### Table 2: The frequency of occurrence of various sub- graphs seen across our benchmark suite.

1999

"... In PAGE 9: ...he add-with-carryin described in Section 2.3. This pattern occurs frequently and having few inputs is highly ranked. It is the most important of the pat- terns found as evidenced by the frequency numbers shown in Table2 . Figure 12 is the conditional add operation also described earlier.... In PAGE 9: ... CPR found several patterns which expose compiler inadequacies and ar- eas for future optimization. Table2 shows the frequency of occurrence of each of these subgraphs in all the kernels. 4.... ..."

Cited by 15

### Table 1. IPs of the Sub-Graph open file of the Fig. 2

"... In PAGE 7: ...e. legal interaction pairs (IP) of inputs as the edges of the FSA ( Table1 , IPs based on the sub-graph in Fig. 2a).... In PAGE 8: ... Please note that the Fig. 2a, Table1 and Table 2 are equivalent; the different representations are supposed to help better understanding. 2.... ..."

### TABLE I Results of applying the slack-based op-chaining algorithm. The first eight columns report benchmark-specific information: total nodes, |V|, the number of concurrent threads, |Vm|, pre-assigned pipeline registers, |V0|, maximal single-output sub-graphs, |Gm| and the total nodes within each sub-graph in Gm. The next four columns report, for each version, the fraction of total nodes that are free of pipeline registers. The final column reports the runtime, in seconds.

2007

Cited by 1

### Table 5. Comparison with the error-correcting sub-graph isomorphism algorithm

2002

"... In PAGE 5: ... This can be very significant when matching large graphs. Table5 shows a comparison with the A*- based error-correcting algorithm over 1000 pairs of graphs generated randomly. The size of each graph is between 2 and 10 nodes.... ..."

Cited by 2

### Table 4: Pre-computed table for random sub-graph attack

"... In PAGE 13: ... For each iteration, the attacker tests for a collision with the set of computed end-points, and if a collision is found for a function fk i , the internal state of A5/1 is the preceding state r 1 i (fk 1 i (start point), which is straightforward to compute. Table4 is intended to be stored on a hard disk and [2] estimate the speed of a collision test to be 6ms. This would make the duration of an attack several days.... ..."

### Table 1: Runtime on three types of circuit struc- tures and sub-graphs of i10.

"... In PAGE 4: ... This ratio grows almost linearly, indicating that the complexity of the sparse LP linear programming algorithm is quadratic to the size of circuit. Table1 depicts the runtime of the two algorithm on sev- eral circuit structures: k-nary trees, the two DAGs in Fig- ure 2 and 4-input subnetworks from i10.blif.... In PAGE 5: ... DAG0 and DAG1 have 2 fanouts. From Table1 , in all structures, the DP algorithm is at least 16 times faster than LP, and with more unknown nodes to be placed, this speedup is even more reflecting a quadratic complexity of LP. Surprisingly, about half of the random samples of DAG0 and DAG1 and 99.... ..."

### Table 1 Change in topological parameters for various ways of increasing network size from n to n2

2005

"... In PAGE 8: ... The various parameters are formulated in terms of the respective parameters for the net- work of size n. As evident from Table1 , swapped networks... ..."

Cited by 3

### Table 3: Summary of odd parity results using linear sub-graph crossover and node de- scription crossover with relative probability 3 and 1, respectively.

1997

Cited by 13

### Table 1. Change in topological parameters for various ways of increasing network size from n to n2. For example, an entry of D for diameter means that there is no change when the network size is squared and 2D means that the diameter is doubled. Squared network refers to the cross product of a network with itself.

"... In PAGE 5: ... In particular, we showed that swapped networks are cost-effective, modular, packageable, and quite robust. Table1 indicates how the main topological parameters of certain classes of interconnection networks change as the network size is scaled from n to n2. The various parameters are formulated in terms of the respective parameters for the network of size n.... In PAGE 5: ... The various parameters are formulated in terms of the respective parameters for the network of size n. As evident from Table1 , swapped networks offer a mechanism for increasing the size of a network with relatively small cost increase, while limiting the deterioration of topological parameters and ensuring strong fault tolerance. Work in progress includes expanding our results on the properties of swapped network to include other topological attributes as well as a deeper analysis of their performance parameters and robustness attributes.... ..."

### Table 1. Performance Characteristics of Different AM Implementations

1997

"... In PAGE 7: ...ficient, buffered writes in the SCI DSM only. Performance measurements on the UCSB SCI cluster show competitive performance behavior of the SCI AM system ( Table1 ). Our own implementation, depicted in the first row of Table 1, adds little over- head to the raw latency of 9.... ..."

Cited by 13