专利快速检索-快速检索全球专利，免费商用专利数据库-IPRDB

1. 发明专利

IL284681D0 Directed interrupt for multilevel virtualization 未知
公开(公告)号：IL284681D0
公开(公告)日：2021-08-31
申请号：IL28468121
申请日：2021-07-07
申请人： IBM , MARCO KRAEMER , CHRISTOPH RAISCH , BERND NERZ , DONALD W SCHMIDT , MATTHIAS KLEIN , SASCHA JUNGHANS , PETER DANA DRIEVER
发明人： MARCO KRAEMER , CHRISTOPH RAISCH , BERND NERZ , DONALD W SCHMIDT , MATTHIAS KLEIN , SASCHA JUNGHANS , PETER DANA DRIEVER
IPC分类号： G06F9/455 , G06F9/48 , G06F9/54 , G06F13/24
摘要： An interrupt signal is provided to a first guest operating system. A bus attachment device receives an interrupt signal from a bus connected module with an interrupt target ID identifying a processor assigned for use by the guest operating system as a target processor for handling the interrupt signal. The bus attachment device checks whether the target processor is scheduled for usage by the guest operating system. If the target processor is not scheduled for usage, the bus attachment device forwards the interrupt signal using broadcasting and updates a forwarding vector entry stored in a memory section assigned to a second guest operating system hosting the first guest operating system. The update is used for indicating to the first operating system that there is a first interrupt signal addressed to the interrupt target ID to be handled.

2. 发明专利

IL282055D0 Sort and merge instruction for a general-purpose processor 未知
公开(公告)号：IL282055D0
公开(公告)日：2021-05-31
申请号：IL28205521
申请日：2021-04-04
申请人： IBM , BRUCE C GIAMEI , MARTIN RECKTENWALD , DONALD W SCHMIDT , TIMOTHY SLEGEL , ADITYA N PURANIK , MARK S FARRELL , CHRISTIAN JACOBI , JONATHAN D BRADBURY , CHRISTIAN ZOELLIN
发明人： BRUCE C GIAMEI , MARTIN RECKTENWALD , DONALD W SCHMIDT , TIMOTHY SLEGEL , ADITYA N PURANIK , MARK S FARRELL , CHRISTIAN JACOBI , JONATHAN D BRADBURY , CHRISTIAN ZOELLIN
IPC分类号： G06F9/30 , G06F16/242
摘要： A Sort Lists instruction is provided to perform a sort and/or a merge operation. The instruction is an architected machine instruction of an instruction set architecture and is executed by a general-purpose processor of the computing environment. The executing includes sorting a plurality of input lists to obtain one or more sorted output lists, which are output.

3. 发明专利

IL281734D0 Saving and restoring machine state between multiple executions of an instruction 未知
公开(公告)号：IL281734D0
公开(公告)日：2021-05-31
申请号：IL28173421
申请日：2021-03-22
申请人： IBM , BRUCE CONRAD GIAMEI , MARTIN RECKTENWALD , DONALD W SCHMIDT , TIMOTHY SLEGEL , ADITYA N PURANIK , MARK S FARRELL , CHRISTIAN JACOBI , JONATHAN D BRADBURY , CHRISTIAN ZOELLIN
发明人： BRUCE CONRAD GIAMEI , MARTIN RECKTENWALD , DONALD W SCHMIDT , TIMOTHY SLEGEL , ADITYA N PURANIK , MARK S FARRELL , CHRISTIAN JACOBI , JONATHAN D BRADBURY , CHRISTIAN ZOELLIN
IPC分类号： G06F9/30 , G06F9/38
摘要： Saving and restoring machine state between multiple executions of an instruction. A determination is made that processing of an operation of an instruction executing on a processor has been interrupted prior to completion. Based on determining that the processing of the operation has been interrupted, current metadata of the processor is extracted. The metadata is stored in a location associated with the instruction and used to re-execute the instruction to resume forward processing of the instruction from where it was interrupted.

4. 发明专利

JPH11353259A METHOD AND DEVICE FOR TREATING COMMUNICATION REQUEST 失效
公开(公告)号：JPH11353259A
公开(公告)日：1999-12-24
申请号：JP7768199
申请日：1999-03-23
申请人： IBM
发明人： DAVID B EMES , DONALD W SCHMIDT
IPC分类号： G06F13/00 , G06F13/14
摘要： PROBLEM TO BE SOLVED: To obtain performance close to a 'streaming' level even in the presence of only small interaction traffic of high frequency through an adapter by writing blocks to an I/O device at specific time and dynamically adjusting a block-processing factor according to the quantity of requests. SOLUTION: In constitution 100, a 1st computer system 102 (local system) communicates with a remote system 104 over a communication channel 106. A requester 114 sends a communication request to a communication stack 118 of an OS kernel 110. Then sent communication requests are stored into a block 120 and a block-processing factor representing specific size obtained before the block 120 is written to the I/O device is generated. Here, when the I/O device obtains the specific size, the block 120 is written to the I/O device and the block-processing factor is dynamically adjusted according to the quantity of the requests.

