Abstract: A method includes determining, for a first thread of execution, a first speculative decoded operands signal and determining, for a second thread of execution, a second speculative decoded operands signal. The method further includes determining, for the first thread of execution, a first constant and determining, for the second thread of execution, a second constant. The method further compares the first speculative decoded operands signal to the second speculative decoded operands signal and uses the first and second constant to detect a wordline collision for accessing the memory array.

Abstract: A pipeline circuit determines a first effective address based a sum of a first value and a second value. The first effective address is based upon an actual value of a carry-in into a bit-wise region of the first and second values. The bit-wise region includes a predefined internal region of bits of the first and second values. The pipeline circuit also determines a second effective address based a sum of a third value and a fourth value. A collision detector circuit receives bits from the bit-wise region of each of the four values and determines a plurality of speculative results based upon the bits of the bit-wise regions and based upon a plurality of speculative carry-in values. A collision indicator is asserted based on at least one result of the plurality of speculative results, and the actual values of the first and second carry-in.

Abstract: A cache management system employs a replacement policy in a manner that manages concurrent accesses to cache. The cache management system comprises a cache, a replacement policy storage for storing replacement statuses of cache lines of the cache, and an update module. The update module, comprising access filtering and a concurrent update handling, determines how updates to the replacement policy storage are handled. In a multi-threaded compute environment, a concurrent access to shared cache causes a selective update to the replacement policy storage.

Abstract: A pipeline circuit determines a first effective address based a sum of a first value and a second value. The first effective address is based upon an actual value of a carry-in into a bit-wise region of the first and second values. The bit-wise region includes a predefined internal region of bits of the first and second values. The pipeline circuit also determines a second effective address based a sum of a third value and a fourth value. A collision detector circuit receives bits from the bit-wise region of each of the four values and determines a plurality of speculative results based upon the bits of the bit-wise regions and based upon a plurality of speculative carry-in values. A collision indicator is asserted based on at least one result of the plurality of speculative results, and the actual values of the first and second carry-in.

Abstract: A method includes determining, for a first thread of execution, a first speculative decoded operands signal and determining, for a second thread of execution, a second speculative decoded operands signal. The method further includes determining, for the first thread of execution, a first constant and determining, for the second thread of execution, a second constant. The method further compares the first speculative decoded operands signal to the second speculative decoded operands signal and uses the first and second constant to detect a wordline collision for accessing the memory array.

Abstract: A cache management system employs a replacement policy in a manner that manages concurrent accesses to cache. The cache management system comprises a cache, a replacement policy storage for storing replacement statuses of cache lines of the cache, and an update module. The update module, comprising access filtering and a concurrent update handling, determines how updates to the replacement policy storage are handled. In a multi-threaded compute environment, a concurrent access to shared cache causes a selective update to the replacement policy storage.

Abstract: Each cacheline of a unified cache storing information is marked as incoherent if the information was acquired incoherently or marked as coherent if the information was acquired coherently. A subsequent incoherent read access to a cacheline can result in a cache hit and a return of the cached information regardless of whether the cacheline is marked as coherent or incoherent. However, a subsequent coherent read access to a cacheline marked as incoherent will be returned as a cache miss regardless of whether the cacheline includes information sought by the coherent read access. In response to a cache miss for a coherent read access, a global snoop is initiated so as to query all other target components within the same coherency domain. In contrast, a cache miss resulting from an incoherent read access is processed using a non-global snoop to a limited set of one or a few target components in the coherency domain.

Abstract: Each cacheline of a unified cache storing information is marked as incoherent if the information was acquired incoherently or marked as coherent if the information was acquired coherently. A subsequent incoherent read access to a cacheline can result in a cache hit and a return of the cached information regardless of whether the cacheline is marked as coherent or incoherent. However, a subsequent coherent read access to a cacheline marked as incoherent will be returned as a cache miss regardless of whether the cacheline includes information sought by the coherent read access. In response to a cache miss for a coherent read access, a global snoop is initiated so as to query all other target components within the same coherency domain. In contrast, a cache miss resulting from an incoherent read access is processed using a non-global snoop to a limited set of one or a few target components in the coherency domain.

Abstract: A method and device for locking a cache line of a cache is disclosed. The method includes automatically changing a state of a cache line from a valid locked state to an invalid locked state in response to receiving an indication that a memory location external to the cache and corresponding to the cache line is associated with an access request by a processor or other data access module. Thus, the locked state of a cache line is maintained even after data in the locked cache line is invalidated. By maintaining the invalid locked state, the cache line is not available for reallocation by the cache. This allows locked cache lines that become invalidated to remain locked without additional software overhead to periodically determine whether the lock has been lost due to invalidation of the cache line.

Abstract: A method and device for locking a cache line of a cache is disclosed. The method includes automatically changing a state of a cache line from a valid locked state to an invalid locked state in response to receiving an indication that a memory location external to the cache and corresponding to the cache line is associated with an access request by a processor or other data access module. Thus, the locked state of a cache line is maintained even after data in the locked cache line is invalidated. By maintaining the invalid locked state, the cache line is not available for reallocation by the cache. This allows locked cache lines that become invalidated to remain locked without additional software overhead to periodically determine whether the lock has been lost due to invalidation of the cache line.