Abstract: An apparatus and method is disclosed for flushing a cache in a computing system. In a multinode computing system a cache in a first node may contain modified data in an address space of a second node. The cache in the first node must be purged prior to shutting down the first node. The computing system uses a random class replacement scheme for the cache. A cache flush routine sets a cache flush mode in a class replace select mechanism, overriding the random class replacement scheme. With the random class replacement scheme overridden, a minimum number of fetches will flush all the cache lines in the cache, each fetch loading the cache with a cache line not already in the cache. No additional delay penalty is incurred in a critical path through which fetches and stores to the cache must pass.

Abstract: A method and apparatus implement redundant memory access using multiple controllers on the same bank of memory. A first memory controller uses the memory as its primary address space, for storage and fetches. A second redundant controller is also connected to the same memory. System control logic is used to notify the redundant controller of the need to take over the memory interface. The redundant controller initializes if required and takes control of the memory. The memory only needs to be initialized if the system has to be brought down and restarted in the redundant mode. This invention allows the system to continue to stay up and continue running during a memory controller or link failure.

Abstract: A method and apparatus implement redundant memory access using multiple controllers on the same bank of memory, and a design structure on which the subject circuit resides is provided. A first memory controller uses the memory as its primary address space, for storage and fetches. A second redundant controller is also connected to the same memory. System control logic is used to notify the redundant controller of the need to take over the memory interface. The redundant controller initializes if required and takes control of the memory. The memory only needs to be initialized if the system has to be brought down and restarted in the redundant mode. This invention allows the system to continue to stay up and continue running during a memory controller or link failure.

Abstract: An apparatus includes a virtual memory manager that moves data from a first block to second block in memory. When the virtual memory manager is ready to transfer data from the first block to the second block, a third, temporary block of memory is defined. The translation table in a DMA controller is changed to point DMA transfers that target the first block to instead target the temporary block. The virtual memory manager then transfers data from the first block to the second block. When the transfer is complete, a check is made to see if the DMA transferred data to the temporary block while the data from the first block was being written to the second block. If so, the data written to the temporary block is written to the second block. A hardware register is preferably used to efficiently detect changes to the temporary block.

Abstract: A memory controller provides page copy logic that assures data coherency when a DMA operation to a page occurs during the copying of the page by the memory controller. The page copy logic compares the page index of the DMA operation to a copy address pointer that indicates the location currently being copied. If the page index of the DMA operation is less than the copy address pointer, the portion of the page that would be written to by the DMA operation has already been copied, so the DMA operation is performed to the physical address of the new page. If the page index of the DMA operation is greater than the copy address pointer, the portion of the page that would be written to by the DMA operation has not yet been copied, so the DMA operation is performed to the physical address of the old page.

Abstract: An apparatus and method is disclosed for providing capacity on demand using control to alter latency and/or bandwidth on a signaling bus in a computer system. If additional capacity is required, authorization is requested for additional capacity. If authorized, bandwidth of the signaling bus is increased to provide additional capacity in the computing system. Alternatively, upon authorization, latency of data transmissions over the signaling bus is reduced. In another alternative, upon authorization, memory timings are adjusted to speed up memory fetches and stores.

Abstract: Methods and apparatus are provided that allow an electronic system having a signaling bus with a fault on a signaling conductor to operate at a degraded performance. A block of data is transferred from a first electronic unit to a second electronic unit over the signaling bus. A transmission sequence sends the block of data using all of the nonfaulty signaling conductors using a minimum number of beats required to complete the transmission.

Abstract: Multiple memory channels of a multi-channel memory architecture are effectively bridged together to enable data traffic associated with various nodes in daisy chain arrangement to be communicated over both memory channels. For example, a daisy chain arrangement of nodes, such as FB-DIMM memory modules disposed in a first memory channel may be coupled to a second memory channel, with support for communicating data associated with one of the nodes over either or both of the first and second memory channels.

Abstract: A memory chip having a data bus having a plurality of bits. The number of bits is apportioned between a read portion and a write portion. The write portion is dedicated to receiving data that is to be written into an array on the memory chip; the read portion is dedicated to driving data that has been read from the array on the memory chip. The apportionment is programmable. Apportionment can be specified by programming signal pins on the memory chip, connecting the signal pins to appropriate logical values. The apportionment can alternatively be specified by scanning apportionment information into the memory chip at bring up time. The apportionment and also alternatively be specified by receiving apportionment information in an address/command word.

Abstract: A memory system having a memory controller and a daisy chain of memory chips. The memory controller is coupled to memory chips in the daisy chain of memory chips by an address/command bus chain. The memory controller is coupled to memory chips in the daisy chain of memory chips by a data bus chain having a number of data bus bits. The data bus chain has a first portion of data bus bits dedicated to transmitting write data from the memory controller to a memory chip. The data bus chain has a second portion of data bus bits dedicated to transmitting read data from a memory chip to the memory controller. Apportionment of data bus bits between the first portion and the second portion is programmable. Programming is done by pin connection, scanning of a value, or by request from a processor coupled to the memory controller.

