An execution architecture, a development architecture and an operations architecture for a netcentric computing system. The execution architecture contains common, run-time services required when an application executes in the netcentric computing system. The development architecture is the production environment for one or several systems development projects as well as for maintenance efforts. The purpose of the development environment is to support the tasks involved in the analysis, design, construction, and maintenance of business systems, as well as the associated management processes. It is important to note that the environment should adequately support all the development tasks, not just the code/compile/test/debug cycle. The operations architecture is a combination of tools and support services required to keep a production system up and running efficiently.

Architectures for netcentric computing systems

We describe the design and implementation of Dynamo, a software dynamic optimization system that is capable of transparently improving the performance of a native instruction stream as it executes on the processor. The input native instruction stream to Dynamo can be dynamically generated (by a JIT for example), or it can come from the execution of a statically compiled native binary. This paper evaluates the Dynamo system in the latter, more challenging situation, in order to emphasize the limits, rather than the potential, of the system. Our experiments demonstrate that even statically optimized native binaries can be accelerated Dynamo, and often by a significant degree. For example, the average performance of -O optimized SpecInt95 benchmark binaries created by the HP product C compiler is improved to a level comparable to their -O4 optimized version running without Dynamo. Dynamo achieves this by focusing its efforts on optimization opportunities that tend to manifest only at runtime, and hence opportunities that might be difficult for a static compiler to exploit. Dynamo's operation is transparent in the sense that it does not depend on any user annotations or binary instrumentation, and does not require multiple runs, or any special compiler, operating system or hardware support. The Dynamo prototype presented here is a realistic implementation running on an HP PA-8000 workstation under the HPUX 10.20 operating system.

/pdf/dynamo-a-transparent-dynamic-optimization-system-18ljty4tkn.pdf

Dynamo: a transparent dynamic optimization system

Until recently, the x86 architecture has not permitted classical trap-and-emulate virtualization. Virtual Machine Monitors for x86, such as VMware ® Workstation and Virtual PC, have instead used binary translation of the guest kernel code. However, both Intel and AMD have now introduced architectural extensions to support classical virtualization.We compare an existing software VMM with a new VMM designed for the emerging hardware support. Surprisingly, the hardware VMM often suffers lower performance than the pure software VMM. To determine why, we study architecture-level events such as page table updates, context switches and I/O, and find their costs vastly different among native, software VMM and hardware VMM execution.We find that the hardware support fails to provide an unambiguous performance advantage for two primary reasons: first, it offers no support for MMU virtualization; second, it fails to co-exist with existing software techniques for MMU virtualization. We look ahead to emerging techniques for addressing this MMU virtualization problem in the context of hardware-assisted virtualization.

/pdf/a-comparison-of-software-and-hardware-techniques-for-x86-abe9cjup21.pdf

A comparison of software and hardware techniques for x86 virtualization

Dynamic classification of sections of software using a profile-based optimization system optimizes management of the sections of software. Software executes under expected usage conditions. After execution, a set of usage profiles describes the dynamic properties of sections of the software. Each usage profile includes information identifying a section of software. Each usage profile maps to an outcome meant to optimize management of the sections of the software during later execution. During such later execution, a usage background describes the dynamic properties of a section of the software. The usage background includes information identifying the section of software. By matching the usage background to a usage profile in the set of usage profiles, the section is dynamically classified during later execution. Based on this dynamic classification, the section maps to the outcome meant to optimize management of the sections of software.

Dynamic classification of sections of software

Software systems have been using "just-in-time" compilation (JIT) techniques since the 1960s. Broadly, JIT compilation includes any translation performed dynamically, after a program has started execution. We examine the motivation behind JIT compilation and constraints imposed on JIT compilation systems, and present a classification scheme for such systems. This classification emerges as we survey forty years of JIT work, from 1960--2000.

