Results 1 - 10 of 18,664

Table 1. Benchmark programs and algorithm running time. Release Date is the latest date stamp in the source file tree. Code blocks in our implementation are essentially program basic blocks with subroutine calls not considered as basic block separators. Build is the time for dissecting the program binary using the ATOM tool. Match is the time for constructing and matching control trees between two versions. All running times are measured using the time system utility in Digital Unix and are average results from five test runs. All standard deviations are under 0.1 second.

in Progressive Profiling: A Methodology based on Profile Propagation and Selective Profile Collection
by Zheng Wang, Michael D. Smith 1969
"... In PAGE 5: ....95 of the gcc suite. We also use versions 5.0 and 5.10 of the GNU ghostscript program as another benchmark case. Table1 summarizes the benchmark programs and lists the running time of our binary matching algorithm for the three test cases. We did not tune the algorithm specifically for the benchmark programs.... ..."
Cited by 1

Table 5. Code Block Matching Rates at Different Stages

in BMAT - A Binary Matching Tool for Stale Profile Propagation
by Zheng Wang, Ken Pierce, Scott McFarling 2000
"... In PAGE 16: ...Table 5. Code Block Matching Rates at Different Stages Table5 lists the total number and percentage of successfully matched code blocks in the newer build for three cases: after using only the first hashing pass (fuzziness level 1), after using all hashing passes, and after using both hashing and control-flow-based matching. These results demonstrate benefits from using multiple matching passes and performing control-flow-based basic block matching.... ..."
Cited by 20

Table 5. Code Block Matching Rates at Different Stages

in BMAT -- A Binary Matching Tool for Stale Profile Propagation
by Zheng Wang, Ken Pierce, Scott McFarling 2000
"... In PAGE 15: ... For code blocks, the Matching Rate may exceed the Maximum Remain because multiple blocks may be matched to the same block during control-flow-based basic block matching. For two cases of binary matching, Table 4 and Table5 list the matching rates on the procedure and code block levels, respectively, after different phases of the matching process. These numbers demonstrate the performance gain of BMAT from using a multiple-pass approach and a combination of different methods.... In PAGE 16: ...Table 5. Code Block Matching Rates at Different Stages Table5 lists the total number and percentage of successfully matched code blocks in the newer build for three cases: after using only the first hashing pass (fuzziness level 1), after using all hashing passes, and after using both hashing and control-flow-based matching. These results demonstrate benefits from using multiple matching passes and performing control-flow-based basic block matching.... ..."
Cited by 20

Table 5. Code Block Matching Rates at Different Stages

in BMAT -- A Binary Matching Tool for Stale Profile Propagation
by Zheng Wang, Ken Pierce, Scott McFarling 2000
"... In PAGE 15: ... For code blocks, the Matching Rate may exceed the Maximum Remain because multiple blocks may be matched to the same block during control-flow-based basic block matching. For two cases of binary matching, Table 4 and Table5 list the matching rates on the procedure and code block levels, respectively, after different phases of the matching process. These numbers demonstrate the performance gain of BMAT from using a multiple-pass approach and a combination of different methods.... In PAGE 16: ...Table 5. Code Block Matching Rates at Different Stages Table5 lists the total number and percentage of successfully matched code blocks in the newer build for three cases: after using only the first hashing pass (fuzziness level 1), after using all hashing passes, and after using both hashing and control-flow-based matching. These results demonstrate benefits from using multiple matching passes and performing control-flow-based basic block matching.... ..."
Cited by 20

Table 1. Building Blocks

in Overview and Outlook on the Semantic Desktop
by Leo Sauermann, Ansgar Bernardi, Andreas Dengel 2005
Cited by 16

Table 1. Building Blocks

in Overview and Outlook on the Semantic Desktop
by Leo Sauermann, Ansgar Bernardi, Andreas Dengel 2005
Cited by 16

Table 1: Building Blocks

in Structuring the Process of Airline Scheduling
by Jürgen Antes 1997
Cited by 2

Table 4.5. Code block matching rates at different stages. This table lists the total number and percentage of successfully matched code blocks in the newer build for three cases: after using only the first hashing pass (fuzziness level 1), after using all hashing passes, and after using both hashing and control flow-based matching. These results demonstrate benefits from using multiple matching passes and performing control flow-based basic block matching.

in Progressive Profiling: A Methodology Based on Profile Propagation and Selective Profile Collection
by Zheng Wang

Table 1: Basic Building Blocks

in FOUNDATIONS FOR THE HARMONIZATION OF INFORMATION TECHNOLOGY SECURITY STANDARDS Item: Deliverable #1 Source: Cooperation on Security of Information Systems Joint Task 01 Status: Revised Draft
by unknown authors
"... In PAGE 5: ... Each of the criteria start with basic building blocks. These building blocks are defined in Table1 for the various efforts. Table 1: Basic Building Blocks... ..."

Table 9. A Tabular Form of the Configuration Structure for a Product Family in the Structural View of PFA

in Design for Mass Customization by Developing Product Family Architecture
by Mitchell M. Tseng 1998
"... In PAGE 15: ... In practical production systems, a part coding system is usually used to represent different components and sub- assemblies for modules and/or end products. As illustrated in Table9 , different indented levels conform to the assembly levels from component to sub-assemblies and to end products. Various building blocks (those with their part codes bolded in Table 9) can be either a component or a sub-assembly and are shared at different levels across the entire product family.... In PAGE 15: ... As illustrated in Table 9, different indented levels conform to the assembly levels from component to sub-assemblies and to end products. Various building blocks (those with their part codes bolded in Table9 ) can be either a component or a sub-assembly and are shared at different levels across the entire product family. Finally, configuration design procedures are formulated according to the PFA in which various building blocks are well established in three consecutive views together with modular structures in each view.... ..."
Cited by 3
Results 1 - 10 of 18,664