Abstract: A self timed memory chip having an apportionable data bus. Access timing to an array on the memory chip is dynamically determined by circuitry on the memory chip. A ring oscillator on the memory chip has a frequency that is indicative of how fast an array on the memory chip can be accessed. The ring oscillator includes a bit line that is periodically charged and a memory element that subsequently discharges the bit line. The memory chip has a data bus interface having a number of bits. The data bus interface has a first number of bits apportioned to write data and a second number of bits apportioned to read data. The first number of bits and the second number of bits is programmable.

Abstract: A memory system having a data bus coupling a memory controller and a memory. The data bus has a number of data bus bits. The data bus is programmably apportioned to a first portion dedicated to transmitting data from the memory controller to the memory and a second portion dedicated to transmitting data from the memory to the memory controller. The apportionment can be assigned by suitable connection of pins on a memory chip in the memory and the memory controller to logical values. Alternatively, the apportionment can be scanned into the memory controller and the memory at bring up time. In another alternative, the apportionment can be changed by suspending data transfer and dynamically changing the sizes of the first portion and the second portion.

Abstract: A method and structure are provided for implementing enhanced cooling of a plurality of memory devices. The memory structure includes a stack of platters. A sub-plurality of memory devices is mounted on each platter. At least one connector is provided with each platter for connecting to the sub-plurality of memory devices. A heat sink is associated with the stack of platters for cooling the plurality of memory devices.

Abstract: A circuit arrangement, method and apparatus utilize communication links that are selectively configurable to operate in both unidirectional and bidirectional modes to communicate data between multiple nodes that are interconnected to one another in a daisy chain configuration. As a result, in many instances communications may be maintained with nodes located both before and after a discontinuity in a daisy chain configuration.

Abstract: Methods, apparatus, and program product are disclosed for use in a computer system to provide for dynamic allocation of a directory memory in a node memory controller in which one or more coherent multiprocessor nodes comprise the computer system. The directory memory in a node is partitioned between a snoop directory portion and a remote memory directory portion. During a predetermined time interval, snoop directory entry refills and remote memory directory entry refills are accumulated. After the time interval has elapsed, a ratio of the snoop directory entry refills to the number of remote memory directory entry refills is computed. The ratio is compared to a desired ratio. Respondent to a difference between the ratio and the desired ratio, adjustments are made to the allocation of the memory directory between the snoop directory and the remote memory directory.

Abstract: Methods, apparatus, and program product are disclosed for use in a computer system in which one or more multiprocessor nodes comprise the computer system. The methods and apparatus provide for configurable allocation of a memory in a node memory controller. In a single node implementation of the computer system, substantially all of the memory is allocated to a snoop directory used to store directory entries for cache lines used by processors in the node. In computer system implementations having more than one node, the amount of the memory allocated to the snoop directory and the amount of the memory allocated to a remote memory directory is controlled respondent to predetermined sizes respondent to the number of nodes in the computer system.

Abstract: A method and apparatus is provided which dynamically alters SDRAM memory interface timings to provide minimum read access latencies for different types of memory accesses in a memory subsystem of a computer system. The dynamic alteration of the SDRAM memory interface timings is based on workload and is determined with information from the memory controller read queue.

Abstract: A circuit arrangement and method are used in connection with a data latch that is coupled to a data source over a source synchronous communications interface to disable the data latch from latching data whenever the data source is not driving the source synchronous data strobe signal. As such, when the data source is not driving the source synchronous data strobe signal, undesired and/or inadvertent latching by the data latch can be avoided. Moreover, in implementations where a data strobe signal line is bidirectional, and capable of being driven either by the data source or by another circuit used to access the data source (e.g., a memory controller), disabling data latching as described herein can minimize the risk of driver damage resulting from conflicting attempts to drive the data strobe signal line at both ends.

Abstract: Embodiments are provided in which cache updating is described for a computer system having at least a first processor and a second processor having a first cache and a second cache, respectively. When the second processor obtains from the first processor a lock to a shared memory region, the first cache pushes to the second cache cache lines for the addresses in the shared memory region accessed by the first processor while the first processor had the lock.

Abstract: Methods and apparatus in a computer system are disclosed for providing a memory controller featuring a dedicated bank sequencer for each memory bank in a memory system. Each bank sequencer controls the dispatch of load and store requests to a central controller such that each request sent to the central controller can be served by the associated memory bank at the time that the central controller receives the request. Since every request received by the central controller is valid from a bank timing standpoint, the central controller is free to process the requests from a predetermined priority basis, without concern for bank availability. This significantly improves the design of the memory controller in the processing system.