49 results about "Memory coherence" patented technology

The invention comprises (i) a compilation method for automatically converting a single-threaded software program into an application-specific supercomputer, and (ii) the supercomputer system structure generated as a result of applying this method. The compilation method comprises: (a) Converting an arbitrary code fragment from the application into customized hardware whose execution is functionally equivalent to the software execution of the code fragment; and (b) Generating interfaces on the hardware and software parts of the application, which (i) Perform a software-to-hardware program state transfer at the entries of the code fragment; (ii) Perform a hardware-to-software program state transfer at the exits of the code fragment; and (iii) Maintain memory coherence between the software and hardware memories. If the resulting hardware design is large, it is divided into partitions such that each partition can fit into a single chip. Then, a single union chip is created which can realize any of the partitions.

A computer system including a first processor, a second processor in communication with the first processor, a memory coupled to the first and second processors (i.e., a shared memory) and including multiple memory locations, and a storage device coupled to the first processor. The first and second processors implement virtual memory using the memory. The first processor maintains a first set of page tables and a second set of page tables in the memory. The first processor uses the first set of page tables to access the memory locations within the memory. The second processor uses the second set of page tables, maintained by the first processor, to access the memory locations within the memory. A virtual memory page replacement method is described for use in the computer system, wherein the virtual memory page replacement method is designed to help maintain paged memory coherence within the multiprocessor computer system.

A graphics system including a custom graphics and audio processor produces exciting 2D and 3D graphics and surround sound. The system includes a graphics and audio processor including a 3D graphics pipeline and an audio digital signal processor. The graphics pipeline processes graphics commands at different rates depending upon the type of operation being performed. This makes it difficult to synchronize pipeline operations with external operations (e.g., a graphics processor with a main processor). To solve this problem, a synchronization token including a programmable data message is inserted into a graphics command stream sent to a graphics pipeline. At a predetermined point near the bottom of the pipeline, the token is captured and a signal is generated indicated the token has arrived. The graphics command producer can look at the captured token to determine which of multiple possible tokens has been captured, and can use the information to synchronize a task with the graphics pipeline. Applications include maintaining memory coherence in memory shared between the 3D graphics pipeline and a graphics command producer.

A method and apparatus to protect memory consistency in a multiprocessor computing system are described, in particular relating to program code conversion such as dynamic binary translation. The exemplary system provides a memory, processors and a controller/translator unit (CTU) arranged to convert subject code into at least first and second target code portions executable on the processors. The CTU comprises an address space allocation unit to provide virtual address space regions and direct the target code portions to access the memory therethough; a shared memory detection unit to detect a request to access a shared memory area, accessible by both target code portions, and to identify at least one group of instructions in the first target code portion which access the shared memory area; and a memory protection unit to selectively apply memory consistency protection in relation to accesses to the shared memory area by the identified group of instructions.

A system for efficiently verifying compliance with a memory consistency model includes a test module and an analysis module. The test module may coordinate an execution of a multithreaded test program on a test platform. If the test platform provides an indication of the order in which writes from multiple processing elements are performed at shared memory locations, the analysis module may use a first set of rules to verify that the results of the execution correspond to a valid ordering of events according to a memory consistency model. If the test platform does not provide an indication of write ordering, the analysis module may use a second set of rules to verify compliance with the memory consistency model. Further, a backtracking search may be performed to find a valid ordering if such ordering exists or show that none exists and, hence, confirm whether or not the results comply with the given memory consistency model.

A data processing system includes an interconnect, a system memory and a number of snoopers coupled to the interconnect, and response logic. In response to a requesting snooper issuing a data request on the interconnect specifying a memory address, the snoopers provide snoop responses. The response logic compiles the snoop responses to obtain a combined response including an indication of a demand-source snooper that will source requested data associated with the memory address to the requesting snooper and an indication of whether additional non-requested data will be supplied to the requesting snooper. This combined response is then transmitted to the snoopers on the interconnect to direct the provision of the requested data, and possibly unrequested prefetch data, to the requesting snooper.

A method, apparatus and program for booting a non-uniform-memory-access (NUMA) machine are provided. The invention comprises configuring a plurality of standalone, symmetrical multiprocessing (SMP) systems to operate within a NUMA system. A master processor is selected within each SMP; the other processors in the SMP are designated as NUMA slave processors. A NUMA master processor is then chosen from the SMP master processors; the other SMP master processors are designated as NUMA slave processors. A unique NUMA ID is assigned to each SMP that will be part of the NUMA system. The SMPs are then booted in NUMA mode in one-pass with memory coherency established right at the beginning of the execution of the system firmware.

