Management system and method for parallel computer system

Claim 2

A management system having a control terminal for performing maintenance and management of a parallel computer system with a plurality of computers constituting nodes being connected together, each of the plurality of nodes comprising: a main processor operated with a main power source of the node for performing parallel processing; a LAN control unit for managing application software of the node under control of said control terminal; and a system control unit for managing hardware of the node, said system control unit and said LAN control unit being connected to said control terminal via separate, respective networks, said system control unit comprising: a sub-processor operated with an auxiliary power source of the node for executing a system control command for controlling the main processor; and a system control adapter operated for communications with the control terminal for the system control and for controlling the sub-processor, said system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective hardware of the node and to supply the generated control data to said hardware; said control terminal comprising a system control adapter for communication with a plurality of system control adapters of the plurality of nodes; and said nodes comprising a system control interface for connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal and for transmitting the system control command for the maintenance and management of a plurality of main processors of the plurality of nodes, from the control terminal to a plurality of system control adapters of the plurality of nodes.

Claim 3

A management system for a parallel computer system according to claim 1, wherein each sub-processor of the plurality of nodes has a function of turning on or off the main power source of each node, and the control terminal comprises means for transmitting a system control command to the system control adapters of the plurality of nodes to collectively or singularly turn on or off the main power sources of the sub-processors.

Claim 4

A management system for a parallel computer system according to claim 2, wherein the control terminal comprises means for singularly transmitting a system control command to each of the system control adapters of the plurality of nodes at a preset time interval to singularly turn on each main power source of the plurality of nodes.

Claim 5

A management system for a parallel computer system according to claim 1, wherein the control terminal comprises abnormal state detecting means for judging, if a normal response to a specific system control command transmitted to the system control adapter of a specific node among the plurality of nodes is not received in a preset time duration, that an abnormal state is occurring at the specific node.

Claim 6

A management system for a parallel computer system according to claim 1, wherein each system control adapter of the plurality of nodes comprises means for storing a node message output from the main processor or sub-processor of the node, and the control terminal comprises means for reading the node message stored at the system control adapter of the node.

Claim 7

A management system for a parallel computer system according to claim 1, wherein each sub-processor of the plurality of nodes comprises means for referring to or changing the contents of the main memory or registers of the node, and the control terminal comprises means for transmitting a system control command to each system control adapter of the plurality of nodes to refer to or change the contents of the main memory or registers of the node.

Claim 8

A management system for a parallel computer system according to claim 1, wherein each sub-processor of the plurality of nodes comprises means for resetting the main processor of the node, and the control terminal comprises means for transmitting a system control command to the system control adapter of the node to reset the main processor of the node.

Claim 9

A management system for a parallel computer system according to claim 1, wherein each sub-processor of the plurality of nodes comprises means for referring to or changing the contents of the main memory or registers of the node and means for resetting the main processor of the node, and the control terminal comprises means for transmitting a system control command for referring to or change the name of a bootstrap device and a system command for resetting the main processor of the node.

Claim 10

A management system for a parallel computer system according to claim 1, wherein a plurality of control terminals are prepared and the management system comprises means for limiting the functions of some control terminals among the plurality of control terminals.

Claim 11

A management system for a parallel computer system according to claim 1, wherein the control terminal comprises a power turn-on logic unit operated with the auxiliary power source for turning on the main power source of the control terminal upon input of a specific signal and means for transmitting a system control command to the system control adapters of the plurality of nodes to collectively or singularly turn on the main power source of each node after the main power source of the control terminal is turned on by the power turn-on logic unit.

Claim 12

A management system for a parallel computer system according to claim 1, wherein said interpreting means has a function of changing a transmission rate of said control data within the node independently of the transmission rate in the system control interface.

Claim 13

A management system for parallel computer system according to claim 11, wherein said interpreting means includes a function of controlling transmission of a message from said main processor based on interpretation of the system control command.

Claim 16

A method of configuring the management system according to claim 13, wherein the system control command to be transmitted from the control terminal includes a system control command to be transmitted to the system control adapters of the plurality of nodes to collectively or singularly turn on or off the main power sources of the plurality of nodes.

