Nvidia Corporation

Altera Corporation

Pts Corporation

Other

Bops, Incorporated

Billions of Operations Per Second

Patrick Rene Marchand

Gerald George Pechanek

David Brian Glasco

Peter B Holmqvist

James P Roberts

George R Lynch

Edwin Franklin Barry

Edward A Wolff

Edwin Frank Barry

Karan Mehra

Charles W Kurak, Jr

David Carl Strube

Larry D Larsen

John H Edmondson

Nikos P Pitsianis

Ricardo E Rodriguez

Grayson Morris

Dale Edward Schneider

Carl Donald Busboom

Marco Jacobs

Marco C Jacobs

Steven E Molnar

Samuel Hammond Duncan

Robert Anthony Alfieri

Cass Watson Everitt

Juan Guillermo Revilla

John Stephen Montrym

Emmett Michael Kilgariff

David Kirk McAllister

Christopher D S Donham

Dennis K Ma

Duncan Andrew Riach

Wei-Je Robert Huang

Dane Mrazek

Brijesh Tripathi

Ravi Kiran Manyam

Yin Fung Tang

Atul Kalambur

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.

Compression status bit cache and backing store

Processor pipeline controlling techniques are described which take advantage of the variation in critical path lengths of different instructions to achieve increased performance. By examining a processor's instruction set and execution unit implementation's critical timing paths, instructions are classified into speed classes. Based on these speed classes, one pipeline is presented where hold signals are used to dynamically control the pipeline based on the instruction class in execution. An alternative pipeline supporting multiple classes of instructions is presented where the pipeline clocking is dynamically changed as a result of decoded instruction class signals. A single pass synthesis methodology for multi-class execution stage logic is also described. For dynamic class variable pipeline processors, the mix of instructions can have a great effect on processor performance and power utilization since both can vary by the program mix of instruction classes. Application code can be given new degrees of optimization freedom where instruction class and the mix of instructions can be chosen based on performance and power requirements.

Methods and apparatus for adapting pipeline stage latency based on instruction type

Hardware and software techniques for interrupt detection and response in a scalable pipelined array processor environment are described. Utilizing these techniques, a sequential program execution model with interrupts can be maintained in a highly parallel scalable pipelined array processing containing multiple processing elements and distributed memories and register files. When an interrupt occurs, interface signals are provided to all PEs to support independent interrupt operations in each PE dependent upon the local PE instruction sequence prior to the interrupt. Processing/element exception interrupts are supported and low latency interrupt processing is also provided for embedded systems where real time signal processing is required. Further, a hierarchical interrupt structure is used allowing a generalized debug approach using debut interrupts and a dynamic debut monitor mechanism.

Methods and apparatus for scalable array processor interrupt detection and response

Details of a highly cost effective and efficient implementation of a manifold array (ManArray) architecture and instruction syntax for use therewith are described herein. Various aspects of this approach include the regularity of the syntax, the relative ease with which the instruction set can be represented in database form, the ready ability with which tools can be created, the ready generation of self-checking codes and parameterized test cases. Parameterizations can be fairly easily mapped and system maintenance is significantly simplified.

System core for transferring data between an external device and memory

One embodiment of the invention sets forth a method for transmitting display data to a display device. The method includes the steps of receiving a contract for a frame of display data, preparing the frame of display data to ensure the timing requirements of the display device can be satisfied based on the contract, and transmitting the frame of display data to the display device while the contract is pending.

Isochronous hub contracts

A method for cleaning dirty data in an intermediate cache is 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 may 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.

Techniques for evicting dirty data from a cache using a notification sorter and count thresholds

A system and method for buffering intermediate data in a processing pipeline architecture stores the intermediate data in a shared cache that is coupled between one or more pipeline processing units and an external memory. The shared cache provides storage that is used by multiple pipeline processing units. The storage capacity of the shared cache is dynamically allocated to the different pipeline processing units as needed, to avoid stalling the upstream units, thereby improving overall system throughput.

Padding buffer requests to avoid reads of invalid data

One embodiment of the invention sets forth a mechanism for evicting data from a data cache based on the data class of that data. The data stored in the cache lines in the data cache is categorized based on data classes that reflect the reuse potential of that data. The data classes are stored in a tag store, where each tag within the tag store corresponds to a single cache line within the data cache. When reserving a cache line for the data associated with a command, a tag look-up unit examines the data classes in the tag store to determine which data to evict. Data that has a low reuse potential is evicted at a higher priority than data that has a high reuse potential. Advantageously, evicting data that belongs to a data class that has a lower reuse potential reduces the number of cache misses within the system.

Multi-class data cache policies

One embodiment of the present invention sets forth a compression status cache configured to store compression information for blocks of memory stored within an external memory. A data cache unit is configured to request, in response to a cache miss, compressed data from the external memory based on compression information stored in the compression status bit cache. The compression status for active buffers is dynamically swapped into the compression status cache as needed. Different compression formats may be specified for one or more tiles within an active buffer. One advantage of the disclosed compression status cache is that a lame amount of attached memory may be allocated as compressible memory blocks, without incurring a corresponding die area cost because a portion of the compression status stored off chip in attached memory is cached in the compression status cache.

Compression status caching

Hardware and software techniques for interrupt detection and response in a scalable pipelined array processor environment are described. Utilizing these techniques, a sequential program execution model with interrupts can be maintained in a highly parallel scalable pipelined array processing containing multiple processing elements and distributed memories and register files. When an interrupt occurs, interface signals are provided to all PEs to support independent interrupt operations in each PE dependent upon the local PE instruction sequence prior to the interrupt. Processing/element exception interrupts are supported and low latency interrupt processing is also provided for embedded systems where real time signal processing is required. Further, a hierarchical interrupt structure is used allowing a generalized debug approach using debug interrupts and a dynamic debut monitor mechanism.

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.

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.

Efficient line and page organization for compression status bit caching

One embodiment of the present invention sets forth a compression status bit cache with deterministic latency for isochronous memory clients of compressed memory. The compression status bit cache improves overall memory system performance by providing on-chip availability of compression status bits that are used to size and interpret a memory access request to compressed memory. To avoid non-deterministic latency when an isochronous memory client accesses the compression status bit cache, two design features are employed. The first design feature involves bypassing any intermediate cache when the compression status bit cache reads a new cache line in response to a cache read miss, thereby eliminating additional, potentially non-deterministic latencies outside the scope of the compression status bit cache. The second design feature involves maintaining a minimum pool of clean cache lines by opportunistically writing back dirty cache lines and, optionally, temporarily blocking non-critical requests that would dirty already clean cache lines. With clean cache lines available to be overwritten quickly, the compression status bit cache avoids incurring additional miss write back latencies.

Compression status bit cache with deterministic isochronous latency

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.

Virtual channels for effective packet transfer

Storing dynamically sized buffers within a cache

Growing community of inventors

Apex, NC, United States of America

Patrick Rene Marchand