Abstract: A system, apparatus, and method are disclosed for increasing the physical memory size accessible to a processor, at least in part, by reclaiming physical address space typically associated with addresses of a restricted linear address space (i.e., addresses that are otherwise unusable by the processor as system memory). In one embodiment, an exemplary memory controller redirects a linear address associated with a range of addresses to access a reclaimed memory hole. The memory controller includes an address translator configured to determine an amount of restricted addresses and to establish a baseline address identified as a first number being a first integer power of 2. The range of addresses can be located at another address identified as a second number being a second integer power of 2. As such, the address translator translates the linear address into a translated address associated with the reclaimed memory hole based on the baseline address.

Abstract: A system, apparatus, and method are disclosed for controlling accesses into memory to minimize sequential accesses to the same bank of memory, at least in part, by characterizing a subset of an address in parallel with address translations associated with those accesses. In one embodiment, an exemplary memory controller can include an address translator configured to translate an address useable by a processor to a first memory address. Also, the memory controller includes a bit characterizer configured to characterize a subset of the address as having a value from a range of values, and a bank separator coupled to the address translator and the bit characterizer for receiving a first portion of the first memory address and the value, respectively. Accordingly, the bank separator is configured to differentiate the first portion from a second portion of a second memory address.

Abstract: A system, apparatus, and method are disclosed for storing predictions as well as examining and using one or more caches for anticipating accesses to a memory. In one embodiment, an exemplary apparatus is a prefetcher for managing predictive accesses with a memory. The prefetcher can include a speculator to generate a range of predictions, and multiple caches. For example, the prefetcher can include a first cache and a second cache to store predictions. An entry of the first cache is addressable by a first representation of an address from the range of predictions, whereas an entry of the second cache is addressable by a second representation of the address. The first and the second representations are compared in parallel against the stored predictions of either the first cache and the second cache, or both.

Abstract: A system, apparatus, and method are disclosed for increasing the physical memory address space accessible to a processor, at least in part, by translating linear addresses associated with a memory hole into a subset of physical memory addresses that otherwise is inaccessible as system memory by a processor. In one embodiment, a memory controller reclaims memory holes in a system memory divided into ranges of linear addresses, where the system memory includes a number of arbitrarily-sized memory devices. The memory controller includes a memory configuration evaluator configured to determine a translated memory hole size for a memory hole, the memory hole including restricted linear addresses that translate into a subset of physical addresses. Also, memory configuration evaluator can be configured to form adjusted ranges to translate at least one linear address into a subset of physical addresses. As such, the system memory increases by at least the subset of physical addresses.

Abstract: A system, apparatus, and method are disclosed for predicting accesses to memory. In one embodiment, an exemplary apparatus comprises a processor configured to execute program instructions and process program data, a memory including the program instructions and the program data, and a memory processor. The memory processor can include a speculator configured to receive an address containing the program instructions or the program data. Such a speculator can comprise a sequential predictor for generating a configurable number of sequential addresses. The speculator can also include a nonsequential predictor configured to associate a subset of addresses to the address and to predict a group of addresses based on at least one address of the subset, wherein at least one address of the subset is unpatternable to the address.

Abstract: A method and system for memory temperature measurement. The method includes the step of monitoring a plurality of accesses to a memory component. A number of accesses occurring to the memory component over a time period is determined. A temperature of the memory component is determined based on the number of accesses occurring over the time period.

Abstract: Embodiments of the invention accelerate at least one special purpose processor, such as a GPU, or a driver managing a special purpose processor, by using at least one co-processor. Advantageously, embodiments of the invention are fault-tolerant in that the at least one GPU or other special purpose processor is able to execute all computations, although perhaps at a lower level of performance, if the at least one co-processor is rendered inoperable. The co-processor may also be used selectively, based on performance considerations.

Abstract: A bus interface unit is adapted to receive transaction requests for at least two different targets. The bus interface unit monitors a capacity of a resource associated with servicing transaction requests to the targets, such as a posted write buffer. If a transaction request would fill the resource beyond a current remaining capacity of the resource such that the execution of other pipelined transactions would become stalled, the bus interface generates a retry response so that the request is retried at a later time, permitting other transactions to proceed while the resource drains.