Abstract:
A processor performance profiler is enabled to for identify specific instructions causing performance issues within a program being executed by a microprocessor through random sampling to find the worst-case offenders of a particular event type such as a cache miss or a branch mis-prediction. Tracking all instructions causing a particular event generates large data logs, creates performance penalties, and makes code analysis more difficult. However, by identifying and tracking the worst offenders within a random sample of events without having to hash all events results in smaller memory requirements for the performance profiler, lower performance impact while profiling, and decreased complexity to analyze the program to identify major performance issues, which, in turn, enables better optimization of the program in shorter developer time.
Abstract:
Systems, methods and/or devices are used to reduce declared capacity of non-volatile memory of a storage device in a storage system. In one aspect, the method includes, detecting an amelioration trigger for reducing declared capacity of non-volatile memory of a storage device of the storage system, and in accordance with the detected amelioration trigger, performing an amelioration process to reduce declared capacity of the non-volatile memory of the storage device, the performing including reducing a range of logical addresses of a logical address space available to a host. In some embodiments, the storage device includes one or more flash memory devices. In some embodiments, the detecting, the performing, or both are performed by the storage device, or by one or more subsystems of the storage system distinct from the storage device, or by the host.
Abstract:
A high-performance-computer system includes a statistics accumulation apparatus configured to efficiently accumulate system performance data from a variety of system components, and periodically write such data to processor local memory for efficient subsequent software processing of the thus acquired data, thereby reducing the system hardware and software overhead needed for collection of such data as compared to prior art systems.
Abstract:
An Extended Input/output (I/O) measurement word facility is provided. Provision is made for emulation of the Extended I/O measurement word facility. The facility provides for storing measurement data associated with a single I/O operation in an extended measurement word associated with an I/O response block. In a further aspect, the stored data may have a resolution of approximately one-half microsecond.
Abstract:
Systems and methods for memory page offloading in multi-processor computer systems. An example method may comprise: detecting, by a computer system, a memory pressure condition on a first node; invalidating a page table entry for a memory page residing on the first node; copying the memory page to a second node; and updating the page table entry for the memory page to reference the second node.
Abstract:
There is provided a method for verifying a full drive as an error check processing of a storage drive, but along with the increase in capacity of the storage drive, there are cases where error check cannot be executed before using the storage area. In the present invention, quantity of use of data is estimated, and the storage areas corresponding to the estimated quantity are verified periodically. By verifying the storage area to be used in advance before actual data use, the occurrence of error can be prevented without having to execute a check of the full drive.
Abstract:
Systems and methods for monitoring hardware observation points within a System on a Chip (SoC) are disclosed. In one embodiment, a monitoring system includes observers electrically coupled to corresponding hardware observation points. For each observation time period, the observer receives events from a corresponding hardware observation point and outputs observation data indicative of the number of events received. Each observer operates in the same clock and power domain as the corresponding hardware observation point. The hardware monitoring system includes collectors; each collector receives the observation data from a corresponding observer for the observation time period(s). Each collector updates a value of a live-counter based on the observation data and outputs the value of the live-counter. By counting the number of events at the hardware observation points, the monitoring system can enable detailed runtime monitoring, which can be used, e.g., for optimizing future hardware iterations and to support software power optimizations.
Abstract:
A quality score for a computer application release is determined using a first number of unique users who have launched the computer application release on user devices and a second number of unique users who have encountered at least once an abnormal termination with the computer application release on user devices. Additionally or optionally, an application quality score can be computed for a computer application based on quality scores of computer application releases that represent different versions of the computer application. Additionally or optionally, a weighted application quality score can be computed for a computer application by further taking into consideration the average application quality score and popularity of a plurality of computer applications.
Abstract:
User interfaces are provided that display application energy usage information to a user. Application energy usage information may be displayed as an energy meter user interface that provides derived information (e.g., power and time used to determine energy usage) that is not available through a traditional activity monitor. For example, energy usage information is provided for an application running in isolation, taking into account energy usage based on what the application is doing and not what the rest of the system is doing (e.g., shows CPU wakes in isolation of any other processes). Each of a series of energy bars is presented as an energy utilization metric that includes at least one of a utilization portion and a wake tax portion, combined to represent to represent total energy utilization.
Abstract:
When performing a read operation on a non-volatile memory device which includes a plurality of memory sections each corresponding to a plurality of data units, the read count of a specific memory section and the error bits of its corresponding data units are monitored for determining whether data relocation should be perform. When the read count of the specific section exceeds a read count threshold and the error bits of any corresponding data unit exceeds an error threshold, data is moved from the specific memory section to another memory section of the non-volatile memory device, thereby preventing read disturbance from occurring in the specific memory sections.