£4 -has its second input connected to the joutput of an 'eight AND control gate 71, the first input of which is connected to the first comparison circuit, its second and third inputs respectively receiving the write markings intermediate stored in the auxiliary memory write bit positions 57 and the write marking selected via the terminal 34 from the instruction memory 17. 5 If a write address e g. A3, selected from the instruction memory, agrees with an intermediate-stored write address, ' he coJ imn in .the auxiliary- memory where the scan has been stopped due to the selected write address is updated so that, for example, the auxiliary memory first column intermediate-stores data information D3/2 in the auxiliary memory data bit positions 72 after updating.
D The control gate 70 has. its output connected to one input of a ninth AND control gate 73, which" has its second input disposed for -receiving the read marking selected via terminal 33 from the instruction memory 17, the output of the gate 73 constituting the above-mentioned category signal output 42, which consequently has transferred a binary ZERO to the above-mentioned control gate 41 and gate device 42 in conjunction with intermediate-storage of the address A4 in the second column of the auxiliary memory. It is assumed that the selection of the read instruction with associated address A4 is repeated a first time after selection of a number of instructions (not illustrated in Figure 3) with addresses other than A3 and A4. There is obtained that the selection repeated for the first time does not result in a binary ONE on the category signal output 42 and a scanning stop due to address likeness determined by the first comparison circuit 68, so that the read instruction repeated a first time is intermediate stored a second time in a column where bit positions 57 and 58 were previously set to ZERO. The second sequence read instructions with associated address Ax are called "first category instructions" if they are selected before the same address Ax is present in the second sequence in conjunction with the selection of a write instruction. An output 42 set to ZERO identifies a first category instruction.
It is assumed that a write instruction with associated address A4 and data D4 is selected after selection of the above mentioned read instruction repeated for the first time and after selection of a number of instructions (not illustated in Figure 3) with addresses other than A3 and A4. There is thus obtained that this write instruction is intermediate-stored in a previously unoccupied column. It is further assumed that selection of the read instruction with associated address A4 is repeated a second time after selection of the mentioned write instruction with the address A4. There is obtained a scanning stop in conjunction upon reading the intermediate-stored write instruction with the address A4. No updating is obtained, i.e. the write marking in the bit position 57 is retained and no read marking is carried out in the bit position 58, and a binary ONE is obtained on the category signal output 42, whereby the read instruction with associated address A4 repeated a second time is identified as a second category instruction. Consequently, the auxiliary memory contains no second category instructions, the selection of which results, however, in that the control gate 41 illustrated in Figure 2 stops the instruction transfer to the bus system 3 and that the data information, e.g. D4, which is read in conjunction with the associated scanning stop from the auxiliary memory data bit positions 72 is transferred to the data processing unit via the gate devices 43 and 44 illustrated in Figure 2. The intermediate storage unit includes a second comparison circuit 74, the output of which is activated on likeness between the address obtained from the
• address bus 8 via a second bus terminal 75 and one of the addresses read from the auxiliary memory address bit positions 69 during a scanning operation. The second comparison circuit is connected to the first input of a tenth AND control gate 76, the second input of which is connected to the above-mentioned control gate 66, its third input receiving the read markings intermediate-stored in the auxiliary memory bit positions 58 and its output, constituting the above mentioned restarting signal output 27, being connected to an erase OR gate 77. The OR gate 26 illustrated in Figure 2 receives a restarting signal from an activated control gate 76. The scanning device 50 receives from an activated erase OR gate 77 an erasure order pulse, due to which all address information intermediate-stored in the auxiliary memory is immediately cancelled.
There is obtained that an erasure of the auxilary memory and a restart of the execution of the second sequence is carried out if an intermediate-stored first category read instruction is encountered, where the associated address of this instruction agrees with the address associated with a write instruction which is selected during the execution of the first sequence. The erasure is required since the associated data processing unit has received data information from the main memory ,due to the encountered first category instruction, which the memory .would not- have received if attentiomhad Jbeen .consistently paid to the order determined in advance, I.e. If instructions in the first and second sequences had not been selected in parallel.
The intermediate storage unit includes a second NOR gate 78 with inputs connected to the above-mentioned identifying signal inputs 16 and 31 and which has its inverting output connected -to ihe-OR gate 51. AfiJ en the selections of the first and the second sequences are terminated, a scanning operation is con¬ sequently ordered, although the scanning is without result in the intermediate storage unit included in the data processing unit which has executed the first sequence. The output of the NOR gate 78 is connected to the first input of an eleventh AND control α,ate 79, the second .input of which .receives the write markings intermediate stored in the auxiliary memory bit positions 57, 'the output of gate 79 being connected to the above mentioned gate device 46 via a second control terminal 80. There is obtained that the scanning operation started by the NOR gate 78 results in that all write instructions intermediate- stored in the auxiliary memory are transferred, via the terminals 45, 47 and 48, the gate devices 46 and 38, and the bus system 3, to the main memory 1.
Finally, the intermediate storage unit according to Figure 3 has a twelfth AND control gate 81, with its inputs connected to the NOR gates 56 and 78 and its output connected to erase OR gate 77. There is obtained an erase order pulse when the scanning operation started by NOR gate 78 is terminated.
As an overall effect, there is achieved with the aid of the intermediate storage unit that the first and the second sequences are faultlessy, finally executed in the order determined in advance., in spite of the execution of both sequences having been initially ordered to be carried out in parallel.