An apparatus and method are disclosed for single-stepping coherence events in a multiprocessor system under software control in order to monitor the behavior of a memory coherence mechanism. Single-stepping coherence events in a multiprocessor system is made possible by adding one or more step registers. By accessing these step registers, one or more coherence requests are processed by the multiprocessor system. The step registers determine if the snoop unit will operate by proceeding in a normal execution mode, or operate in a single-step mode.

Techniques are described for memory coherence in a multi-core system with a heterogeneous memory architecture comprising one or more hardware-managed caches and one or more software-managed caches. According to one embodiment, a set of one or more buffers are allocated in memory, and each respective buffer is associated with a respective metadata tag. The metadata tag may be used to store metadata that identifies a state associated with the respective buffer. The multi-core system may enforce coherence for the one or more hardware-managed caches and the one or more software-managed caches based on the metadata stored in the metadata tag for each respective buffer in the set of one or more buffers. The multi-core system may read the metadata to determine whether a particular buffer is in a hardware-managed or a software-managed cacheable state. Based on the current state of the particular buffer, the multi-core system may perform coherence operations.

A memory coherence protocol is provided for using cache line access frequencies to dynamically switch from an invalidation protocol to an update protocol. A frequency access count (FAC) is associated with each line of data in a memory area, such as each cache line in a private cache corresponding to a CPU in a multiprocessor system. Each time the line is accessed, the FAC associated with the line is incremented. When the CPU, or process, receives an invalidate signal for a particular line, the CPU checks the FAC for the line. If the CPU, or process, determines that it is a frequent accessor of a particular line that has been modified by another CPU, or process, the CPU sends an update request in order to obtain the modified data. If the CPU is not a frequent accessor of a line that has been modified, the line is simply invalidated in the CPU's memory area. By dynamically switching from an invalidate protocol to an update protocol, based on cache line access frequencies, efficiency is maintained while cache misses are minimized. Preferably, all FACs are periodically reset in order to ensure that the most recent cache line access data in considered.

The invention provides a design method of a distributed virtual machine self-adaptive memory consistency protocol, which comprises the following steps of: intercepting synchronous operation: aiming atdifferent applications, flexibly switching a vCPU (Virtual Central Processing Unit) of a client in a sequential consistency mode and a TSO (Transport Storage and Offloading) mode; the state description of the synchronization protocol: on the basis of the original memory consistency protocol, adding Dirty and a corresponding atomic operation state, and realizing state transition. Meanwhile, the invention provides a distributed virtual machine self-adaptive memory consistency protocol obtained based on the design method and a terminal used for executing the design method. According to the method, on the basis of the distributed virtual machine, the distributed shared memory can obtain better performance. According to the method, the memory synchronization protocol of the distributed virtualmachine is flexibly switched in order consistency and x86-TSO; for different application scenarios and restrictions, the adaptive consistency protocol relaxes the original sequence consistency to x86-TSO, so that the performance of the distributed shared memory is improved.

Techniques are disclosed relating to memory consistency in a memory hierarchy with relaxed ordering. In some embodiments, an apparatus includes a first level cache that is shared by a plurality of shader processing elements and a second level cache that is shared by the shader processing elements and at least a texture processing unit. In some embodiments, the apparatus is configured to execute operations specified by graphics instructions that include (1) an attribute of the operation that specifies a type of memory consistency to be imposed for the operation and (2) scope information for the attribute that specifies whether the memory consistency specified by the attribute should be enforced at the first level cache or the second level cache. In some embodiments, the apparatus is configured to determine whether to sequence memory accesses at the first level cache and the second level cache based on the attribute and the scope.

Memory consistency is provided in an emulated processing environment. A processor architected with a weak memory consistency emulates an architecture having a firm memory consistency. This memory consistency is provided without requiring serialization instructions or special hardware.

The invention discloses a distributed memory management method based on network and page granularity management. The method is characterized in that a node, page and block memory management method isadopted for a global address; partial memories from different nodes are packaged in a global page-based memory (GPM), and data transmission under page granularity is supported, memory management is performed based on the updated memory consistency protocol, so that the data on the local cache page is consistent with the data on the GPM, and a high-level application program deployed in the PDMM transparently accesses the GPM, thereby realizing low-delay and high-throughput inter-node access. Compared with the prior art, the inter-node access delay with low delay and high throughput is achieved,the method is simple and convenient to use, the problem that cache is invalid due to write operation in data intensive work is effectively solved, and the performance of the PDMM is superior to thatof other products of the same type.