### Table 6: Lins triality

2008

"... In PAGE 16: ... Other operations were introduced in [Li82]: mapping (a, b, c) to (a, b, ac) or (ac, b, c) produce the maps called phial(M) and skew(M). In [JoTh87] it is proved that there is no other good notions of dualities for maps on surfaces than the six ones given in Table6 .... In PAGE 16: ... By using operation phial, we see that, moreover, the graph of zigzags Z(M) is connected. The six operations depicted in Table6 form a group isomorphic to Sym(3). In particular, each of operations dual, skew and phial is a reflexion.... ..."

### Table 17: Results using the BNC with the Lin Measure

2005

"... In PAGE 53: ... There are small differences in the results when using Laplace (Add1) smoothing or Resnik counting, and the use of smoothing or Resnik counting does seem to increase both precision and recall, but the differences are not large. Table17 shows the results of using the same information content data with the Lin measure on the same five SemCor files. The differences in this case are smaller than the differences with the Jiang amp; Conrath measure.... ..."

### Table 1. The LinCa syntax

"... In PAGE 2: ... In order to investigate the introduction of quantitative information in the Linda paradigm, [5] introduces a minimal language, called Linda Calculus or LinCa, which includes the Linda core calculus expressed via only three con- structs: prefix, parallel composition and replication. The syntax is presented in Table1 . nil represents the inactive process.... ..."

### Table 1: 8 Processor Communication Schedule for Lin-

1992

"... In PAGE 8: ...iteration = iteration + 1 fori=1tonprocsdo P i selects the next available P j among the processors it has to send to If P j also sends to P i then do an exchange Mark P i and P j as unavailable for this iter Decrement msgs to send appropriately end for end while Figure 12: Greedy Scheduling Algorithm Table1 0: Communication Schedule for Pattern apos;P apos; using Greedy Scheduling Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 0 $ 1 0 $ 3 0 ! 5 0 $ 6 1 ! 6 1 $ 7 2 $ 3 1 $ 2 1 $ 4 1 $ 5 3 ! 5 6 ! 2 4 $ 5 4 $ 7 3 $ 6 3 $ 4 2 4 6 $ 7 5 $ 6 7 ! 0 densities of 10%, 25%, 50% and 75% of complete ex- change and studied the performance of the aboveal- gorithms on these patterns for message sizes of 256 and 512 bytes on a 32 processor system. The results are given in Table 11.... In PAGE 9: ...Table1 1: Performance of Scheduling Algorithms for Synthetic Irregular Patterns on 32 Processors Time (ms.) Algorithms 10% Pattern 25% Pattern 50% Pattern 75% Pattern 256 bytes 512 bytes 256 bytes 512 bytes 256 bytes 512 bytes 256 bytes 512 bytes Linear 4.... In PAGE 9: ...241 12.29 Table1 2: Performance of Scheduling Algorithms for Real Irregular Patterns on 32 Processors Time (ms.) Algorithms Conj.... ..."

Cited by 11

### Table 1 Circuit Const. model Lin. model

1997

"... In PAGE 6: ... Two din0berent test sequences n28of 200 vectors eachn29 were usedn3a the n0crst one with the same input statistics of that used for characterizationn2c the second with a lower average input activity n28of 20n25 instead of 50n25n29. Experimental results are reported in Table1 . The n0crst 2 columns contain the name of the circuit and the number of In2fO signals.... In PAGE 7: ...9 6.1 Table1 n3a Comparison between constant and linear power estimators. For each benchmarkn2c estimates reported on the n0crst row have been obtained with the same input statistics used for characterizationn2c estimates on the second rowhave been obtained with completely din0berent input statistics.... ..."

Cited by 4

### Table 1: Pos + # SH + Free + Lin vs Pos + SH + Free + Lin: precision.

1999

"... In PAGE 4: ...that arbitrary precision losses are possible with # SH , Table1 illustrates how heavy precision penalties can be incurred by # SH even on real programs. Most notably, for bryant we see a precision loss as high as 28% on GI analysis and 42% on GD analysis.... ..."

Cited by 4

### Table 1. Pos + # SH + Free + Lin vs Pos + SH + Free + Lin: precision.

1999

"... In PAGE 4: ... In other words, Fecht has the whole merit for having trusted on this domain from a precision/e ciency perspective. The results for the remaining 19 programs are summarized in Table1... In PAGE 6: ...that arbitrary precision losses are possible with # SH , Table1 illustrates how heavy precision penalties can be incurred by # SH even on real programs. Most notably, for bryant we see a precision loss as high as 28% on goal-independent analysis (GI) and 42% on goal-dependent analysis.... ..."

Cited by 4