Claims

C L A IMS
1. Method of executing two instruction sequences in an order determined in advance, the executions including selection of read instructions each containing its read address for retrieval of data information stored in one of a plurality of main memory locations accessible by their respective addresses, as well as 5 selection of write instructions each containing its write address and data information for transferring this data information to a main memory location accessible by this write address, and the data information used in conjunction with the execution of the sequence which is second due to the order not being guaranteed in advance independent of the data Information obtained in 10 conjunction with the execution of the sequence which is first due to the order, characterized by the steps a) executing the instructions in both sequences in parallel without regard to the order, b) preventing, during the execution of the first sequence, data information 15 transfer to the main memory locations, if this data information is obtained due to the execution of the second sequence write instructions, c) intermediate-storing a write address and data information,- obtained due to a write instruction selected during the execution of the second sequence,
. „.d_) -intermediate^stαring^an addressϊwhich is- obtained as÷a read addressid e to a "20 read instruction which Is selected during the execution of the second sequence, if this address has previously not been selected in conjunction with the execution of the second sequence write instructions, e) carrying out a first comparison between a read address selected during the execution of the second sequence and the intermediate-stored write address, 25 f) -preventing,- idue-tα likeness. etermined' ith the -aid^of- sai 'first -asomparison between a read address and a write address, data information retrieval from the main memory location which is accessible with the aid of this read address, and reading instead the intermediate-stored information associated with this write address, 30 g) carrying out a second comparison between the intermediate-stored __read address-and each of-the write -addresses obtained due'to the execution of the first sequence write instructions, h) restarting the executions of the second sequence if an address likeness is determined by said second comparison, and i) transferring when execution of the first sequence is terminated, the intermediate-stored data information with the aid of the likewise intermediate- stored write address to the main memory location accessible by this write address.
2. Apparatus for executing, according to claim 1, two instruction sequences in an order determined in advance, including a first starting circuit (13, 25, 28) for starting the execution of the sequence, which is first due to the order, and a first and second, respectively, instruction memory device (4) for storing the sequence which is first and second, respectively, due to the order and for selecting the stored instructions one at a time, said instruction memory units being connected via a bus system (3) to a main memory unit (1, 11) having its main memory locations accessible for data information reading/writing with the aid of read/write addresses selected from read/write address bit positions (19, 20, 21) in the instruction memory devices, in conjunction with the execution of the second s. juence, there being used data information which is not guaranteed in advance to be independent of the data information obtained in conjunction with executing the first sequence, characterized in that the apparatus includes a) a second starting circuit (14, 25, 26) for starting the second sequence, with its first activating input (14) connected to said first starting circuit, b) a first instruction transfer circuit (37, 38, 39, 41) which on its input side and output side is respectively connected to the second instruction memory device and bus system, and which is de-activated during execution of the first sequence due to the selection of a write instruction included in the second sequence, c) an auxiliary memory unit (49, 50) with its input terminals (33-36) connected to the second instruction memory device, and which includes data bit positions (72) and write address bit positions (57, 69) for intermediate-storing a write address and data information obtained due to a write instruction selected during the execution of the second sequence, and read address bit positions (58, 69) for intermediate-storing an address obtained as a read address due to a read instruction selected during the execution of the second sequence, if this address has previously not been selected in conjunction with the execution of the second sequence write instructions, d) a switching comparison circuit (68, 70, 73) the first comparison terminals of which are connected to the read address bit positions (19, 20) of the second instruction memory device, and the second comparison terminals of which .are connected to said write address' bit positions (57, 69) of the auxiliary memory unit, the output of said switching comparison circuit sending a first and a second logical switching signal state respectively due to difference and likeness between the addresses received on its switching comparison terminals. e) a data switch (37, 38, 39, 41, 44) with its first data input connected to the bus system and its second data input connected to the data bit positions (72) of the auxiliary memory unit and with its control inputs connected to the output of the switching comparison circuit, for retrieving, due to said first switching signal state, data information from the main memory location which is accessible with the aid of the read address received by the switching comparison circuit, and for retrieving, due to said second switching signal state, from the. auxiliary memory
Figure imgf000018_0001
information associated with the -write address received by the switching comparison circuit, f) comparison criterion circuit {66, 74, 76) having first comparison terminals (16, 75) for receiving the write addresses selected during the execution of the first sequence, second comparison terminals connected to said read address bit positions (58, 69) of the auxiliary memory unit and its output (27) connected to a second activating input of said second starting circuit, which output is activated on agreement between the addresses received on the comparison - criterion terminals- and g) a second instruction transfer circuit (38, 46, 78, 79) which on its input side and output side is respectively connected to the data bit positions (72) and write address bit positions (57, 69) of the auxiliary memory unit and to the bus system (3), and which is activated when the execution of the first sequence is terminated.
PCT/SE1987/000437 1986-10-03 1987-09-28 Method and device to execute two instruction sequences in an order determined in advance WO1988002513A1 (en)

