Abstract: An accelerator 120 is tightly coupled to the normal execution unit 110. The operand store, which could be a register file 130, a stack based operand store or other operand store is shared by the execution unit and the accelerator unit. Operands may also be accessed as immediate values within the instructions themselves. The sequences of individual program instructions corresponding to computational subgraphs remain within a program but can be recognized by the accelerator as suitable for acceleration and when encountered are executed by the accelerator instead of by the normal execution unit. Within such tightly coupled arrangement problems can arise due to a lack of register resources within the system. The present technique provides that at least some intermediate operand values which are generated within the accelerator, but are determined not to be referenced outside of the computational subgraph concerned, are not written to the operand store.

Abstract: An automatic and retargetable computer design system is using a combination of simulation and performance prediction to investigate a plurality of target computer systems. A high-level specification and a predetermined application are used by the computer design system to provide inputs into a computer evaluator. The computer evaluator has reference computer system dependent and independent systems for producing a reference representation and dynamic behavior information, respectively, of the application. The reference representation and information are mated to produce further information to drive a simulator. The simulator provides performance information of a reference computer system. The performance information is provided to another computer evaluator, which has a target computer system dependent system for producing a target representation of the application for the plurality of target computer systems.

Abstract: A system is provided which simplifies and speeds up the process of designing a computer system by evaluating the components of the memory hierarchy for any member of a broad family of processors in an application-specific manner. The system uses traces produced by a reference processor in the design space for a particular cache design and characterizes the differences in behavior between the reference processor and an arbitrarily chosen processor. The differences are characterized as a series of dilation parameters which relate to how much the traces would expand because of the substitution of a target processor. In addition, the system characterizes the reference trace using a set of trace parameters that are part of a cache behavior model. The dilation and trace parameters are used to determine the factors for estimating the performance statistics of target processors with specific memory hierarchies.

Abstract: An automated design system for VLIW processors explores a parameterized design space to assist in identifying candidate processor designs that satisfy desired design constraints, such as processor cost and performance. A VLIW synthesis process takes as input a specification of processor parameters and synthesizes a datapath specification, an instruction format design, and a control path specification. The synthesis process also extracts a machine description suitable to re-target a compiler. The re-targeted compiler generates operation issue statistics for an application program or set of programs. Using these statistics, a procedure for searching the design space can extract internal resources utilization information that is used to determine new candidate processors for evaluation.

Abstract: A compiler includes a branch statistics data analyzer to analyze branch statistics data of a branch instruction to construct a branch predictor function for the branch instruction. A branch prediction instruction generator is coupled to the branch statistics data analyzer to generate at least one prediction instruction to implement the branch predictor function. A main compiling engine is coupled to the branch prediction instruction generator to insert the prediction instruction before the branch instruction. A method of dynamically predicting a branch instruction of a program is also described.