5. 发明专利

JP2008047126A Affinity dispatching load balancer with precise cpu consumption data 有权
标题翻译：具有精确CPU消耗数据的AFFINITY DISPATCHING LOAD BALANCER
公开(公告)号：JP2008047126A
公开(公告)日：2008-02-28
申请号：JP2007211128
申请日：2007-08-13
申请人： Internatl Business Mach Corp , インターナショナル・ビジネス・マシーンズ・コーポレーションInternational Business Maschines Corporation
发明人： DILLENBERGER DONNA N , DYCK GREG A , HEISIG STEPHEN J , PIERCE BERNARD R , DONALD W SCHMIDT , SU GONG
IPC分类号： G06F9/50
CPC分类号： G06F9/505 , G06F2209/5022
摘要： PROBLEM TO BE SOLVED: To provide a processor load balancing method to overcome drawbacks of the prior art.
SOLUTION: A computer-implemented method for distributing a plurality of tasks to a plurality of processing nodes in a processor network includes: a step of calculating a task process consumption value with respect to the tasks; a step of calculating a measured node processor consumption value for the nodes; a step of calculating a target node processor consumption value for the nodes, the target node processor consumption value indicating optimal node processor consumption; a step of calculating a load index value as a difference between the calculated node processor consumption value for a node i and the target node processor consumption value for the node i; and a step of distributing the tasks among the nodes so as to balance a processor workload among the nodes, according to the calculated load index value, such that the calculated load index value of each node is substantially zero. Further, the method comprises a multidimensional balancing matrix with each dimension thereof representing a node corresponding to a different processor type and each cell representing tasks assigned to multiple nodes.
COPYRIGHT: (C)2008,JPO&INPIT
摘要翻译：要解决的问题：提供一种处理器负载平衡方法来克服现有技术的缺点。解决方案：一种用于将多个任务分配到处理器网络中的多个处理节点的计算机实现的方法包括：计算相对于所述任务的任务处理消耗值的步骤; 计算节点的测量节点处理器消耗值的步骤; 计算节点的目标节点处理器消耗值的步骤，指示最佳节点处理器消耗的目标节点处理器消耗值; 计算负载指标值作为节点i的计算节点处理器消耗值与节点i的目标节点处理器消耗值之间的差的步骤; 以及根据所计算的负载指数值，在所述节点之间分配所述任务以平衡所述节点之间的处理器工作负荷的步骤，使得所计算的每个节点的负载指数值基本上为零。此外，该方法包括多维平衡矩阵，其每个维度表示对应于不同处理器类型的节点，并且每个小区表示分配给多个节点的任务。版权所有（C）2008，JPO＆INPIT

6. 发明专利

JPH11331265A METHOD AND DEVICE FOR USING INPUT/OUTPUT BUFFER 失效
公开(公告)号：JPH11331265A
公开(公告)日：1999-11-30
申请号：JP7763499
申请日：1999-03-23
申请人： IBM
发明人： DAVID B EMES , DONALD W SCHMIDT
IPC分类号： G06F13/00 , H04L1/16 , H04L12/56
摘要： PROBLEM TO BE SOLVED: To provide the method and the device to use an input/output(I/O) buffer for a re-transmission means in a client/server system. SOLUTION: When it is discriminated that I/O buffers (112, 114) are not to be used as a re-transmission means based on factors such as a packet size and a round trip time(RTT), data are copied to a transmission buffer 124 maintained by the system with respect to a specific requester. An RTT threshold level, a critical threshold level and a tight buffer threshold level are set, a connection with a longer RTT than the set value or a connection conducted when number of remaining I/O buffers is lower than the critical threshold level is not allowed to reserve the I/O buffers as the re-transmission means. In the case that number of the I/O buffers having already been assigned exceeds the tight buffer threshold level during a critical stabilizing interval that is started when number of remaining I/O buffers is lower than the critical threshold level, the I/O buffer cannot be used for the re-transmission means.

你已经成功收藏专利！

检索式保存成功!

IPRDB

热门服务

关于我们

友情链接

联系方式