Intel Corporation

Josep M Codina

Fernando Latorre

Pedro Lopez

Enric Gibert Codina

Raul Martinez

Antonio Gonzalez

Carlos Madriles

Grigorios Magklis

Pedro Marcuello

Enric Gibert

Antonio González

Marc Lupon

Polychronis Xekalakis

Daniel Ortega

Kyriakos A Stavrou

Demos Pavlou

Georgios Tournavitis

Mirem Hyuseinova

Christos E Kotselidis

Crispin Gomez Requena

Alejandro Martinez Vincente

Alejandro Martinez Vicente

Alejandro Vicente Martinez

Ravishankar Iyer

Donald K Newell

Srihari Makineni

Jaume Abella

Jaideep Moses

Tanausu Ramirez

Jose-Alejandro Pineiro

Ramesh G Illikkal

Carlos Madriles Gimeno

F Jesús Sánchez

José Alejandro Piñeiro

Pedro López

Antonio Gonzelez

Marti Torrents Lapuerta

Rekai Gonzalez-Alberquilla

Systems, methods, and apparatuses for decomposing a sequential program into multiple threads, executing these threads, and reconstructing the sequential execution of the threads are described. A plurality of data cache units (DCUs) store locally retired instructions of speculatively executed threads. A merging level cache (MLC) merges data from the lines of the DCUs. An inter-core memory coherency module (ICMC) globally retires instructions of the speculatively executed threads in the MLC.

Merging level cache and data cache units having indicator bits related to speculative execution

A computer-readable storage medium, method and system for optimization-level aware branch prediction is described. A gear level is assigned to a set of application instructions that have been optimized. The gear level is also stored in a register of a branch prediction unit of a processor. Branch prediction is then performed by the processor based upon the gear level.

Instruction and logic for optimization level aware branch prediction

A combination of hardware and software collect profile data for asynchronous events, at code region granularity. An exemplary embodiment is directed to collecting metrics for prefetching events, which are asynchronous in nature. Instructions that belong to a code region are identified using one of several alternative techniques, causing a profile bit to be set for the instruction, as a marker. Each line of a data block that is prefetched is similarly marked. Events corresponding to the profile data being collected and resulting from instructions within the code region are then identified. Each time that one of the different types of events is identified, a corresponding counter is incremented. Following execution of the instructions within the code region, the profile data accumulated in the counters are collected, and the counters are reset for use with a new code region.

Propagating a prefetching profile bit from a prefetch queue to a data cache to indicate that a line was prefetched in response to an instruction within a code region

An apparatus and method are described for improved thread selection. For example, one embodiment of a processor comprises: first logic to maintain a history table comprising a plurality of entries, each entry in the table associated with an instruction and including history data indicating prior hits and/or misses to a cache level and/or a translation lookaside buffer (TLB) for that instruction; and second logic to select a particular thread for execution at a particular processor pipeline stage based on the history data.

Method and apparatus for improved thread selection

Methods and apparatus are disclosed for using a register checkpointing mechanism to resolve multithreading mis-speculations. Valid architectural state is recovered and execution is rolled back. Some embodiments include memory to store checkpoint data. Multiple thread units concurrently execute threads. They execute a checkpoint mask instruction to initialize memory to store active checkpoint data including register contents and a checkpoint mask indicating the validity of stored register contents. As register contents change, threads execute checkpoint write instructions to store register contents and update the checkpoint mask. Threads also execute a recovery function instruction to store a pointer to a checkpoint recovery function, and in response to mis-speculation among the threads, branch to the checkpoint recovery function. Threads then execute one or more checkpoint read instructions to copy data from a valid checkpoint storage area into the registers necessary to recover a valid architectural state, from which execution may resume.

Utilization of register checkpointing mechanism with pointer swapping to resolve multithreading mis-speculations

Disclosed is an apparatus and method to manage instruction cache prefetching from an instruction cache. A processor may comprise: a prefetch engine; a branch prediction engine to predict the outcome of a branch; and dynamic optimizer. The dynamic optimizer may be used to control: identifying common instruction cache misses and inserting a prefetch instruction from the prefetch engine to the instruction cache.

Managed instruction cache prefetching

Systems, methods, and apparatuses for decomposing a sequential program into multiple threads, executing these threads, and reconstructing the sequential execution of the threads are described. A plurality of data cache units (DCUs) store locally retired instructions of speculatively executed threads. A merging level cache (MLC) merges data from the lines of the DCUs. An inter-core memory coherency module (ICMC) globally retire instructions of the speculatively executed threads in the MLC.

Systems, methods, and apparatuses to decompose a sequential program into multiple threads, execute said threads, and reconstruct the sequential execution

An apparatus and method for profiling program code. In particular, an apparatus according to one embodiment comprises a filtering component identifying a first set of instructions for which profiling is desired wherein, in response to detecting that an instruction has been retired, the filtering component determines whether the instruction is within the first set of instructions for which profiling is desired; an event selection component detecting an event in response to the instruction retiring, the event selection component generating event signals in response to a designated event; and a profiling component recording the occurrence or not occurrence of the event within a first storage device responsive to signals from the filtering component and/or the event selection component.

Method and apparatus for flexible, accurate, and/or efficient code profiling

Methods and apparatus relating to a replacement policy for hot code detection are described. In some embodiments, it may be determined which entry amongst a plurality of entries stored in storage unit is to be replaced next. The entries may correspond to hot code and may store age and execution frequency information corresponding to the hot code. Other embodiments are also described and claimed.

Replacement policy for hot code detection

In one embodiment, the present invention includes a method for accessing registers associated with a first thread while executing a second thread. In one such embodiment a method may include preventing an instruction of a first thread that is to access a source operand from a register file of a second thread from executing if a synchronization indicator associated with the source operand indicates incompletion of a producer operation of the second thread, and executing the instruction if the synchronization indicator indicates completion of the producer operation of the second thread. Other embodiments are described and claimed.

Access of register files of other threads using synchronization

Techniques for achieving coherence between dynamically optimized code and original code are disclosed. In an embodiment, a search is performed for a first entry for a first page containing a first code region in a first data structure. The first code is used to determine whether a first indicator in the first entry is set to a first value. The first entry is added to the first data structure, in response to failing to find the first entry in the first data structure. A second search may be performed for a second entry for the first code region in a second data structure, in response to determining that the first indicator is set to the first value. Other embodiments are also disclosed and claimed.

Achieving coherence between dynamically optimized code and original code

Apparatus and computing systems associated with cache sharing based thread control are described. One embodiment includes a memory to store a thread control instruction and a processor to execute the thread control instruction. The processor is coupled to the memory. The processor includes a first unit to dynamically determine a cache sharing behavior between threads in a multi-threaded computing system and a second unit to dynamically control the composition of a set of threads in the multi-threaded computing system. The composition of the set of threads is based, at least in part, on thread affinity as exhibited by cache-sharing behavior. The thread control instruction controls the operation of the first unit and the second unit.

Cache sharing based thread control

Growing community of inventors

Hospitalet de Llobregat, Spain

Josep M Codina