### Table 7 presents comparative results, obtained on the PC, for the accuracy of the calculations using SLICOT algorithms, as well as the available MATLAB codes. Standard QR factorization has been used in all these calculations, but exactly the same accuracy results have been obtained using the Cholesky factorization algorithm, except for Application # 4, for which the Cholesky algorithm failed, and the QR algorithm was automatically called. The input and output trajec- tories for Application # 4 are plotted in the Figure 2 and Figure 3, respectively. The reported relative output error has been computed with the following MATLAB formula 20

"... In PAGE 22: ... A default value is used when ldwork is not speci ed, or if it is less than the minimal workspace needed. Table7 : Relative output errors using the QR or Cholesky factorization algorithms.... ..."

### Table 2. Performance in MFlops of the double precision Modi ed Cholesky Factorization.

1997

"... In PAGE 16: ... to 10. We show, in Table2 , the performance of the Level 1 and Level 2 BLAS version of the Modi ed Cholesky factorization (called dsymd2) and the performance of the block version (called dsymdf) on two RISC workstations: the HP 715/64 and the IBM RS/6000-750. We give the performance achieved using di erent block sizes (16, 32, and 64) and for matrices of order 100 and 500.... ..."

Cited by 2

### Table 5. Branching factor reductions

"... In PAGE 18: ...17. A search tree Table5 illustrates this by giving b0 (with 1% and 5% of good moves missed) and the corresponding upper bound on the average branching factor we obtained by solving the equation mentioned above using the ordering obtained with the heuristics presented in Section 4.... ..."

### Table 1: Clustering error and speed before and after performing incomplete Cholesky decomposition. err1, t1: clustering error and time using the full kernel matrix; err2, t2: clustering error and time using the incomplete Cholesky factor. col#: number of columns in the incomplete Cholesky factor. (m, d): sample size and dimension. c: number of clusters. Breastcancer Iris Wine Soybean Vehicle Glass Segment USPS Vowel

### Table 2: Damage Reduction Factor using

2002

"... In PAGE 8: ...amage reductions of 1.4 to 3.8 times can be achieved with Fractional/Equal IAS/DS, depending upon topology (see Table 2). Table2 shows the damage reduction factors that can be achieved by switching from a Weighted/Proportial IAS/DS to a Fractional/Equal IAS/DS for all of the topologies considered with the malicious node in the most threatening position. For example, employing Frac- tional/Equal IAS/DS for the power-law topology results in reducing damage by about a factor of two as compared to Weighted/Proportional IAS/DS when the malicious node is highly connected.... ..."

Cited by 28

### Table 2: Damage Reduction Factor using

"... In PAGE 8: ...amage reductions of 1.4 to 3.8 times can be achieved with Fractional/Equal IAS/DS, depending upon topology (see Table 2). Table2 shows the damage reduction factors that can be achieved by switching from a Weighted/Proportial IAS/DS to a Fractional/Equal IAS/DS for all of the topologies considered with the malicious node in the most threatening position. For example, employing Frac- tional/Equal IAS/DS for the power-law topology results in reducing damage by about a factor of two as compared to Weighted/Proportional IAS/DS when the malicious node is highly connected.... ..."

### Table 1: Overview of factors influencing the dimensions of GQM goals Factors

"... In PAGE 5: ...n the context is used to make environmental influential factors explicit, e.g., team structure and experience, application domain. These five dimensions are summarized in Table1 . They specify completely a measurement goal [BCR94].... In PAGE 5: ... They specify completely a measurement goal [BCR94]. An example of a measurement goal using the GQM goal template is: Analyze the final product for the purpose of characterization with respect to reliability from the viewpoint of the tester in the context of Project X Table1 : Dimensions of the measurement goal templates Dimension Definition Examples Object of Study What will be analyzed development process, system test, design document, final product,.... ..."

### Table 2: Performance of out-of-core QR factorization on 64 processors using MB=NB=50.

1997

"... In PAGE 12: ... Note that without this extra reordering cost and assuming perfect speedup from 64 to 256 processors, the out-of-core solver incurs approximately a 18% overhead over in-core solvers ((3502 ? 290)=(681 4) 1:18). Table2 shows the runtime (in seconds) for the out-of-core QR factorization on the Intel Paragon. The eld lwork is the amount of temporary storage (number of double precision numbers) available to the out-of-core routine for panels X and Y.... ..."

Cited by 17

### Table 2: factors influencing decisions on dimensions of frameworks Factor Reason

2005

"... In PAGE 8: ... The simplicity of the image can reduce the sensitivity of the framework to the complexities of the qualifications system. The exercise of choice in NQF design is often therefore limited in scope by factors like those in Table2 . Having looked at the architecture of qualifications frameworks from a broad policy perspective it may be useful to look at some of the technical design features that could form part of a NQF.... ..."

### Table 2: Performance of out-of-core QR factorization on 64 processors using MB=NB=50.

"... In PAGE 12: ... Note that without this extra reordering cost and assuming perfect speedup from 64 to 256 processors, the out-of-core solver incurs approximately a 18% overhead over in-core solvers ((3502 ? 290)=(681 4) 1:18). Table2 shows the runtime (in seconds) for the out-of-core QR factorization on the Intel Paragon. The eld lwork is the amount of temporary storage (number of double precision numbers) available to the out-of-core routine for panels X and Y.... ..."