Abstract: A system and method for cleaning dirty data in an intermediate cache are disclosed. A dirty data notification, including a memory address and a data class, is transmitted by a level 2 (L2) cache to frame buffer logic when dirty data is stored in the L2 cache. The data classes include evict first, evict normal and evict last. In one embodiment, data belonging to the evict first data class is raster operations data with little reuse potential. The frame buffer logic uses a notification sorter to organize dirty data notifications, where an entry in the notification sorter stores the DRAM bank page number, a first count of cache lines that have resident dirty data and a second count of cache lines that have resident evict_first dirty data associated with that DRAM bank. The frame buffer logic transmits dirty data associated with an entry when the first count reaches a threshold.

Abstract: One embodiment of the present invention sets forth a technique for performing a memory access request to compressed data within a virtually mapped memory system comprising an arbitrary number of partitions. A virtual address is mapped to a linear physical address, specified by a page table entry (PTE). The PTE is configured to store compression attributes, which are used to locate compression status for a corresponding physical memory page within a compression status bit cache. The compression status bit cache operates in conjunction with a compression status bit backing store. If compression status is available from the compression status bit cache, then the memory access request proceeds using the compression status. If the compression status bit cache misses, then the miss triggers a fill operation from the backing store. After the fill completes, memory access proceeds using the newly filled compression status information.

Abstract: The invention sets forth a crossbar unit that includes multiple virtual channels, each virtual channel being a logical flow of data within the crossbar unit. Arbitration logic coupled to source client subsystems is configured to select a virtual channel for transmitting a data request or a data packet to a destination client subsystem based on the type of the source client subsystem and/or the type of data request. Higher priority traffic is transmitted over virtual channels that are configured to transmit data without causing deadlocks and/or stalls. Lower priority traffic is transmitted over virtual channels that can be stalled.

Abstract: A graphics system utilizes page table entries to provide information on the storage format used to store graphics data. The page table entries, in turn, may be used for address translation. Exemplary kinds of storage format information include compression mode, a packing mode for storing Z data in tiles or color data in tiles, and a mode for allocating tile data among partitions in a partitioned memory.

Abstract: A technique for caching page access parameters, in accordance with one embodiment of the present invention, includes translating a given virtual address to a particular physical address using an address translation data structure. One or more page access parameters related to the particular physical address is stored in a separate page access data structure. The technique may further include accessing the page access data structure to manage access to memory as a function of the page access data.

Abstract: Deadlocks are avoided by marking read requests issued by a parallel processor to system memory as “special.” Read completions associated with read requests marked as special are routed on virtual channel 1 of the PCIe bus. Data returning on virtual channel 1 cannot become stalled by write requests in virtual channel 0, thus avoiding a potential deadlock.

Abstract: A method for executing processing operations using data stored in a memory. The method includes generating a snoop request configured to determine whether first data stored in a local memory is coherent relative to second data stored in a data cache, the snoop request including destination information that identifies the data cache on a bus, and a cache line address identifying where in the data cache the second data is located. The method further includes causing the snoop request to be transmitted over the bus to the second processor, extracting the cache line address from the snoop request, determining whether the second data is coherent, generating a complete message that includes completion information indicating that the first data is coherent with the second data, and causing the complete message to be transmitted over the bus to the first processor.

Abstract: One embodiment of the present invention sets forth a technique for increasing available storage space within compressed blocks of memory attached to data processing chips, without requiring a proportional increase in on-chip compression status bits. A compression status bit cache provides on-chip availability of compression status bits used to determine how many bits are needed to access a potentially compressed block of memory. A backing store residing in a reserved region of attached memory provides storage for a complete set of compression status bits used to represent compression status of an arbitrarily large number of blocks residing in attached memory. Physical address remapping (“swizzling”) used to distribute memory access patterns over a plurality of physical memory devices is partially replicated by the compression status bit cache to efficiently integrate allocation and access of the backing store data with other user data.

Abstract: A shared memory management system and method are described. In one embodiment, a memory management system includes a memory management unit for concurrently managing memory access requests from a plurality of engines. The shared memory management system independently controls access to the context memory without interference from other engine activities. In one exemplary implementation, the memory management unit tracks an identifier for each of the plurality of engines making a memory access request. The memory management unit associates each of the plurality of engines with particular translation information respectively. This translation information is specified by a block bind operation. In one embodiment the translation information is stored in a portion of instance memory. A memory management unit can be non-blocking and can also permit a hit under miss.