Priority Applications (7)

Application Number Priority Date Filing Date Title
KR1019880700625A KR920006769B1 (en) 1986-10-03 1987-09-28 Method and device to execute two instruction sequences in an order determined in advance
AT87906632T ATE73940T1 (en) 1986-10-03 1987-09-28 METHOD OF EXECUTING TWO SEQUENCES OF COMMANDS IN A PRE-DETERMINED ORDER.
BR8707473A BR8707473A (en) 1986-10-03 1987-09-28 PROCESS AND APPARATUS TO PERFORM TWO INSTRUCTION SEQUENCES IN A PARTICULAR ORDER OF ANTEMAO
DE8787906632T DE3777632D1 (en) 1986-10-03 1987-09-28 METHOD FOR EXECUTING TWO ORDERS IN A PRE-DETERMINED ORDER.
FI882468A FI93907C (en) 1986-10-03 1988-05-25 A method and apparatus for executing two instruction sequences in a predetermined order
NO882413A NO173207C (en) 1986-10-03 1988-06-01 PROCEDURE AND DEVICE FOR PERFORMING TWO INSTRUCTION SEQUENCES IN A PRESENT ORDER PREFERRED
DK300788A DK168135B1 (en) 1986-10-03 1988-06-02 Method for executing two instruction sequences in a predetermined order

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
SE8604223-1 1986-10-03
SE8604223A SE454921B (en) 1986-10-03 1986-10-03 SET AND DEVICE FOR EXECUTIVE ORDER IN ORDER TO EXECUTE TWO INSTRUCTION SEQUENCES

Publications (1)

Publication Number Publication Date
WO1988002513A1 true WO1988002513A1 (en) 1988-04-07

Family

ID=20365824

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/SE1987/000437 WO1988002513A1 (en) 1986-10-03 1987-09-28 Method and device to execute two instruction sequences in an order determined in advance

Country Status (20)