Claim 17

A method of configuring a management system according to claim 14, wherein the control terminal singularly transmits a system control command to each of the system control adapters of the plurality of nodes at a preset time interval to singularly turn on each main power source of the plurality of nodes.

Claim 18

A method of configuring a management system according to claim 13, wherein the control terminal judges, if a normal response to a specific system control command transmitted to the system control adapter of a specific node among the plurality of nodes is not received in a preset time duration, that an abnormal state is occurring at the specific node.

Claim 19

A method of configuring a management system according to claim 13, wherein each system control adapter of the plurality of nodes stores a node message output from the main processor or sub-processor of the node, and the control terminal reads the node message stored at the system control adapter of the node.

Claim 20

A method of configuring a management system according to claim 13, wherein each sub-processor of the plurality of nodes refers to or changes the contents of the main memory or registers of the node, and the control terminal transmits a system control command to each system control adapter of the plurality of nodes to refer to or change the contents of the main memory or registers of the node.

Claim 21

A method of configuring a management system according to claim 13, wherein each sub-processor of the plurality of nodes resets the main processor of the node, and the control terminal transmits a system control command to the system control adapter of the node to reset the main processor of the node.

Claim 22

A method of configuring a management system according to claim 13, wherein each sub-processor of the plurality of nodes refers to or changes the contents of the main memory or registers of the node and resets the main processor of the node, and the control terminal transmits a system control command for referring to or change the name of a bootstrap device and a system command for resetting the main processor of the node.

Claim 23

A method of configuring a management system according to claim 13, wherein a plurality of control terminals are prepared and the management system limits the functions of some control terminals among the plurality of control terminals.

Claim 24

A method of configuring a management system according to claim 13, wherein the control terminal comprises a power turn-on logic unit operated with the auxiliary power source for turning on the main power source of the control terminal upon input of a specific signal and transmits a system control command to the system control adapters of the plurality of nodes to collectively or singularly turn on the main power source of each node after the main power source of the control terminal is turned on by the power turn-on logic unit.

Claim 25

A management system for managing a parallel computer system having a plurality of computers constituting nodes, system control interface for configuring the parallel computer system by connecting the nodes together, and a control terminal for performing maintenance and management of the parallel computer system, each of the plurality of nodes comprising: a main processor operated with a main power source of the node for performing parallel processing; a LAN control unit for managing application software of the node under control of said control terminal; and a system control unit for controlling the hardware of the node, said system control unit and said LAN control unit being connected to said control terminal via separate, respective networks, said system control unit including: a sub-processor operated with an auxiliary power source of the node for executing a system control command for controlling the main processor, and a system control adapter operated for communications with the control terminal for the hardware control and for controlling the sub-processor, said system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective hardware of the node and supply the generated control data to said hardware; said control terminal comprising a system control adaptor for communication with a plurality of system control adapters of the plurality of nodes; and said system control interface connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal, and said control terminal transmitting the system control command for the maintenance and management of a plurality of main processors of the plurality of nodes, to a plurality of system control adapters of the plurality of nodes under the control of the node system control adapters.

Claim 26

A management system for managing a parallel computer system having a plurality of computers constituting nodes, system control interface for configuring the parallel computer system by connecting the nodes together, and a control terminal for performing maintenance and management of the parallel computer system, each of the plurality of nodes comprising: a main processor operated with a main power source of the node for performing parallel processing; a LAN control unit for managing application software of the node under control of said control terminal; and a system control unit for controlling the hardware of the node, said system control unit and said LAN control unit being connected to said control terminal via separate, respective networks, said system control unit including: a sub-processor operated with an auxiliary power source of the node for executing a system control command for controlling the main processor, and a system control adapter operated for communications with the control terminal for the hardware control and for controlling the sub-processor, said system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective hardware of the node and supply the generated control data to said hardware; said control terminal comprising a system control adaptor for communication with a plurality of system control adapters of the plurality of nodes; and said system control interface connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal, and said control terminal transmitting the system control command for the maintenance and management of a plurality of main processors of the plurality of nodes, to a plurality of system control adapters of the plurality of nodes under the control of the node system control adapters.