Abstract: According to the present invention, methods and apparatus are provided for increasing the efficiency of data access in a multiple processor, multiple cluster system. A home cluster of processors receives a cache access request from a request cluster. The home cluster includes mechanisms for instructing probed remote clusters to respond to the request cluster instead of to the home cluster. The home cluster can also include mechanisms for reducing the number of probes sent to remote clusters. Techniques are also included for providing the requesting cluster with information to determine the number of responses to be transmitted to the requesting cluster as a result of the reduction in the number of probes sent at the home cluster.

Abstract: A graphics system utilizes page table entries to provide information on the storage format used to store graphics data. The page table entries, in turn, may be used for address translation. Exemplary kinds of storage format information include compression mode, a packing mode for storing Z data in tiles or color data in tiles, and a mode for allocating tile data among partitions in a partitioned memory.

Abstract: According to the present invention, methods and apparatus are provided to allow dynamic multiple cluster system configuration changes. In one example, processors in the multiple cluster system share a virtual address space. Mechanisms for dynamically introducing and removing processors, I/O resources, and clusters are provided. The mechanisms can be implemented during reset or while a system is operating. Links can be dynamically enabled or disabled.

Abstract: A graphics system utilizes page table entries to provide information on the storage format used to store graphics data. The page table entries, in turn, may be used for address translation. Exemplary kinds of storage format information include compression mode, a packing mode for storing Z data in tiles or color data in tiles, and a mode for allocating tile data among partitions in a partitioned memory.

Abstract: Improved techniques are provided for detecting and correcting errors and skew in inter-cluster communications within computer systems having a plurality of multi-processor clusters. The local nodes of each cluster include a plurality of processors and an interconnection controller. Intra-cluster links are formed between the local nodes, including the interconnection controller, within a cluster. Inter-cluster links are formed between interconnection controllers of different clusters. Intra-cluster packets may be serialized and encapsulated as inter-cluster packets for transmission on inter-cluster links, preferably with link-layer encapsulation. Each inter-cluster packet may include a sequence identifier and error information computed for that packet. Clock data may be embedded in symbols sent on each bit lane of the inter-cluster links. Copies of transmitted inter-cluster packets may be stored until an acknowledgement is received.

Abstract: According to the present invention, methods and apparatus are provided for increasing the efficiency of data access in a multiple processor, multiple cluster system. A home cluster of processors receives a cache access request from a request cluster. The home cluster includes mechanisms for instructing probed remote clusters to respond to the request cluster instead of to the home cluster. The home cluster can also include mechanisms for reducing the number of probes sent to remote clusters. Techniques are also included for providing the requesting cluster with information to determine the number of responses to be transmitted to the requesting cluster as a result of the reduction in the number of probes sent at the home cluster.

Abstract: Improved techniques are provided for reducing latency in inter-cluster communications within computer systems having a plurality of multi-processor clusters. The local nodes of each cluster include a plurality of processors and an interconnection controller. Intra-cluster links are formed between the local nodes, including the interconnection controller, within a cluster. Inter-cluster links are formed between interconnection controllers of different clusters. Intra-cluster packets may be encapsulated as inter-cluster packets and stored in a transmission buffer pending transmission on an inter-cluster link. When the transmission buffer is empty, a control character is transmitted on an inter-cluster link. The control character is not stored in the transmission buffer or in a reception buffer, but instead is dropped. Clock data may be embedded in symbols sent on each bit lane of the inter-cluster links, including the symbol(s) of the control character.

Abstract: Method and apparatus for processing one or more fragment data. In one embodiment, the method includes processing one or more fragment data to generate one or more texture map addresses for one or more texels, determining relevance information that correspond to the texture map addresses, and translating the relevance information into a rendering constraint data structure.

Abstract: According to the present invention, methods and apparatus are provided for increasing the efficiency of data access in a multiple processor, multiple cluster system. Mechanisms for reducing the number of transactions in a multiple cluster system are provided. In one example, owning node information is used to limit the number of probes transmitted in a particular cluster.

Abstract: A computer system having a plurality of processing nodes interconnected by a first point-to-point architecture is described. Each processing node has a cache memory associated therewith. A probe filtering unit is operable to receive probes corresponding to memory lines from the processing nodes and to transmit the probes only to selected ones of the processing nodes with reference to probe filtering information. The probe filtering information is representative of states associated with selected ones of the cache memories.

Abstract: According to the present invention, methods and apparatus are provided for increasing the efficiency of data access in a multiple processor, multiple cluster system. Mechanisms for improving the accuracy of information available to a cache coherence controller are provided in order to allow the cache coherence controller to reduce the number of transactions in a multiple cluster system. Cache state information is provided to a home cluster cache coherence controller to allow silent evictions of shared memory lines and change to dirty associated intervening requests to be efficiently handled.