Country Link
US (1) US4956770A (en)
EP (1) EP0285634B1 (en)
JP (1) JPH01500935A (en)
KR (1) KR920006769B1 (en)
CN (2) CN87106625A (en)
AU (1) AU596234B2 (en)
BR (1) BR8707473A (en)
CA (1) CA1289670C (en)
DE (1) DE3777632D1 (en)
DK (1) DK168135B1 (en)
ES (1) ES2005370A6 (en)
FI (1) FI93907C (en)
GR (1) GR871511B (en)
IE (2) IE61307B1 (en)
MA (1) MA21073A1 (en)
MX (1) MX159991A (en)
PT (1) PT85811B (en)
SE (1) SE454921B (en)
TN (1) TNSN87108A1 (en)
WO (1) WO1988002513A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6662203B1 (en) 1998-11-16 2003-12-09 Telefonaktiebolaget Lm Ericsson (Publ) Batch-wise handling of signals in a processing system
US6714961B1 (en) 1998-11-16 2004-03-30 Telefonaktiebolaget Lm Ericsson (Publ) Multiple job signals per processing unit in a multiprocessing system

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5075844A (en) * 1989-05-24 1991-12-24 Tandem Computers Incorporated Paired instruction processor precise exception handling mechanism
US5163139A (en) * 1990-08-29 1992-11-10 Hitachi America, Ltd. Instruction preprocessor for conditionally combining short memory instructions into virtual long instructions
US5420990A (en) * 1993-06-17 1995-05-30 Digital Equipment Corporation Mechanism for enforcing the correct order of instruction execution
JPH07334372A (en) * 1993-12-24 1995-12-22 Seiko Epson Corp System and method for emulation
US8738822B2 (en) * 2005-05-03 2014-05-27 Flexera Software Llc System and method for controlling operation of a component on a computer system
JP5350677B2 (en) * 2008-05-19 2013-11-27 株式会社東芝 Bus signal control circuit and signal processing circuit having bus signal control circuit
JP2021015384A (en) * 2019-07-10 2021-02-12 富士通株式会社 Information processing circuit, information processing apparatus, information processing method and information processing program

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB1218656A (en) * 1968-03-27 1971-01-06 Int Computers Ltd Improvements in or relating to computer system
GB1441458A (en) * 1972-06-28 1976-06-30 Texas Instruments Inc Stored programme data processing for parallel processing of programme segment
US3969702A (en) * 1973-07-10 1976-07-13 Honeywell Information Systems, Inc. Electronic computer with independent functional networks for simultaneously carrying out different operations on the same data
US4626989A (en) * 1982-08-16 1986-12-02 Hitachi, Ltd. Data processor with parallel-operating operation units

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3787673A (en) * 1972-04-28 1974-01-22 Texas Instruments Inc Pipelined high speed arithmetic unit
SE378690B (en) * 1973-12-13 1975-09-08 Ellemtel Utvecklings Ab
SE387763B (en) * 1975-10-23 1976-09-13 Ellemtel Utvecklings Ab COMPUTER AT A COMPUTER MEMORY TO POSSIBLE A SUCCESSIVE TRANSFER DURING OPERATION OF AN AVAILABLE MEMORY FIELD
JPS57162165A (en) * 1981-03-30 1982-10-05 Fanuc Ltd Re-editing system for storage area
US4466061A (en) * 1982-06-08 1984-08-14 Burroughs Corporation Concurrent processing elements for using dependency free code
JPS60146350A (en) * 1984-01-11 1985-08-02 Hitachi Ltd Communication controller
US4720779A (en) * 1984-06-28 1988-01-19 Burroughs Corporation Stored logic program scanner for a data processor having internal plural data and instruction streams
SE454920B (en) * 1986-10-03 1988-06-06 Ellemtel Utvecklings Ab SETTING AND DEVICE FOR EXECUTIVE ORDER IN ORDER TO EXECUTE TWO INSTRUCTION SEQUENCES BY MEANS OF Separate memory

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB1218656A (en) * 1968-03-27 1971-01-06 Int Computers Ltd Improvements in or relating to computer system
GB1441458A (en) * 1972-06-28 1976-06-30 Texas Instruments Inc Stored programme data processing for parallel processing of programme segment
US3969702A (en) * 1973-07-10 1976-07-13 Honeywell Information Systems, Inc. Electronic computer with independent functional networks for simultaneously carrying out different operations on the same data
US4626989A (en) * 1982-08-16 1986-12-02 Hitachi, Ltd. Data processor with parallel-operating operation units

