CiteSeerX — Table search results

Results 31 - 40 of 22,422

Table 2: Program Versions: This table lists the versions of the programs we used.

in Measuring Link Bandwidths Using a Deterministic Model of Packet Delay

by Kevin Lai, Mary Baker 2000

"... In PAGE 9: ... The results in this sec- tion are preliminary and are intended as a proof of concept and not an evaluation of the full potential of the technique. We compare the results of the latest publicly available ver- sions (listed in Table2 ) of pathchar [7], clink [4], pchar [11], and nettimer. pathchar, clink, and pchar implement the one-packet technique described earlier.... ..."

Cited by 161

Table 4.1 contains time complexity and memory overhead for the different versions of the polynomial expansion algorithm described above. Included are also the case three-dimensional data with full certainty. This case will be used in an experiment in section 6.1.2.

in Multiscale Curvature Detection in Computer Vision

by Björn Johansson, Bjorn Johansson 2001

Cited by 15

Table 1. Computational complexity of full-band/subband NLMS and full-band LMS algorithms

in Architectural Synthesis of Computational Engines for Subband Adaptive Filtering

by S. Ramanathan, V. Visvanathan, S. K. Nandy

"... In PAGE 3: ... De ne the weight vector wi(m), con- taining the coe cients of the i-th band adaptive sub lter Wi(zL), and the corresponding input vec- tor Xi(m) as: wi(m) = [wi;0(m) wi;1(m) wi;2(m) ::: wi;(NL ?1)(m)] (1) Xi(m) = [Xi(m) Xi(m ? 1) Xi(m ? 2) ::: Xi(m ? (N L ? 1))] (2) where, N is the number of coe cients in the full- band version of the adaptive lter and Xi(m) is the i-th band analysis lter output. The coe - cient update equation based on the NLMS algo- rithm for the i-th band adaptive sub lter Wi(zL) is given by: wi(m + 1) = wi(m) + kXi(m)k2 Ei(m) Xi(m) (3) where, is the step-size, and Ei(m) = Di(m) ? Yi(m) (4) Yi(m) = wi(m) XT i (m) (5) kXi(m)k2 = Xi(m) XT i (m) (6) The computational complexity (per input sam- ple period) of the full-band/subband NLMS1 and the full-band LMS are listed in Table1 [13]-[14], where, Nfb is the number of coe cients in each of the analysis and synthesis sub lters. For large N, which is the case with many of the applica- tions, the computational savings in the subband NLMS vis- a-vis the full-band LMS is substantial, since Nfb is small and L is greater than 1.... ..."

Table 3: Cache performance of the hyperplane and pipeline versions of LU, measured with the perfex tool on the Origin2000. Cycles and cache misses are given in seconds. Numbers were obtained for 1/10th of the full iteration for Class A. 1 CPU 4 CPUs

in The OpenMP Implementation of NAS Parallel Benchmarks and its Performance

by H. Jin, H. Jin, M. Frumkin, M. Frumkin, J. Yan, J. Yan 1999

"... In PAGE 16: ...and the results are listed in Table3 measurements done on 1 CPU and 4 CPUs. With one CPU, the pipeline version has slightly less L1 and L2 cache misses, but the TLB miss is significantly less.... ..."

Cited by 82

Table 3 Twelve Nearest Neighbors to Eat, Car, Reading, and Slowly in Context, Order, and Composite Spaces

in Representing word meaning and order information in a composite holographic lexicon

by Michael N. Jones, Douglas J. K. Mewhort 2007

"... In PAGE 12: ... Contextual and order information complement each other. Table3 shows the top 12 neighbors (and their cosines) to eat, car, reading, and slowly in the full versions of each space (i.e.... ..."

Cited by 4

Table 1 shows the result of these experiments.

in NESSIE Document NES/DOC/SAG/WP3/027/1 Statistical Attacks on the Stream Cipher LEVIATHAN

by Marcus Schafheutle

"... In PAGE 1: ... Table1 . s-box matching attack on LEVIATHAN 3 Conclusion The s-box matching distinguisher works very ne even for the full version of LEVIATHAN.... ..."

