J. S. Cox, D. P. Howell and T. M. Conte. Commercializing profile-driven optimization. In Trevor N. Mudge and Bruce D. Shriver, editors, Proceedings of the 28th Annual Hawaii International Conference on System Sciences. Volume 1: Architecture, pages 221--228, Los Alamitos, CA, USA, January 1995. IEEE Computer Society Press. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Dynamic Optimization through the use of Automatic Runtime.. - Whaley (1999) (3 citations) (Correct)
....can come up with a nearperfect hash function that is suited to the actual data that the program will be using. Static compiler writers have realized the benefit of using actual performance data, and profile directed optimization has already been incorporated into some modern commercial compilers [88, 131, 89]. However, profile directed static compilation has its faults. Profile directed static compilers require a number of test runs to collect profile information to prime the compilation. If the profile of the program during the test runs is different than from an actual execution, then the ....
.... utilize profile information from prior test runs to perform better optimizations, for example, trace scheduling [79] improving cache locality [87, 117, 48, 128, 127] or traditional optimizations [35, 34, 36, 23] Profile driven optimizations have shown up in recent commercial products [89, 88, 131]. There has been work on using profile information in dynamic compilers for Scheme [18, 17, 19] Self [61] Cecil [26] and ML [96, 101] More recently, the upcoming HotSpot dynamic compiler for Java claims to perform profile driven optimizations [84] 2.4 Dynamic code specialization The SELF ....
J. S. Cox, D. P. Howell and T. M. Conte. Commercializing profile-driven optimization. In Trevor N. Mudge and Bruce D. Shriver, editors, Proceedings of the 28th Annual Hawaii International Conference on System Sciences. Volume 1: Architecture, pages 221--228, Los Alamitos, CA, USA, January 1995. IEEE Computer Society Press.
Accurate and Practical Profile-Driven Compilation Using the.. - Thomas Conte (1996) (18 citations) Self-citation (Conte) (Correct)
....the slowdown to a maximum of 2.05 times for the SPEC92 benchmarks. Unfortunately, this overhead is still too large for integrated software vendors to readily absorb. Commercial software vendors in general can tolerate a negligible amount of additional execution overhead (i.e. approximately 5 ) [10]. Code instrumentation techniques, therefore, may not be appropriate for real time, interactive, and transaction processing applications. If profiling is to gain commercial acceptance, profiling must be smoothly integrated into the software development cycle. Unfortunately, this requires the ....
....and the optimizations are based on actual program usage. Profiling in this manner is commercially appealing because vendors alpha and beta testing processes are often very well defined, and the hardware based style of profiling leverages their existing investment to produce better optimized code [10]. Some profiling methods today utilize the hardware support for debugging to provide limited profiling capabilities. For example, the PA RISC architecture provides for a performance monitoring coprocessor that may be employed for collecting profile data [14] These techniques require frequent ....
J. S. Cox, D. P. Howell, and T. M. Conte. Commercializing profile-driven optimization. In Proc. 28th Hawaii Int'l. Conf. on System Sciences, volume 1, Maui, HI, Jan. 1995.
Hardware-Based Profiling: An Effective Technique for.. - Conte, Patel, Menezes, .. (1996) (6 citations) Self-citation (Cox Conte) (Correct)
....the object code by the compiler, and instructions to increment table entries are placed inside each transition block. After execution, node weights can be recovered from arc weights by summing all arc counts into or out of a basic block. Empirically, arc based profiling results in a 50 slowdown [20]. This paper presents a third, hardware based profiling approach that uses the contents of hardware branch prediction buffers to construct weights for the WCFG. The branch prediction hardware is sampled periodically using kernel mode instructions which are implemented in several commercial ....
J. S. Cox, D. P. Howell, and T. M. Conte, "Commercializing profile-driven optimization, " in Proc. 28th Hawaii Int'l. Conf. on System Sciences, vol. 1, (Maui, HI), pp. 221-- 228, Jan. 1995.
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