Unisys Corporation

Eugene A Rodi

Mitchell A Bauman

Justin S Neils

John S Jensen

Douglas E Morrissey

Ferris T Price, Deceased

Merrill J Nelson

Roger Lee Gilbertson

Michael L Haupt

Aaron C Peterson

Marvin W Theis

Bart E Reigstad

Robert M Rice

Error correction and error detection related to DRAM chip failures, particularly adapted server memory subsystems. It uses &#xd7;4 bit DRAM devices organized in a code word of 128 data bit words and 16 check bits. These 16 check bits are generated in such a way as to provide a code capable of 4 bit adjacent error correction within a family (i.e., in a &#xd7;4 DRAM) and double bit non-adjacent error detection across the entire 128 bit word, with single bit correction across the word as well. Each device can be thought of as a separate family of bits, errors occurring in more than one family are not correctable, but may be detected if only one bit in each of two families is in error. Syndrome generation and regeneration are used together with a specific large code word. Decoding the syndrome and checking it against the regenerated syndrome yield data sufficient for providing the features described.

Familial correction with non-familial double bit error detection

In general, techniques are described for initializing a memory module in accordance with a programmable initialization sequence. A memory controller, for example, includes a programmable computer-readable medium that stores configuration data to control initialization of one or more memory modules. The memory controller includes an initialization control unit that outputs a sequence of commands to initialize the memory modules in accordance with the configuration data. The initialization control unit may select the sequence of commands from a set of predefined initialization sequences based on the configuration data.

Memory controller having programmable initialization sequence

A system and method for increasing the throughput of a directory-based storage system is provided. The storage system includes a data storage system to store data signals, and a directory to store state information for the data signals. Requests issued to the storage system are grouped into sets. The requests within a same set are issued in succession to the data storage system to initiate read and/or write memory operations. At the same time, directory entries are read from the directory for each request in the set. Each directory entry is updated as it is retrieved to reflect the requested memory operation. After all directory entries are retrieved, the updated entries are stored back to the directory in succession so that the bi-directional interface to the directory undergoes only a single direction change during the processing of the set.

System and method for increasing bandwidth in a directory based high speed memory system

A system and method for testing and/or initializing a Directory Store in a directory-based coherent memory. In one illustrative embodiment, the directory-based coherent memory includes a Main Store for storing a number of data entries, a Directory Store for storing the directory state for at least some of the data entries in the Main Store, and a next state block for determining a next directory state for a requested data entry in response to a memory request. To provide access to the Directory Store, and in one illustrative embodiment, a selector is provided for selecting either the next directory state value provided by the next state block or another predetermined value. The other predetermined value may be, for example, a fixed data pattern, a variable data pattern, a specified value, or any other value suitable for initializing and/or testing the Directory Store. The output of the selector may be written to the Directory Store.

System and method for testing and initializing directory store memory

A program and method enables easy creation and manipulation of timing charts. The preferred embodiment employs off-the-shelf commercial software and uses Visual Basic commands to get timing chart drawing commands into the drawing program and out of the spreadsheet program to order the drawing program to produce a displayable and print/plotable file. The user can easily see changes needed and even if they require ripple-through redrawing, because the user manipulates data in the spreadsheet file instead of directly manipulating drawing commands, the spread sheet will carry through ripple-through calculations to modify all lines related to the recalculated data.

Efficient timing chart creation and manipulation

Error correction and error detection related to DRAM chip failures, particularly adapted server memory subsystems. The application of a code for 128 bit memories is applied to a 20 bit directory store to improve reliability of the directory store memory of the computer system. The code uses &#xd7;4 bit DRAM devices organized in a code word of 20 data bit words and 12 check bits. These 12 check bits provide a code capable of 4 bit adjacent error correction within a family (i.e., in a &#xd7;4 DRAM) and double bit non-adjacent error detection across the entire 20 bit word, with single bit correction across the word as well. Each device can be though of as a separate family of bits, errors occurring in more than one family are not correctable, but may be detected if only one bit in each of two families is in error. Syndrome generation and regeneration are used together with a specific large code word. Decoding the syndrome and checking it against the regenerated syndrome yield data sufficient for providing the features described.

Familial correction with non-familial double bit error detection for directory storage

A directory-based cache coherency system is disclosed for use in a data processing system having multiple Instruction Processors (IP) and multiple Input/Output (I/O) units coupled through a shared main memory. The system includes one or more IP cache memories, each coupled to one or more IPs and to the shared main memory for caching units of data referred to as cache lines. The system further includes one or more I/O memories within ones of the I/O units, each I/O memory being coupled to the shared main memory for storing cache lines retrieved from the shared main memory. Coherency is maintained through the use of a central directory which stores status for each of the cache lines in the system. The status indicates the identity of the IP caches and the I/O memories having valid copies of a given cache line, and further identifies a set of access privileges, that is, the cache line &ldquo;state&rdquo;, associated with the cache line. The cache line states are used to implement a state machine which tracks the cache lines and ensures only valid copies of are maintained within the memory system. According to another aspect of the system, the main memory performs continuous tracking and control functions for all cache lines residing in the IP caches. In contrast, the system maintains tracking and control functions for only predetermined cache lines provided to the I/O units so that system overhead may be reduced. The coherency system further supports multiple heterogeneous instruction processors which operate on cache lines of different sizes.

