Abstract: A Very Long Instruction Word (VLIW) processor having a plurality of functional units includes a multi-ported register file that is divided into a plurality of separate register file segments, each of the register file segments being associated to one of the plurality of functional units. The register file segments are partitioned into local registers and global registers. The global registers are read and written by all functional units. The local registers are read and written only by a functional unit associated with a particular register file segment. The local registers and global registers are addressed using register addresses in an address space that is separately defined for a register file segment/functional unit pair. The global registers are addressed within a selected global register range using the same register addresses for the plurality of register file segment/functional unit pairs.

Abstract: One embodiment of the present invention provides a system that facilitates deferring execution of instructions with unresolved data dependencies as they are issued for execution in program order. During a normal execution mode, the system issues instructions for execution in program order. Upon encountering an unresolved data dependency during execution of an instruction, the system generates a checkpoint that can subsequently be used to return execution of the program to the point of the instruction. Next, the system executes subsequent instructions in an execute-ahead mode, wherein instructions that cannot be executed because of an unresolved data dependency are deferred, and wherein other non-deferred instructions are executed in program order.

Abstract: One embodiment of the present invention provides a system which creates multiple checkpoints in a processor that supports speculative-execution. The system starts by issuing instructions for execution in program order during execution of a program in a normal-execution mode. Upon encountering a launch condition during an instruction which causes a processor to enter execute-ahead mode, the system performs an initial checkpoint and commences execution of instructions in execute-ahead mode. Upon encountering a predefined condition during execute-ahead mode, the system generates an additional checkpoint and continues to execute instructions in execute-ahead mode. Generating the additional checkpoint allows the processor to return to the additional checkpoint, instead of the previous checkpoint, if the processor subsequently encounters a condition that requires the processor to return to a checkpoint.

Abstract: One embodiment of the present invention provides a system which performs simultaneous speculative threading. The system starts by executing instructions in normal execution mode using a first thread. Upon encountering a data-dependent stall condition, the first thread generates an architectural checkpoint and commences execution of instructions in execute-ahead mode. During execute-ahead mode, the first thread executes instructions that can be executed and defers instructions that cannot be executed into a deferred queue. When the data dependent stall condition has been resolved, the first thread generates a speculative checkpoint and continues execution in execute-ahead mode. At the same time, the second thread commences execution in a deferred mode, wherein the second thread executes instructions deferred by the first thread.

Abstract: One embodiment of the present invention provides a system that facilitates selectively unmarking load-marked cache lines during transactional program execution, wherein load-marked cache lines are monitored during transactional execution to detect interfering accesses from other threads. During operation, the system encounters a release instruction during transactional execution of a block of instructions. In response to the release instruction, the system modifies the state of cache lines, which are specially load-marked to indicate they can be released from monitoring, to account for the release instruction being encountered. In doing so, the system can potentially cause the specially load-marked cache lines to become unmarked. In a variation on this embodiment, upon encountering a commit-and-start-new-transaction instruction, the system modifies load-marked cache lines to account for the commit-and-start-new-transaction instruction being encountered.

Abstract: One embodiment of the present invention provides a system that facilitates selectively unmarking load-marked cache lines during transactional program execution, wherein load-marked cache lines are monitored during transactional execution to detect interfering accesses from other threads. During operation, the system encounters a release instruction during transactional execution of a block of instructions. In response to the release instruction, the system modifies the state of cache lines, which are specially load-marked to indicate they can be released from monitoring, to account for the release instruction being encountered. In doing so, the system can potentially cause the specially load-marked cache lines to become unmarked. In a variation on this embodiment, upon encountering a commit-and-start-new-transaction instruction, the system modifies load-marked cache lines to account for the commit-and-start-new-transaction instruction being encountered.

Abstract: One embodiment of the present invention provides a system that predicts a result produced by a section of code in order to support speculative program execution. The system begins by executing the section of code using a head thread in order to produce a result. Before the head thread produces the result, the system generates a predicted result to be used in place of the result. Next, the system allows a speculative thread to use the predicted result in speculatively executing subsequent code that follows the section of code. After the head thread finishes executing the section of code, the system determines if a difference between the predicted result and the result generated by the head thread has affected execution of the speculative thread. If so, the system executes the subsequent code again using the result generated by the head thread. If not, the system performs a join operation to merge state associated with the speculative thread with state associated with the head thread.

Abstract: One embodiment of the present invention provides a system that facilitates interleaved execution of a head thread and a speculative thread within a single processor pipeline. The system operates by executing program instructions using the head thread, and by speculatively executing program instructions in advance of the head thread using the speculative thread, wherein the head thread and the speculative thread execute concurrently through time-multiplexed interleaving in the single processor pipeline.

Abstract: A graphics system and method for increasing efficiency of decompressing blocks of compressed geometry data and reducing redundant transformation and lighting calculations is disclosed. Multiple decompression pipelines are used to increases the decompression speed. A control unit receives blocks of compressed geometry data information and selectively routes them to a plurality of decompression pipelines. Each decompression pipeline is configured to decompress the blocks into a set of vertices. The reduction in redundant calculations is accomplished by delaying the formation of geometric primitives until after transformation and lighting has been performed on the vertices. Transformation and/or lighting are performed independently on a vertex-by-vertex basis without reference to which geometric primitives the vertices belong to. After transformation and or lighting, geometric primitives may be formed utilizing previously generated connectivity information.

Abstract: Object simulation and interaction of and between computer-generated or graphical objects in a virtual space includes neutral scene graphs, data structures and procedures for using such graphs and data structures.