Table 1: Riboswitch sub-families in Rfam database (version 7.0). Average length and %identity are based on the information in Rfam database. #seed is the number of sequences in the seed alignment. #total is the number of full family sequences.

in V.: A sequence-based filtering method for ncRNA identification and its application to searching for riboswitch elements

by Shaojie Zhang, Ilya Borovok, Yair Aharonowitz, Roded Sharan, Vineet Bafna 2006

"... In PAGE 7: ... In contrast, the riboswitches, with 12 distinct sub-families (and new sub-families being continuously discovered) are quite diverse, and relatively difficult to filter. Table1 summarizes known riboswitches from Rfam v.... ..."

Cited by 2

Table 3: Measure 1000 iterations of the Split-C primitive get. Three versions of Split-C are used: one without tracing, one with tracing (where data is flushed on barriers), and one with tracing (where data is flushed only when the buffer is full).

in Profiling a Parallel Language Based on Fine-Grained Communication

by Björn Haake, Klaus E. Schauser, Chris J. Scheiman 1996

"... In PAGE 10: ... The simplest strategy would be to just flush a buffer when it gets full. This does not perform well in practice, as we can see from Table3 . In this experiment, we repeatedly measure the time it takes to perform 1000 roundtrip get operations.... ..."

Cited by 1

Table 3: Measure 1000 iterations of the Split-C primitive get. Three versions of Split-C are used: one without tracing, one with tracing (where data is flushed on barriers), and one with tracing (where data is flushed only when the buffer is full).

in Profiling Techniques for a Fine-Grained Parallel Language

by Chris J. Scheiman, Björn Haake, Klaus E. Schauser

"... In PAGE 13: ... The simplest strategy would be to just flush a buffer when it gets full. This does not perform well in practice, as we can see from Table3 . In this experiment, we repeatedly measure the time it takes to perform 1000 roundtrip get operations.... ..."

Table 5.1: Output from Split-C program measuring the roundtrip times of Split-C primitives. Three versions are shown: one without tracing, one with tracing (where data is ushed on barriers), and one with tracing (where data is ushed only when the bu er is full).

in Design, Implementation, and Analysis of a Split-C Profiler

by Björn Haake, Advisor Prof, Klaus Erik Schauser, Ph. D

Results 31 - 40 of 22,422

Tools

Table 2: Program Versions: This table lists the versions of the programs we used.

Table 4.1 contains time complexity and memory overhead for the different versions of the polynomial expansion algorithm described above. Included are also the case three-dimensional data with full certainty. This case will be used in an experiment in section 6.1.2.

Table 1. Computational complexity of full-band/subband NLMS and full-band LMS algorithms

Table 3: Cache performance of the hyperplane and pipeline versions of LU, measured with the perfex tool on the Origin2000. Cycles and cache misses are given in seconds. Numbers were obtained for 1/10th of the full iteration for Class A. 1 CPU 4 CPUs

Table 3 Twelve Nearest Neighbors to Eat, Car, Reading, and Slowly in Context, Order, and Composite Spaces

Table 1 shows the result of these experiments.

Table 1: Riboswitch sub-families in Rfam database (version 7.0). Average length and %identity are based on the information in Rfam database. #seed is the number of sequences in the seed alignment. #total is the number of full family sequences.

Table 3: Measure 1000 iterations of the Split-C primitive get. Three versions of Split-C are used: one without tracing, one with tracing (where data is flushed on barriers), and one with tracing (where data is flushed only when the buffer is full).

Table 3: Measure 1000 iterations of the Split-C primitive get. Three versions of Split-C are used: one without tracing, one with tracing (where data is flushed on barriers), and one with tracing (where data is flushed only when the buffer is full).

Table 5.1: Output from Split-C program measuring the roundtrip times of Split-C primitives. Three versions are shown: one without tracing, one with tracing (where data is ushed on barriers), and one with tracing (where data is ushed only when the bu er is full).