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6662203B1 (en) 1998-11-16 2003-12-09 Telefonaktiebolaget Lm Ericsson (Publ) Batch-wise handling of signals in a processing system
US6714961B1 (en) 1998-11-16 2004-03-30 Telefonaktiebolaget Lm Ericsson (Publ) Multiple job signals per processing unit in a multiprocessing system

Also Published As

Publication number Publication date
MA21073A1 (en) 1988-07-01
AU8036587A (en) 1988-04-21
SE8604223D0 (en) 1986-10-03
JPH01500935A (en) 1989-03-30
FI882468A0 (en) 1988-05-25
SE454921B (en) 1988-06-06
IE872505L (en) 1988-04-03
FI93907C (en) 1995-06-12
US4956770A (en) 1990-09-11
DE3777632D1 (en) 1992-04-23
GR871511B (en) 1987-10-01
CA1289670C (en) 1991-09-24
FI882468A (en) 1988-05-25
IE61307B1 (en) 1994-10-19
SE8604223L (en) 1988-04-04
CN87106765A (en) 1988-04-20
BR8707473A (en) 1988-09-13
KR920006769B1 (en) 1992-08-17
IE61306B1 (en) 1994-10-19
PT85811A (en) 1988-11-30
DK168135B1 (en) 1994-02-14
PT85811B (en) 1993-08-31
MX159991A (en) 1989-10-23
AU596234B2 (en) 1990-04-26
IE872506L (en) 1988-04-03
FI93907B (en) 1995-02-28
DK300788A (en) 1988-06-02
EP0285634A1 (en) 1988-10-12
ES2005370A6 (en) 1989-03-01
KR880701913A (en) 1988-11-07
EP0285634B1 (en) 1992-03-18
TNSN87108A1 (en) 1990-01-01
DK300788D0 (en) 1988-06-02
CN87106625A (en) 1988-04-13

Similar Documents

Publication Publication Date Title
EP0386935B1 (en) Apparatus capable of varying number of wait states for access
JP2539199B2 (en) Digital processor controller
WO1987005417A1 (en) Instruction prefetch control apparatus
KR100335785B1 (en) Execution of data processing instructions
US4755936A (en) Apparatus and method for providing a cache memory unit with a write operation utilizing two system clock cycles
US4348722A (en) Bus error recognition for microprogrammed data processor
US3740722A (en) Digital computer
US5371857A (en) Input/output interruption control system for a virtual machine
WO1988002513A1 (en) Method and device to execute two instruction sequences in an order determined in advance
EP0247604B1 (en) Apparatus and method for interprocessor communication
US5594880A (en) System for executing a plurality of tasks within an instruction in different orders depending upon a conditional value
AU589047B2 (en) Method and device to execute two instruction sequences in an order determined in advance
JP2003515805A (en) Processor system
US4800490A (en) Buffer storage control system having a priority circuit
US6209001B1 (en) Back-up system capable of performing back-up operation at optional time
JP2594600B2 (en) Single chip microcomputer
US3761893A (en) Digital computer
US6085297A (en) Single-chip memory system including buffer
JPS5819094B2 (en) Priority vector interrupt device
JPH0756633B2 (en) Task switching method
EP0559220A2 (en) Method and apparatus for storing DOS in high memory area
EP0098170B1 (en) Access control processing system in computer system
JP3144979B2 (en) Program processing device and processing method
JP2639988B2 (en) Data driven data processor
JPS6146545A (en) Input and output instruction control system

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AU BR DK FI HU JP KR NO SU US

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE FR GB IT LU NL SE

WWE Wipo information: entry into national phase

Ref document number: 1987906632

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 882468

Country of ref document: FI

WWP Wipo information: published in national office

Ref document number: 1987906632

Country of ref document: EP

WWG Wipo information: grant in national office

Ref document number: 1987906632

Country of ref document: EP

WWG Wipo information: grant in national office

Ref document number: 882468

Country of ref document: FI