Abstract: Techniques and structures are described which allow the detection of certain dependency conditions, including evil twin conditions, during the execution of computer instructions. Information used to detect dependencies may be stored in a logical map table, which may include a content-addressable memory. The logical map table may maintain a logical register to physical register mapping, including entries dedicated to physical registers available as rename registers. In one embodiment, each entry in the logical map table includes a first value usable to indicate whether only a portion of the physical register is valid and whether the physical register includes the most recent update to the logical register being renamed. Use of this first value may allow precise detection of dependency conditions, including evil twin conditions, upon an instruction reading from at least two portions of a logical register having an entry in the logical map table whose first value is set.

Abstract: Techniques and systems are described herein to maintain a mapping of logical to physical registers—for example, in the context of a multithreaded processor that supports renaming. A mapping unit may have a plurality of entries, each of which stores rename information for a dedicated one of a set of physical registers available to the processor for renaming. This physically-indexed mapping unit may support multiple threads, and may comprise a content-addressable memory (CAM) in certain embodiments. The mapping unit may support various combinations of read operations (to determine if a logical register is mapped to a physical register), write operations (to create or modify one or more entries containing mapping information), thread flush operations, and commit operations. More than one of such operations may be performed substantially simultaneously in certain embodiments.

Abstract: Various techniques for mitigating dependencies between groups of instructions are disclosed. In one embodiment, such dependencies include “evil twin” conditions, in which a first floating-point instruction has as a destination a first portion of a logical floating-point register (e.g., a single-precision write), and in which a second, subsequent floating-point instruction has as a source the first portion and a second portion of the same logical floating-point register (e.g., a double-precision read). The disclosed techniques may be applicable in a multithreaded processor implementing register renaming. In one embodiment, a processor may enter an operating mode in which detection of evil twin “producers” (e.g., single-precision writes) causes the instruction sequence to be modified to break potential dependencies. Modification of the instruction sequence may continue until one or more exit criteria are reached (e.g., committing a predetermined number of single-precision writes).

Abstract: In one embodiment, a processor comprises a cache and a cache miss unit coupled to the cache. The cache miss unit is configured to initiate a cache fill of a cache line for the cache responsive to a first cache miss in the cache, wherein the first cache miss corresponds to a first thread of a plurality of threads in execution by the processor. Furthermore, the cache miss unit is configured to record an additional cache miss corresponding to a second thread of the plurality of threads, wherein the additional cache miss occurs in the cache prior to the cache fill completing for the cache line. The cache miss unit is configured to inhibit initiating an additional cache fill responsive to the additional cache miss.

Abstract: In one embodiment, a processor comprises a plurality of instruction buffers, an instruction cache coupled to supply instructions to the plurality of instruction buffers, and a cache miss unit coupled to the instruction cache. Each of the plurality of instruction buffers is configured to store instructions fetched from a respective thread of a plurality of threads. The cache miss unit is configured to monitor cache misses in the instruction cache. Particularly, the cache miss unit is configured to detect which of the plurality of threads experience a cache miss to a cache line. Responsive to a return of the cache line for storage in the instruction cache, the cache miss unit is configured to concurrently cause at least one instruction from the cache line to be stored in each of the plurality of instruction buffers that corresponds to one of the plurality of threads which experienced the cache miss to the cache line.

Abstract: In one embodiment, a processor comprises a cache shared by a plurality of threads in execution by the processor, an error detection unit coupled to the cache, and a fetch control unit. The error detection unit is configured to detect an error in data output by the cache responsive to an access corresponding to a first thread of a plurality of threads. Coupled to receive an indication of the error, the fetch control unit is configured to inhibit fetching for the first thread responsive to the error until the thread is redirected in response to the error and until the error is eliminated from the cache that includes the data.

Abstract: In one embodiment, a processor comprises a plurality of pipeline stages and a first circuit operable at a first pipeline stage of the plurality of pipeline stages. The first circuit is configured to maintain a plurality of program counters (PCs), each of which corresponds to one of a plurality of threads that the processor is configured to have concurrently in process with respect to the plurality of pipeline stages. The first circuit is configured to provide a first PC to a second pipeline stage of the plurality of pipeline stages. The first PC is derived from one of the plurality of PCs that corresponds to a first thread of the plurality of threads, and a first instruction entering the second pipeline stage is from the first thread.

Abstract: In one embodiment, a processor comprises an instruction cache and a fetch generator circuit coupled thereto. The fetch generator circuit is configured to generate at least one fetch request to the instruction cache for at least one of the plurality of threads. The fetch generator circuit is also configured to monitor for a plurality of conditions for each thread, wherein each of the plurality of conditions defined to inhibit the thread from being fetched. The fetch generator circuit is configured to speculatively generate a first fetch request for a first thread of the plurality of threads if each thread is inhibited from fetching and the first thread is inhibited from fetching only due to a first predetermined condition of the plurality of conditions. In one particular implementation, the first predetermined condition is a lack of room in a corresponding one of a plurality of instruction buffers.