Abstract: A system that facilitates error correction within a register file in a central processing unit (CPU). During execution of an instruction by the CPU, the system retrieves a dataword and an associated syndrome from a source register in the register file. Next, the system uses information in the dataword and the associated syndrome to detect, and if necessary correct, an error in the dataword or associated syndrome. This error detection and correction takes place in parallel with using the dataword to perform a computational operation specified by the instruction. If an error is detected, the system prevents the instruction from performing a writeback to a destination register in the register file. The system also writes a corrected dataword to the source register in the register file. Next, the system flushes the instruction pipeline, and restarts execution of the instruction so that the corrected dataword is retrieved for the computational operation.

Abstract: One embodiment of the present invention provides a system that predicts a result produced by a section of code in order to support speculative program execution. The system begins by executing the section of code using a head thread in order to produce a result. Before the head thread produces the result, the system generates a predicted result to be used in place of the result. Next, the system allows a speculative thread to use the predicted result in speculatively executing subsequent code that follows the section of code. After the head thread finishes executing the section of code, the system determines if a difference between the predicted result and the result generated by the head thread has affected execution of the speculative thread. If so, the system executes the subsequent code again using the result generated by the head thread. If not, the system performs a join operation to merge state associated with the speculative thread with state associated with the head thread.

Abstract: Object simulation and interaction of and between computer-generated or graphical objects in a virtual space includes neutral scene graphs, data structures and procedures for using such graphs and data structures.

Abstract: One embodiment of the present invention supports execution of a start transactional execution (STE) instruction, which marks the beginning of a block of instructions to be executed transactionally. Upon encountering the STE instruction during execution of a program, the system commences transactional execution of the block of instructions following the STE instruction. Changes made during this transactional execution are not committed to the architectural state of the processor until the transactional execution successfully completes.

Abstract: By encoding an exception triggering value in storage referenced by an instruction in an otherwise unused slot (e.g., the delay slot of a delayed control transfer instruction or an unused instruction position in a VLIW-based architecture) coinciding with a safe point, an efficient coordination mechanism can be provided for multi-threaded code. Because the mechanism(s) impose negligible overhead when not employed and can be engaged in response to an event (e.g., a start garbage collection event), safe points can be defined at call, return and/or backward branch points throughout mutator code to reduce the latency between the event and suspension of all threads. Though particularly advantageous for thread suspension to perform garbage collection at safe points, the techniques described herein are more generally applicable to program suspension at coordination points coinciding with calls, returns, branches or calls, returns and branches therein.

Abstract: A processor including a large register file utilizes a dirty bit storage coupled to the register file and a dirty bit logic that controls resetting of the dirty bit storage. The dirty bit logic determines whether a register or group of registers in the register file has been written since the process was loaded or the context was last restored and, if written generates a value in the dirty bit storage that designates the written condition of the register or group of registers. When the context is next saved, the dirty bit logic saves a particular register or group of registers when the dirty bit storage indicates that a register or group of registers was written. If the register or group of registers was not written, the context is switched without saving the register or group of registers. The dirty bit storage is initialized when a process is loaded or the context changes.

Abstract: One embodiment of the present invention provides a processor which selectively fetches cache lines for store instructions during speculative-execution. During normal execution, the processor issues instructions for execution in program order. Upon encountering an instruction which generates a launch condition, the processor performs a checkpoint and begins the execution of instructions in a speculative-execution mode. Upon encountering a store instruction during the speculative-execution mode, the processor checks an L1 data cache for a matching cache line and checks a store buffer for a store to a matching cache line. If a matching cache line is already present in the L1 data cache or if the store to a matching cache line is already present in the store buffer, the processor suppresses generation of the fetch for the cache line. Otherwise, the processor generates a fetch for the cache line.

Abstract: The present invention provides an efficient implementation of multiprecision arithmetic, such as for a microprocessor. For example, an implementation of multiprecision arithmetic is provided that eliminates condition codes, such as condition codes for a carry bit and a borrow bit, and eliminates an add-with-carry instruction for multiprecision addition and a subtract-with-borrow instruction for multiprecision subtraction. In one embodiment, a method includes separately performing a first one or more arithmetic operations and a second one or more arithmetic operations. The second arithmetic operations indicate if the first arithmetic operations cause a carry condition or if the first arithmetic operations cause a borrow condition. The one or more results of the first and second arithmetic operations are then provided. The first and second arithmetic operations can be executed in parallel on a microprocessor.

Abstract: One embodiment of the present invention provides a system which selectively executes deferred instructions following a return of a long-latency operation in a processor that supports speculative-execution. During normal-execution mode, the processor issues instructions for execution in program order. When the processor encounters a long-latency operation, such as a load miss, the processor records the long-latency operation in a long-latency scoreboard, wherein each entry in the long-latency scoreboard includes a deferred buffer start index. Upon encountering an unresolved data dependency during execution of an instruction, the processor performs a checkpointing operation and executes subsequent instructions in an execute-ahead mode, wherein instructions that cannot be executed because of the unresolved data dependency are deferred into a deferred buffer, and wherein other non-deferred instructions are executed in program order.

Abstract: One embodiment of the present invention provides a system which facilitates eliminating a restart penalty when reissuing deferred instructions in a processor that supports speculative-execution. During a normal execution mode, the system issues instructions for execution in program order. Upon encountering an unresolved data dependency during execution of an instruction, the processor performs a checkpointing operation and executes subsequent instructions in an execute-ahead mode, wherein instructions that cannot be executed because of the unresolved data dependency are deferred, and wherein other non-deferred instructions are executed in program order. When an unresolved data dependency is resolved during execute-ahead mode, the processor begins to execute the deferred instructions in a deferred mode. In doing so, the processor initially issues deferred instructions, which have already been decoded, from a deferred queue.