University Of Washington

Compaq Computer Corporation, Inc.

Susan Jane Eggers

Henry M Levy

Mark Henry Oskin

Steven James Swanson

Joel S Emer

Robert H Halstead, Jr

George Z Chrysos

Dean M Tullsen

Jack Lo

Trggve Fossum

Rebecca Stamm

Dean Tullsen

Luis Henrique Ceze

Andrew Richard Putnam

Craig D Chambers

Matthai Philipose

Emily Anne Fortuna

Andrew Michalski Schwerin

Owen C Anderson

Kenneth Alan Michelson

Andrew Kinoshita Petersen

Markus Mock

Martha Allen Mercaldi

Brian K Grant

Devices and methods of providing hardware support for dynamic type checking are provided. In some embodiments, a processor includes a type check register and support for one or more checked load instructions. In some embodiments, normal load instructions are replaced by a compiler with the checked load instructions. In some embodiments, to perform a checked load, an error handler instruction location is stored in the type check register, and a type tag operand is compared to a type tag stored in the loaded memory location. If the comparison succeeds, execution may proceed normally. If the comparison fails, execution may be transferred to the error handler instruction. In some embodiments, type prediction is performed to determine whether a checked load instruction is likely to fail.

Systems and methods for hardware-assisted type checking

A dataflow instruction set architecture and execution model, referred to as WaveScalar, which is designed for scalable, low-complexity/high-performance processors, while efficiently providing traditional memory semantics through a mechanism called wave-ordered memory. Wave-ordered memory enables 'real-world' programs, written in any language, to be run on the WaveScalar architecture, as well as any out-of-order execution unit. Because it is software-controlled, wave-ordered memory can be disabled to obtain greater parallelism. Wavescalar also includes a software-controlled tag management system.

Wavescalar architecture having a wave order memory

A microarchitecture and instruction set that supports multiple, simultaneously executing threads. The approach is disclosed in regard to its applicability in connection with a recently developed microarchitecture called 'WaveScalar.' WaveScalar is a compiler that breaks a control flow graph for a program into pieces called waves having instructions that are partially ordered (i.e., a wave contains no back-edges), and for which control enters at a single point. Certain aspects of the present approach are also generally applicable to executing multiple threads on a more conventional microarchitecture. In one aspect of this approach, instructions are provided that enable and disable wave-ordered memory. Additional memory access instructions bypass wave-ordered memory, exposing additional parallelism. Also, a lightweight, interthread synchronization is employed that models hardware queue locks. Finally, a simple fence instruction is used to allow applications to handle relaxed memory consistency.

Building a wavecache

A multi-threaded processor comprising a pipeline including a number of stages processing instructions belonging to a plurality of threads. A buffer stores instructions from different ones of the threads. Count logic stores count information relating to each of the threads to indicate the number of instructions in each of the corresponding threads that have a particular attribute. A selection logic circuit has an output coupled to the buffer to determine which instruction is to be read from the buffer based on the count information stored by the count logic. The count information may, for example, provide information relating to a likelihood that one or more instructions belonging to each of the threads will be cancelled; relating to a count of unresolved branch instructions; or relating to a count of outstanding data cache misses. In operation, a thread may be selected for execution based on a selected attribute to enhance processing performance.

Pipelined multi-thread processor selecting thread instruction in inter-stage buffer based on count information

Selective dynamic compilation of source code is performed using run-time information. A system is disclosed that implements a declarative, annotation based dynamic compilation of the source code, employing a partial evaluation, binding-time analysis (BTA), and including program-point-specific polyvariant division and specialization and dynamic versions of traditional global and peephole optimizations. The system allows programmers to declaratively specify policies that govern the aggressiveness of specialization and caching, providing fine control over the dynamic compilation process. The policies include directions for controlling specialization at promotion points and merge points, and further define caching policies, and speculative-specialization policies. The system also enables programmers to specialize programs across arbitrary edges, both at traditional locations, such as procedure boundaries, but also within procedures. Programmers are enabled to conditionally specialize programs based on evaluation of arbitrary compile-time and run-time conditions.

System and method for performing selective dynamic compilation using run-time information

A system and a method is described for freeing renaming registers that have been allocated to architectural registers prior to another instruction redefining the architectural register. Renaming registers are used by a processor to dynamically execute instructions out-of-order. The present invention may be employed by any single or multi-threaded processor that executes instructions out-of-order. A mechanism is described for freeing renaming registers that consists of a set of instructions, used by a compiler, to indicate to the processor when it can free the physical (renaming) register that is allocated to a particular architectural register. This mechanism permits the renaming register to be reassigned or reallocated to store another value as soon as the renaming register is no longer needed for allocation to the architectural register. There are at least three ways to enable the processor with an instruction that identifies the renaming register to be freed from allocation: (1) a user may explicitly provide the instruction to the processor that refers to a particular renaming register; (2) an operating system may provide the instruction when a thread is idle that refers to a set of registers associated with the thread; and (3) a compiler may include the instruction with the plurality of instructions presented to the processor. There are at least five embodiments of the instruction provided to the processor for freeing renaming registers allocated to architectural registers: (1) Free Register Bit; (2) Free Register; (3) Free Mask; (4) Free Opcode; and (5) Free Opcode/Mask. The Free Register Bit instruction provides the largest speedup for an out-of-order processor and the Free Register instruction provides the smallest speedup.

Mechanism for freeing registers on processors that perform dynamic out-of-order execution of instructions using renaming registers

A method and organization for implementing the registers required in a computer system supporting multithreading and dynamic out-of-order execution. Multithreaded computer systems are those in which the processor supports multiple contexts (threads), and either rapid context switching from thread to thread or scheduling of instructions from different threads within a single cycle. An important component of processors for such systems is the register file; the processor needs a large register file or resource to provide the registers used for the threads. One form of the invention maintains a set of private architecturally specified registers, and a set of private renaming register for each different thread. In the other three embodiments, sharing of renaming registers between different threads is permitted, to enable a reduction in the total number of registers required. One of these three embodiments enables any of the architecturally specified registers that are private to a thread but are not in use, to be employed as renaming registers. Another of the embodiments treats all registers as sharable and enables any register from the register file or resource to be used as a renaming register for any thread.

Shared register storage mechanisms for multithreaded computer systems

A technique is provided for selecting a preferred thread from a plurality of threads executing within a simultaneous multithreaded, out-of-order execution computer system, the preferred thread possessing those instructions which, while in flight within the pipeline of the computer system provide, in contrast to those instructions belonging to other threads, a more beneficial performance of the central processing unit of the computer system. To determine the preferred thread, a technique is provided to evaluate attributes of each thread which indicate whether the thread includes a number of instructions which are likely to be cancelled while in flight or whether a thread includes instructions which will remain in the instruction queue for a number of cycles, unable to execute, thus stalling the execution of the thread to which the instruction belongs.

Thread properties attribute vector based thread selection in

Growing community of inventors

Seattle, WA, United States of America

Susan Jane Eggers