K. Driesen. Software and Hardware Techniques for Efficient Polymorphic Calls. PhD thesis, University of California at Santa Barbara, June 1999. Supervised by Urs Holzle. TRCS99-24. Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Microarchitectural and Compile-Time Optimizations for.. - Kalamatianos (2000) (1 citation) (Correct)
....across a larger number of procedures is a possible reason behind the smaller basic block size, since loops and if then else statements are frequently partitioned across procedure boundaries. Polymorphism is one feature that tends to substitute if then else constructs with a procedure call [31]. A direct consequence of this is the lower static and dynamic frequency of conditional branches in OOP [12] the dynamic frequency of conditional branches was 61 for C and 80 for C in [13] Designing object oriented software, with code reuse and modularity in mind, generated a smaller ....
....C and Java. In both domains, however, polymorphic calls support the generation of high level reusable software components that fit different applications, and are therefore frequently encountered. Dispatch code overhead depends on the underlying run time mechanisms supporting polymorphic calls [31]. For C (which favors the Virtual Function Table (VFT) as a supporting run time mechanism for polymorphic calls) the dispatch code sequence consists of a few non branch instructions and an indirect branch [42] The non branch instructions are usually loads, though add operations may also be ....
[Article contains additional citation context not shown here]
K. Driesen. Software and Hardware Techniques for Efficient Polymorphic Calls. PhD thesis, University of California, Santa Barbara, 1999.
Multi-stage Cascaded Prediction - Driesen, Hölzle (1999) (5 citations) Self-citation (Driesen) (Correct)
....path lengths, fewer bits are extracted from each target in order to fit into the 24 bit key pattern 2 . A cascaded predictor consists of several stages, each containing a two level predictor with its own history buffer and predictor table. Successive stages use increasing path lengths (in [Dri99], we demonstrate the inferior accuracy of decreasing path lengths) The use of separate tables allows all stages to predict in parallel. In a final step, the predictor chooses the prediction from the last stage that did not encounter a table miss. This ensures that its target prediction is based ....
Karel Driesen. Software and Hardware Techniques for Efficient Polymorphic Calls. PhD dissertation, University of California, Santa Barbara (in preparation).
On the Syllogistic Structure of Object-Oriented Programming - Derek Rayside And (2001) (Correct)
No context found.
K. Driesen. Software and Hardware Techniques for Efficient Polymorphic Calls. PhD thesis, University of California at Santa Barbara, June 1999. Supervised by Urs Holzle. TRCS99-24.
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