METHOD FOR ORGANIZATION OF MULTI-PROCESSOR COMPUTER Russian patent published in 2007 - IPC G06F15/16 G06F15/173 

Abstract RU 2312388 C2

FIELD: computer engineering, possible use for creating multi-processor multi-thread computers.

SUBSTANCE: method for organization of multi-processor computer includes parallel execution of a thread of computations by means of distributed representation of thread descriptor stored in virtual memory, execution of primary selection of architecture commands by means of thread monitors, generation of graph for information dependencies of transactions, which are serially outputted through network into execution clusters, active thread is transferred to resident queue of transactions awaiting completion and next active thread is selected, by sequencers of execution clusters transactions are received and their commands and aforementioned graph are copied to registry file of cluster, execution-ready commands are copied to priority-ordered secondary selection queues, aforementioned selection and transfer of complete commands to the cluster are performed, graph is corrected based on these, on basis of correction results, the finalized command is added to either secondary selection queue or transaction completion result is transferred to monitor, thread is moved to queue for completed threads with correction of thread descriptor representation root, where completed thread is removed from waiting queue, and completion reason is outputted as a result available for software analysis.

EFFECT: fully hardware-based realization of multi-program control over threads with priority-based exclusion with precision up to an individual command.

Similar patents RU2312388C2

Title Year Author Number
METHOD AND DEVICE FOR PAUSING EXECUTION OF A STREAM UNTIL A CERTAIN MEMORY ACCESS IS PERFORMED 2002
  • Marr Debora
  • Rodzhers Skott
  • Khill Dehvid
  • Kaushik Shivnandan
  • Krosslehnd Dzhejms
  • Kaufehjti Dehvid
RU2308754C2
INSTRUCTION AND LOGIC FOR MEMORY ACCESS IN A CLUSTERED WIDE-EXECUTION MACHINE 2013
  • Leshenko Anton U.
  • Efimov Andrej
  • Shishlov Sergej I.
  • Ier Dzhajesh
  • Babayan Boris A.
RU2662394C2
INSTRUCTION AND LOGICAL SCHEME FOR SORTING AND LOADING OF SAVE INSTRUCTIONS 2014
  • Lechenko, Anton
  • Efimov, Andrey
  • Shishlov, Sergey Y
  • Kluchnikov, Andrey
  • Garifullin, Kamil
  • Burovenko, Igor
  • Babayan, Boris A.
RU2663362C1
UNLIMITED TRANSACTIONAL MEMORY WITH ASSURANCES OF MOVEMENT DURING TRANSFER, USING HARDWARE GLOBAL LOCK 2014
  • Gottshlikh Dzhastin E.
  • Kalchu Irina
  • Shpejsmen Tatyana
  • Pokam Zhil A.
RU2597506C2
USE OF AUTHENTICATED MANIFESTS TO ENSURE EXTERNAL CERTIFICATION OF MULTIPROCESSOR PLATFORMS 2014
  • Skarlata Vinsent R.
  • Dzhonson Sajmon P.
  • Beker Vladimir
  • Uolker Dzhess
  • Sentoni Emi L.
  • Anati Ittaj
  • Makaram Ragkhunandan
  • Makkin Frensis Ks.
  • Rozas Karlos V.
  • Savagaonkar Udaj R.
RU2599340C2
METHOD AND SYSTEM FOR CONTROLLING EXECUTION WITHIN COMPUTING ENVIRONMENT 2012
  • Dan F. Grejner
  • Timoti Dzh. Slidzhl
  • Kristian Jakobi
  • Piter Dzheremi Relson
  • Randall Uilljam Filli
RU2577487C2
FORCED TRANSACTION EXECUTION 2012
  • Dan F. Grejner
  • Timoti Dzh. Slidzhl
  • Kristian Jakobi
RU2549112C2
METHOD OF FUNCTIONING COMPUTER DEVICE OPERATING SYSTEM OF SOFTWARE AND HARDWARE COMPLEX 2016
  • Molyakov Andrej Sergeevich
RU2626350C1
PROGRAMME INTERRUPTION FILTERING IN TRANSACTIONAL EXECUTION 2012
  • Dan F. Grejner
  • Kristian Jakobi
  • Timoti Dzh. Slidzhl
  • Marsel' Mitran
RU2568923C2
CONVERTING MESSAGE-INITIATED INTERRUPTION INTO INPUT/OUTPUT ADAPTER-GENERATED EVENT NOTIFICATION 2010
  • Gustav Sittmann Iii
  • Dejvid Kraddok
  • Tomas Gregg
  • Mark Farrell
  • Dzhanet Iston
  • Ehrik Norman Lejs
RU2546561C2

RU 2 312 388 C2

Authors

Efimov Andrej Igorevich

Dates

2007-12-10Published

2005-09-22Filed