Claim 27

A node used for a management system for a parallel computer system having a plurality of computers constituting nodes, a system control interface for configuring the parallel computer system by connecting the nodes together, and a control terminal for the maintenance and management of the parallel computer system, said node comprising: a main processor operated with a main power source of the node for performing parallel processing, a LAN control unit for managing application software of the node under control of said control terminal, and a system control unit for managing the hardware of the node, said system control unit and said LAN control unit being connected to said control terminal via separate, respective networks, said system control unit including: a sub-processor operated with an auxiliary power source of the node for executing a system control command for controlling the main processor, and an intelligent system control adapter operated for communications with the control terminal for the hardware management and for controlling the sub-processor, said intelligent system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective hardware of the node and supply the generated control data to said hardware; said intelligent system control adapter connected via the system control interface connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal, receives the system control command for the maintenance and management of a plurality of main processors from the control terminal, analyzes the system control command to generate control data, and transmits the system control command to the intelligent system control adapter.

Claim 28

A node used for a management system for a parallel computer system having a plurality of computers constituting nodes, a system control interface for configuring the parallel computer system by connecting the nodes together, and a control terminal for the maintenance and management of the parallel computer system, said node comprising: a main processor operated with a main power source of the node for performing parallel processing, a LAN control unit for managing application software of the node under control of said control terminal, and a system control unit for managing the hardware of the node, said system control unit and said LAN control unit being connected to said control terminal via separate, respective networks, said system control unit including: a sub-processor operated with an auxiliary power source of the node for executing a system control command for controlling the main processor, and an intelligent system control adapter operated for communications with the control terminal for the hardware management and for controlling the sub-processor, said intelligent system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective hardware of the node and supply the generated control data to said hardware; said intelligent system control adapter connected via the system control interface connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal, receives the system control command for the maintenance and management of a plurality of main processors from the control terminal, analyzes the system control command to generate control data, and transmits the system control command to the intelligent system control adapter.

Claim 29

A management system having a control terminal for performing maintenance and management of a parallel computer system with a plurality of computers constituting nodes being connected together, each of the plurality of nodes comprising: a LAN control unit for managing application software for the node under control by said terminal controller; a main processor operated with a main power source of the node for performing parallel processing; and a system control unit for managing hardware of the node, said system control unit and said LAN control unit being connected to said terminal controller via separate, respective networks, said system control unit comprising: a sub-processor executing a system control command from said control terminal to control at least start and reset operations of the main processor; and a system control adapter operated with the control terminal for the system control and for controlling the sub-processor, said system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective resources of the node and supply the generated control data to said resources and changing a transmission rate of said control data within the node independently of that in the system control interface; said control terminal comprising a system control adapter for communication with a plurality of system control adapters of the plurality of nodes; and said node comprising a system control interface for connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal and for transmitting the system control command for the maintenance and management of a plurality of main processors of the plurality of nodes, from the control terminal to a plurality of system control adapters of the plurality of nodes.

Claim 30

A management system having a control terminal for performing maintenance and management of a parallel computer system with a plurality of computers constituting nodes being connected together, each of the plurality of nodes comprising: a LAN control unit for managing application software for the node under control by said terminal controller; a main processor operated with a main power source of the node for performing parallel processing; and a system control unit for managing hardware of the node, said system control unit and said LAN control unit being connected to said terminal controller via separate, respective networks, said system control unit comprising: a sub-processor executing a system control command from said control terminal to control at least start and reset operations of the main processor; and a system control adapter operated with the control terminal for the system control and for controlling the sub-processor, said system control adapter including means for interpreting the control command from the control terminal to generate control data adapted for respective resources of the node and supply the generated control data to said resources and changing a transmission rate of said control data within the node independently of that in the system control interface; said control terminal comprising a system control adapter for communication with a plurality of system control adapters of the plurality of nodes; and said node comprising a system control interface for connecting the plurality of system control adapters of the plurality of nodes to the system control adapter of the control terminal and for transmitting the system control command for the maintenance and management of a plurality of main processors of the plurality of nodes, from the control terminal to a plurality of system control adapters of the plurality of nodes.