摘要:
In a NUMA-topology computer system that includes multiple nodes and multiple logical partitions, some of which may be dedicated and others of which are shared, NUMA optimizations are enabled in shared logical partitions. This is done by specifying a home node parameter in each virtual processor assigned to a logical partition. When a task is created by an operating system in a shared logical partition, a home node is assigned to the task, and the operating system attempts to assign the task to a virtual processor that has a home node that matches the home node for the task. The partition manager then attempts to assign virtual processors to their corresponding home nodes. If this can be done, NUMA optimizations may be performed without the risk of reducing the performance of the shared logical partition.
摘要:
In a NUMA-topology computer system that includes multiple nodes and multiple logical partitions, some of which may be dedicated and others of which are shared, NUMA optimizations are enabled in shared logical partitions. This is done by specifying a home node parameter in each virtual processor assigned to a logical partition. When a task is created by an operating system in a shared logical partition, a home node is assigned to the task, and the operating system attempts to assign the task to a virtual processor that has a home node that matches the home node for the task. The partition manager then attempts to assign virtual processors to their corresponding home nodes. If this can be done, NUMA optimizations may be performed without the risk of reducing the performance of the shared logical partition.
摘要:
Mechanisms are provided, for implementation in a data processing system having at least one physical processor and at least one associated cache memory, for allocating cache resources of the at least one cache memory to virtual processors of the data processing system. The mechanisms identify a plurality of high priority virtual processors in the data processing system. The mechanisms further determine a percentage of cache lines of the at least one cache memory to be assigned to high priority virtual processors. Moreover, the mechanisms mark a portion of the cache lines in the at least one cache memory as being evictable by only high priority virtual processors based on the determined percentage of cache lines to be assigned to high priority virtual processors. The marked portion of the cache lines cannot be evicted by lower priority virtual processors having a priority lower than the high priority virtual processors.
摘要:
Mechanisms are provided, for implementation in a data processing system having at least one physical processor and at least one associated cache memory, for allocating cache resources of the at least one cache memory to virtual processors of the data processing system. The mechanisms identify a plurality of high priority virtual processors in the data processing system. The mechanisms further determine a percentage of cache lines of the at least one cache memory to be assigned to high priority virtual processors. Moreover, the mechanisms mark a portion of the cache lines in the at least one cache memory as being evictable by only high priority virtual processors based on the determined percentage of cache lines to be assigned to high priority virtual processors. The marked portion of the cache lines cannot be evicted by lower priority virtual processors having a priority lower than the high priority virtual processors.
摘要:
Functionality is implemented to determine that a plurality of multi-core processing units of a system are configured in accordance with a plurality of operating performance modes. It is determined that a first of the plurality of operating performance modes satisfies a first performance criterion that corresponds to a first workload of a first logical partition of the system. Accordingly, the first logical partition is associated with a first set of the plurality of multi-core processing units that are configured in accordance with the first operating performance mode. It is determined that a second of the plurality of operating performance modes satisfies a second performance criterion that corresponds to a second workload of a second logical partition of the system. Accordingly, the second logical partition is associated with a second set of the plurality of multi-core processing units that are configured in accordance with the second operating performance mode.
摘要:
According to one aspect of the present disclosure, a method and technique for managing rollback in a transactional memory environment is disclosed. The method includes, responsive to detecting a begin transaction directive by a processor supporting transactional memory processing, detecting an access of a first memory location not needing rollback and indicating that the first memory location does not need to be rolled back while detecting an access to a second memory location and indicating that a rollback will be required. The method also includes, responsive to detecting an end transaction directive after the begin transaction directive and a conflict requiring a rollback, omitting a rollback of the first memory location while performing rollback on the second memory location.
摘要:
A storage subsystem combining solid state drive (SSD) and hard disk drive (HDD) technologies provides low access latency and low complexity. Separate free lists are maintained for the SSD and the HDD and blocks of file system data are stored uniquely on either the SSD or the HDD. When a read access is made to the subsystem, if the data is present on the SSD, the data is returned, but if the block is present on the HDD, it is migrated to the SSD and the block on the HDD is returned to the HDD free list. On a write access, if the block is present in the either the SSD or HDD, the block is overwritten, but if the block is not present in the subsystem, the block is written to the HDD.
摘要:
According to one aspect of the present disclosure, a method and technique for using processor registers for extending a cache structure is disclosed. The method includes identifying a register of a processor, identifying a cache to extend, allocating the register as an extension of the cache, and setting an address of the register as corresponding to an address space in the cache.
摘要:
A method, system and computer-usable medium are disclosed for managing prefetch streams in a virtual machine environment. Compiled application code in a first core, which comprises a Special Purpose Register (SPR) and a plurality of first prefetch engines, initiates a prefetch stream request. If the prefetch stream request cannot be initiated due to unavailability of a first prefetch engine, then an indicator bit indicating a Prefetch Stream Dispatch Fault is set in the SPR, causing a Hypervisor to interrupt the execution of the prefetch stream request. The Hypervisor then calls its associated operating system (OS), which determines prefetch engine availability for a second core comprising a plurality of second prefetch engines. If a second prefetch engine is available, then the OS migrates the prefetch stream request from the first core to the second core, where it is initiated on an available second prefetch engine.
摘要:
A system and method for efficiently managing lock contention for a central processing unit (CPU) of a computer system. The present invention uses both spinning and blocking (or undispatching) to manage threads when they are waiting to acquire a lock. In addition, the present invention intelligently determines when the program thread should spin and when the program thread should become undispatched. If it is determined that the program thread should become undispatched, the present invention provides efficient undispatching of program threads that improves throughput by reducing wait time to acquire the lock. A lock contention management system includes a dispatcher for managing the execution of threads on CPUs as well as threads that are currently ready to run but not executing because they are waiting for an available CPU, a dispatch management module that determines when a program thread should become undispatched to wait on a lock and when the program thread should spin, and low-priority execution module for undispatching the program thread. The present invention also includes a lock contention management method using the above system.