Directory-based cache coherency system supporting multiple instruction processor and input/output caches

A directory-based cache coherency system is disclosed for use in a data processing system having multiple Instruction Processors (IP) and multiple Input/Output (I/0) units coupled through a shared main memory. The system includes one or more IP cache memories, each coupled to one or more IPs and to the shared main memory for caching units of data referred to as cache lines. The system further includes one or more I/O memories within ones of the I/O units, each I/O memory being coupled to the shared main memory for storing cache lines retrieved from the shared main memory. Coherency is maintained through the use of a central directory which stores status for each of the cache lines in the system. The status indicates the identity of the IP caches and the I/O memories having valid copies of a given cache line, and further identifies a set of access privileges, that is, the cache line &ldquo;state&rdquo;, associated with the cache line. The cache line states are used to implement a state machine which tracks the cache lines and ensures only valid copies of are maintained within the memory system. According to another aspect of the system, the main memory performs continuous tracking and control functions for all cache lines residing in the IP caches. In contrast, the system maintains tracking and control functions for only predetermined cache lines provided to the I/O units so that system overhead may be reduced. The coherency system further supports multiple heterogeneous instruction processors which operate on cache lines of different sizes.

Directory based cache coherency system supporting multiple instruction processor and input/output caches

A high-speed memory system is disclosed for use in supporting a directory-based cache coherency protocol. The memory system includes at least one data system for storing data, and a corresponding directory system for storing the corresponding cache coherency information. Each data storage operation involves a block transfer operation performed to multiple sequential addresses within the data system. Each data storage operation occurs in conjunction with an associated read-modify-write operation performed on cache coherency information stored within the corresponding directory system. Multiple ones of the data storage operations may be occurring within one or more of the data systems in parallel. Likewise, multiple ones of the read-modify-write operations may be performed to one or more of the directory systems in parallel. The transfer of address, control, and data signals for these concurrently performed operations occurs in an interleaved manner. The use of block transfer operations in combination with the interleaved transfer of signals to memory systems prevents the overhead associated with the read-modify-write operations from substantially impacting system performance. This is true even when data and directory systems are implemented using the same memory technology.

High-speed memory storage unit for a multiprocessor system having integrated directory and data storage subsystems

A system and method for testing and initializing a memory including multiple memory banks or a memory module partitioned into logical memory units. A plurality of memory exerciser testers are provided, one for each of the plurality of memory banks. Each of the memory exerciser testers includes an address generator to generate a sequence of memory bank addresses to successively address each of the memory banks in a cyclic manner, while each of the address generators concurrently addresses a different one of the memory banks. A data pattern generator is coupled to a corresponding one of the address generators to receive a data pattern control signal upon each output of each of the memory bank addresses generated by its corresponding address generator. The data pattern generator outputs a unique data pattern to the memory bank identified by the memory bank address in response to each occurrence of the data pattern control signal. A plurality of address initialization registers are provided, one for each of the plurality of exerciser testers. Each of the address initialization registers stores an initial memory bank address for one of the memory banks such that each of the address generators is preset to initially address a different one of the memory banks. In this manner, each memory bank is addressed by a different one of the address generators at any given time, which provides for concurrent testing of all memory banks and memory interfaces.

System and method for performing parallel initialization and testing of multiple memory banks and interfaces in a shared memory module

A data processing system and method for substituting selected requests with substitute requests that perform the same or similar end function but achieve increased system performance are disclosed. Those requests that have a selected request characteristic are identified and converted or replaced with a predetermined substitute request. The substitute requests perform at least part of the function of the identified requests. The data processing system may include two or more processors, and the selected request characteristic may be that a write data packet of an identified write request was not changed by a first processor. A substitute request may update directory information associated with the identified write request but may not write to associated data packet to memory. The directory information can indicate whether identified memory locations are currently owned by a processor. The substitute request may also send the associated write data packet from the first processor to a second processor via a processor-to-processor interface. It may be determined whether a processor block has changed selected data by examining a function code of a write type request.

Data processing system and method for substituting one type of request for another for increased performance when processing back-to-back requests of certain types

The disclosure relates to a high performance fault tolerant queuing system. Multiple processors share access to one or more queues which are stored in an addressable memory. A storage controller provides general access to the addressable memory and includes queue functions for maintaining the queues. Queue access is provided in a first-come/first-served basis. In addition to the get and put queue functions, queue control within the storage control saves a queue item which is read from the queue in a location in the addressable memory which is associated with the processor making a get request, thereby alleviating the requesting processor from having to save the queue item.

Fault tolerant apparatus and method for maintaining one or more queues

A memory system providing capability for correction of multiple bit errors. The storage elements of the memory system are divided into four-bit nibbles, wherein storage of a single 32-word requires access to eight separate storage elements. A ninth storage element stores a four-bit error syndrome. All nine storage elements have single bit error correction/multiple bit error detection. All single bit errors are corrected directly within the individual storage element. Multiple bit errors within a single storage element are signaled to the interface controller which corrects the error using the stored four-bit error syndrome.

Growing community of inventors

Minneapolis, MN, United States of America

Eugene A Rodi