/pdf/a-brief-history-of-just-in-time-d2108jerqh.pdf

A brief history of just-in-time

Building high-performance virtual machines is a complex and expensive undertaking; many popular languages still have low-performance implementations. We describe a new approach to virtual machine (VM) construction that amortizes much of the effort in initial construction by allowing new languages to be implemented with modest additional effort. The approach relies on abstract syntax tree (AST) interpretation where a node can rewrite itself to a more specialized or more general node, together with an optimizing compiler that exploits the structure of the interpreter. The compiler uses speculative assumptions and deoptimization in order to produce efficient machine code. Our initial experience suggests that high performance is attainable while preserving a modular and layered architecture, and that new high-performance language implementations can be obtained by writing little more than a stylized interpreter.

/pdf/one-vm-to-rule-them-all-2lvuhbbcre.pdf

One VM to rule them all

The Java programming language promises portable, secure execution of applications. Early Java implementations relied on interpretation, leading to poor performance compared to compiled programs. Compiling Java programs to the native machine instructions provides much higher performance. Because traditional compilation would defeat Java's portability and security, another approach is necessary. This article describes some of the important issues related to just-in-time, or JIT, compilation techniques for Java. We focus on the JIT compilers developed by Sun for use with the JDK (Java Development Kit) virtual machine running on SPARC and Intel processors. (Access the Web at www.sun. com/workshop/java/jit for these compilers and additional information.) We also discuss performance improvements and limitations of JIT compilers. Future Java implementations may provide even better performance, and we outline some specific techniques that they may use.

Compiling Java just in time

The analysis of the lifetime of objects in a garbage-collected system may be accomplished quickly and effectively using reference counts and cyclic garbage analysis. A reference count is maintained for each of the objects to indicate the number of incoming pointers. Each time the graph structure is altered, the reference counts are updated. Timestamps are recorded each time the reference count for objects change. If a reference count goes to zero, the corresponding object may be indicated as dead. A garbage collection need only be run once (perhaps at the end), and after it is run the system may indicate which objects are cyclic garbage. The timestamps for objects which are cyclic garbage are then reviewed in reverse chronological order. For each timestamp found, the corresponding object and any object reachable from the corresponding object are indicated as dead. These objects are then removed from the set of cyclic garbage.

Fast lifetime analysis of objects in a garbage-collected system

Apparatus, methods, systems and computer program products are disclosed describing a data structure and associated processes that optimize garbage collection techniques. The disclosed data structure can be used as an instantiated object in an object-oriented programming environment. The data structure uses a data structure header to separate the portion of the data structure that contains pointer values from the portion of the data structure that contains non-pointer values. The contents of the first word of the data structure header is distinguishable from any pointer value. Thus, a garbage collection procedure can more rapidly locate the pointer values in the data structure. Another advantage of this data structure organization, when applied to instantiated objects, is that the position of an instance variable (with respect to the object header structure) remains constant for all instantiated objects including those based on subclasses of the original class.

Method and apparatus for optimizing exact garbage collection using a bifurcated data structure

One embodiment of the present invention presents a method and apparatus for efficiently performing garbage collection on objects defined within an object-oriented programming system. Garbage collection typically involves following pointers to determine which objects are presently being referenced so that other objects, that are not being referenced, can be removed. To this end, the present invention maintains a bitmap for each object that indicates which variables in the object are pointer variables and which variables are non-pointer variables. A garbage collection process examines the bitmap, and on the basis of the pattern contained in the bitmap jumps to a particular routine that is tailored to garbage collect the particular pattern of pointer and non-pointer values in the object. Note that the system includes a routine tailored for each possible bitmap pattern. This technique speeds up the garbage collection process by eliminating the need to read type information one variable at a time to determine which variables within an object contain pointers.

Mario I. Wolczko

Papers

One VM to rule them all

Compiling Java just in time

Fast lifetime analysis of objects in a garbage-collected system

Method and apparatus for optimizing exact garbage collection using a bifurcated data structure

Performing exact garbage collection using bitmaps that identify pointer values within objects