"... In PAGE 3: ... (In [iii] other traffic profiles, such as bursty traffic were also considered). Table2 shows the sum of the objective functions for each case for the CCO and EP methods. For the 3-node problem there was exact agreement between CCO and GENOCOP algorithms.... ..."

"... In PAGE 8: ...2 Isolating Memory and Network Contention We consider three levels of nite bandwidth for the network and memories (cache bandwidth is as- sumed higher than both memory and network bandwidth). The bandwidth and latency parameters used in our experiments are described in Table1 and 2. We simulate machines that represent a range of \balanced quot; architectures, where the memory bandwidth is equal to the uni-directional bandwidth of a network link.... In PAGE 21: ...8 5.7 Table1 0: Broadcasting vs Eager Combining for Gaussian Elimination ancestor to copy the data, before it can copy the data. Under eager combining, a read operation by any client causes the data to be multicast to the servers, and also forwarded to the client; no processor is forced to wait for an ancestor in the tree.... ..."

"... In PAGE 21: ... Such signaling protocol is not required on the ring where the circular topology with simple implicit token passing is su cient; this is summarized in Table 1. Moreover, as summarized in Table2 , the bandwidth allocation on... ..."

"... In PAGE 21: ... Table1 shows the worst case achievable utilizations of the synchronous band- width allocation schemes introduced above. These results were derived using the assumption that D i = T i , i.... In PAGE 21: ...y these allocation schemes. The remaining 0.67 of the usable network bandwidth can be used for the transmission of asynchronous messages. From Table1 , it can be seen that the normalized proportional scheme and the local scheme both have the same worst case achievable utilization. An advantage of the local allocation scheme, however, is that it only uses information local to node i in calculating the synchronous bandwidth H i .... In PAGE 22: ... Robustness.Asshown in Table1 , the worst case achievable utilization of many synchronous bandwidth allocation schemes is known. The characteristics of syn- chronous message schemes can be altered and all messages will still meet their deadlines, provided that the utilization of the synchronous message streams remains below the worst case achievable utilization.... ..."

"... In PAGE 11: ... For each run of the algorithm we collected the actual and predicted year of rst value in the sequence, the number of hits which con rmed the match, the probability of the rst guess and the collective probability of the match. Table1 presents the results compressed in a single line for each group of ten test sequences of the same length. For each group we present: the length of sequences (SeqL); the size of window used for the construction of theory (WinL); the number of correct matches (NumM); the average, minimum and maximum number of hits (AvgH, MinH, and MaxH) expressed in percentages; the average probability of a guess (AvgGP) that leaded to matching; and the average collective probability of a correct match (AvgMP).... In PAGE 12: ...Table 1: Evaluation of matching algorithm for tree-ring dating thresholds Th and Tm were set relatively low in order to allow the matchings based on the partial agreement between the subsequence and the theory; the values of AvgH in Table1 denote the average proportion of the agreement by matchings. Finally, the prediction supported with the largest number of hits is chosen among the results of Algorithm 1.... ..."

"... In PAGE 13: ... For each run of the algorithm we collected the actual and predicted year of rst value in the sequence, the number of hits which con rmed the match, the probability of the rst guess and the collective probability of the match. Table1 presents the results compressed in a single line for each group of ten test sequences of the same length. For each group we present: the length of sequences (SeqL); the size of window used for the construction of theory (WinL); the number of correct matches (NumM); the average, minimum and maximum number of hits (AvgH, MinH, and MaxH) expressed in percentages; the average probability of a guess (AvgGP) that leaded to matching; and the average collective probability of a correct match (AvgMP).... In PAGE 14: ... The required percent of classi ers which agreed on the prediction Th were 50%, the required probability of the initial guess Tg was 80%, and the required collective probability of the match Tm was 40%. Note that the thresholds Th and Tm were set relatively low in order to allow the matchings based on the partial agreement between the subsequence and the theory; the values of AvgH in Table1 denote the average proportion of the agreement by matchings. Finally, the prediction supported with the largest number of hits is chosen among the results of Algorithm 1.... ..."

"... In PAGE 32: ... The BSE has to reject a lot of connection requests. The test results are shown in Table8 . Uavg denotes the ratio of the average utilized bandwidth to the maximum bandwidth of an FDDI ring calculated for the rst 500 token tours.... ..."