公开(公告)号：GB2381092A

公开(公告)日：2003-04-23

申请号：GB0125314

申请日：2001-10-19

Applicant: IBM

Inventor： MCGUIRE RICHARD KENNETH ,  SLATTERY EDWARD JOHN ,  WEBSTER MATTHEW ALEXANDER

IPC: G06F9/445 ,  G06F9/46

Abstract: A system is provided which comprises multiple Java virtual machines (VMs) . Each VM is capable of running a plurality of threads. The system includes at least one object which is shared between the VMs so that it is accessible from two or more VMs, and at least one non-shared object which is accessible only from a single VM. In order to perform a lock operation on a shared object, it is first detecting that an object is a shared object. In a preferred embodiment, this is accomplished by locating a predefined value in a particular lock data word in the object. In response to the detection, a monitor is used to control local access to the shared object for the thread that is requesting the object. The monitor is on the same VM as the requesting thread. In addition, a data structure, which may be a proxy object, is updated in order to associate the monitor with the shared object.

公开(公告)号：GB2381092B

公开(公告)日：2005-10-19

申请号：GB0125314

申请日：2001-10-19

Applicant: IBM

Inventor： MCGUIRE RICHARD KENNETH ,  SLATTERY EDWARD JOHN ,  WEBSTER MATTHEW ALEXANDER

IPC: G06F9/445 ,  G06F9/46

Abstract: A system is provided which comprises multiple Java virtual machines (VMs). Each VM is capable of running a plurality of threads. The system includes at least one object which is shared between the VMs so that it is accessible from two or more VMs, and at least one non-shared object which is accessible only from a single VM. In order to perform a lock operation on a shared object, it is first detecting that an object is a shared object. In a preferred embodiment, this is accomplished by locating a predefined value in a particular lock data word in the object. In response to the detection, a monitor is used to control local access to the shared object for the thread that is requesting the object. The monitor is on the same VM as the requesting thread. In addition, a data structure, which may be a proxy object, is updated in order to associate the monitor with the shared object.

公开(公告)号：GB2504491A

公开(公告)日：2014-02-05

申请号：GB201213532

申请日：2012-07-30

Applicant: IBM

Inventor： DEAKIN HANNAH JASMIN ,  MITCHELL SARA LOUISE ,  SLATTERY EDWARD JOHN

IPC: G06F11/36 ,  G06F9/455

Abstract: A debug service 350 is provided in a distributed network environment, such as a cloud computing environment. A debug agent 341 is connected with a workload process 311 to operate alongside it in a deployed application 310. The debug service receives 351 a request from the process that debugging is required. The service gathers information required for a debug session, including the source code 321 of the process, and then attaches itself to the process to carry out the debug session, interacting with the agent to do so. The session may be called based on failure criteria, or by user request. The service may operate across boundaries between workload processes where they collaborate to form the application, and may check that it has the required information for the next process to be debugged. The arrangement allows the operation of debugging methods in environments where normal methods are unable to function effectively.

公开(公告)号：DE102013207861A1

公开(公告)日：2013-11-21

申请号：DE102013207861

申请日：2013-04-30

Applicant: IBM

Inventor： SLATTERY EDWARD JOHN

IPC: G06F12/02

Abstract: Ein Verfahren und eine Vorrichtung zum Verwalten von Speicher in einem Computersystem werden offenbart, bei dem Metadaten, die dereferenzierte Speicherobjekte betreffen, für einen späteren Abruf gesichert werden.

公开(公告)号：GB2502076A

公开(公告)日：2013-11-20

申请号：GB201208434

申请日：2012-05-15

Applicant: IBM

Inventor： SLATTERY EDWARD JOHN

IPC: G06F12/02

Abstract: Method for computer memory management where metadata of dereferenced memory objects is saved for later retrieval. Memory locations in use are memory-objects (variable, function, data structure) 201 in first object graph 203 comprising system root object 204, the method comprises: creating dead root object 206 forming second object graph 205; in response to de-referencing first object 201, associating said dereferenced object with second graph 205 so that second graph comprises de-referenced object 209. Metadata object is created in second graph providing reference from second graph to a second object in first graph that referenced said first object immediately prior to dereferencing. Data of objects removed from memory after garbage collection are preserved, comprising identification of removed objects or details of objects referencing the removed objects. The identification is useful for monitoring, measuring, modifying or improving the memory usage. The data is used to reconstruct system object graphs at selected points.