C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. SoftwarePractice and Experience, 31(8):717--738, 2001.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....with dynamically generated bytecode sequences. Dynamic compilation has proven essential for efficient implementation of high level languages [13, 1] Some just in time compilers perform profiling to identify which methods to spend more optimization time on [20] The Jalape no Java virtual machine [3, 23] utilizes idle processors in an SMP system to optimize code at runtime. Jalape no optimizes all code at an initial low level of optimization, embedding profiling information that is used to trigger reoptimization of frequently executed code at higher levels. Self [19] uses a similar adaptive ....

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the overhead of dynamic compilation. Software: Practice and Experience, 31(8), Mar. 2001.

....extensible systems that can be re targeted. Dynamic compilation has proven essential for efficient implementation of high level languages [14, 1] Some justin time compilers perform profiling to identify which methods to spend more optimization time on [22] The Jalape no Java virtual machine [3, 26] utilizes idle processors in an SMP system to optimize code at runtime. Jalape no optimizes all code at an initial low level of optimization, embedding profiling information that is used to trigger reoptimization of frequently executed code at higher levels. Self [21] uses a similar adaptive ....

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the overhead of dynamic compilation. Software: Practice and Experience, 31(8), Mar. 2001.

....with that of the extremely efficient SCCP constant propagator. If the two speeds are comparable, then together with the well known results from dynamic compilation, which has shown to have low overheads, we would show an encouraging result. The low overhead dynamic optimization, such as [2, 21, 22], reduces overhead by means of multiple threads of control that share the processor with the optimized application. Arnold et al. [2] for example, perform feedback directed inlining with re compilation overheads of 6 . Program Scope Specialize Scopes Versions Loops Building dotproduct 0.5 40 ....

Chandra J. Krintz, David Grove, Vivek Sarkar, and Brad Calder. Reducing the overhead of dynamic compilation. Software-Practice and Experience, 31(8):717--738, 2001.

....compiles methods using the baseline compiler, but methods whose execution is a significant part of the program execution time are recompiled using the Opt compiler. As can be seen in Figure 1, the quasi static compiler is one of these compilers. We use the lazy method compilation framework [27] of Jalape no, in which the JVM compiles a method only when it is actually invoked, rather than compiling each method of a class when it is linked and resolved. When the Jalape no system determines that a method needs to be compiled, it invokes the compiler driver. The compiler chosen by the JVM ....

....are allocated in the global table pointed to by the JTOC register. Thus these references need the proper offset in the current JVM instance. 3 The GC map for a method invocation site contains the invoked method s id in addition to GC information. This method id is used for lazy method invocation [27] and must be stitched. ffl Runtime field and method accesses: The Jalape no boot image is built with several predefined Java classes that mainly consist of runtime library methods, since Jalape no itself is written in Java [3] Thus the boot image is a frozen core, i.e. its classes are always ....

[Article contains additional citation context not shown here]

C. Krintz, D. Grove, D. Lieber, V. Sarkar, and B. Calder. Reducing the overhead of dynamic compilation. Technical Report CSE2000-0648, University of California, San Diego, April 2000.

No context found.

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. SoftwarePractice and Experience, 31(8):717--738, 2001.

No context found.

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software-Practice and Experience, 31(8):717--738, 2001.

No context found.

KRINTZ, C., GROVE, D., SARKAR, V., AND CALDER, B. Reducing the overhead of dynamic compilation. Software: Practice and Experience 32, 8 (2000), 717--738.

No context found.

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software-Practice and Experience, 31(8):717-738, 2001.

No context found.

Krintz, C., Grove, D., Sarkar, V., Calder, B.: Reducing the Overhead of Dynamic Compilation. Software---Practice and Experience 31 (2001) 717--738

No context found.

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software---Practice and Experience, 31(8):717--738, 2001.

No context found.

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software-Practice and Experience, 31(8):717--738, 2001.

....two dynamic compilers, a fast, non optimizing compiler, and a second optimizing compiler [5, 7] The program is compiled first with the fast compiler, and then frequently executed methods are compiled later with the optimizing compiler based on dynamic information gathered during execution. In [17], we examine the use of off line profile information to decide which compiler to use initially, however, we provide no automatic mechanism for the communication of this information to compilation system. In this paper, we present such a mechanism that introduces annotation into the bytecode stream ....

....our work, we perform off line instrumentation and analysis to determine important methods to optimize and pass this information to the compiler using our selective optimization annotation. This obviates the need to spend time for both the instrumentation, on line measurement, and recompilation. In [17], we proposed using a technique similar to selective optimization annotation to optimize important methods on a background processor. If a method is encountered before it has been optimized then it is compiled with a fast compiler and possibly replaced with an optimized version if deemed ....

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software---Practice and Experience, 31(8):717--738, 2001.

....and prove to reduce both compilation delay and execution performance. As part of future work though, we will provide and empirical evaluation of the performance advantages to using off line static and profile based analysis communicated via annotations or adaptive compilation or both. In [16], the authors use a technique similar to our selective optimization annotation to optimize important methods on a background processor. If a method is encountered before it has been optimized then it is compiled with a fast compiler and possibly replaced with an optimized version if deemed ....

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. In Software---Practice and Experience, 2000. To appear.

....uses two dynamic compilers, a fast, non optimizing compiler, and a second optimizing compiler [5, 7] The program is compiled rst with the fast compiler, and then frequently executed methods are compiled later with the optimizing compiler based on dynamic information gathered during execution. In [17], we examine the use of o line pro le information to decide which compiler to use initially, however, we provide no automatic mechanism for the communication of this information to compilation system. In this paper, we present such a mechanism that introduces annotation into the bytecode stream ....

....our work, we perform o line instrumentation and analysis to determine important methods to optimize and pass this information to the compiler using our selective optimization annotation. This obviates the need to spend time for both the instrumentation, on line measurement, and recompilation. In [17], we proposed using a technique similar to selective optimization annotation to optimize important methods on a background processor. If a method is encountered before it has been optimized then it is compiled with a fast compiler and possibly replaced with an optimized version if deemed ....

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software|Practice and Experience, 31(8):717-738, 2001.

....text of Chapter X is in part a reprint of the material as it appears in the Journal of Software: Practice and Experience, Software: Practice and Experience, Volume 31, Issue 8, pp. 717 738. The dissertation author was the primary researcher and author and the co authors listed on this publication ([50]) directed and supervised the research which forms the basis for Chapter X. The text of Chapter XI is in part a reprint of the material as it is to appear in the proceedings of the 2001 ACM SIGPLAN Conference on Programming Language Design and Implementation (PLDI) The dissertation author was ....

C. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the Overhead of Dynamic Compilation. Software---Practice and Experience, 31(8):717--738, 2001.

No context found.

C. J. Krintz, D. Grove, V. Sarkar, and B. Calder. Reducing the overhead of dynamic compilation. Software-Practice and Experience, 31(8):717--738, 2001.

No context found.

C. Krintz, D. Grove, D. Lieber, V. Sarkar, and B. Calder. Reducing the overhead of dynamic compilation. Software { Practice and Experience, 31(8):717-738, Dec. 2000.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC