### Table 2: Objective function costs and run time results of the adaptive graph partitioners for various RCF values on problems derived from a particles-in-cells simulation on a Cray T3E. The numbers at the top of each column indicate the RCF of the experiments in that column.

2000

"... In PAGE 8: ...5 2 10 100 1000 auto mdual2 mrng3 WF LMSR 128-way Results Normalized by URA Figure 5: The cost function results obtained from the Uni ed Repartitioning Algorithm compared to the results obtained from optimized scratch-remap (LMSR) and multilevel di usion (WF) algorithms on 128 processors of a Cray T3E. Results on the Mesh from a Simulation of a Diesel Combustion Engine Table2 gives the cost function and average run time results for URA, the optimized scratch-remap algorithm (LMSR), and the optimized multilevel di usion algorithm (WF) on the diesel combustion engine test set for 8, 16, and 32 processors of a Cray T3E. The numbers at the top of each column indicate the Relative Cost Factor of the experiments in that column.... In PAGE 8: ... The numbers at the top of each column indicate the Relative Cost Factor of the experiments in that column. Table2 shows that URA minimizes the cost functions better than either of the other two schemes across the board. In this case, URA obtains somewhat better results than the scratch-remap scheme even when the Relative Cost Factor is set low.... In PAGE 8: ... These results con rm that the URA scheme is able to minimize the cost function as well as or better than current repartitioning schemes. Table2 shows that all three repartitioning schemes obtained similar average run time results on 8 and 16 processors. However, URA obtains signi cantly worse average run time results on 32 processors than either of the other schemes.... ..."

Cited by 23

### Table 2: The edge-cut, total amount of data redistribution, and cost function results of the adaptive graph partitioners WF, LMSR, and URA for various RCF values on problems derived from a particles-in-cells simulation on a Cray T3E. The numbers at the top of each column indicate the RCF of the experiments in that column.

2000

"... In PAGE 8: ... The other two schemes perform well only for limited ranges of values of the RCF. Results from a Simulation of a Diesel Combustion Engine Table2 gives the edge-cut, the total amount of data redistribution, and the cost function results for the optimized multilevel di usion algorithm (WF), the optimized scratch-remap algorithm (LMSR), and URA on the diesel combustion engine test set with 16 processors of a Cray T3E. The numbers at the top of each column indicate the Relative Cost Factor of the experiments in that column.... In PAGE 8: ... The numbers at the top of each column indicate the Relative Cost Factor of the experiments in that column. Table2 shows that across the board, URA computes partitionings with... In PAGE 9: ... These results con rm the trends shown in Figures 3 through 5. Table2 also shows that URA is able to tradeo one objective for another as the input value for RCF is changed. WF and LMSR compute the same partitioning regardless of the value for RCF.... ..."

Cited by 23

### Table 2: The edge-cut, total amount of data redistribution, and cost function results of the adaptive graph partitioners WF, LMSR, and URA for various RCF values on problems derived from a particles-in-cells simulation on a Cray T3E. The numbers at the top of each column indicate the RCF of the experiments in that column.

2000

"... In PAGE 7: ... The other twoschemes perform well only for limited ranges of values of the RCF. Results from a Simulation of a Diesel Combustion Engine Table2 gives the edge-cut, the total amount of data redistribution, and the cost function results for the optimized multilevel di usion algorithm (WF), the optimized scratch-remap algorithm (LMSR), and URA on the diesel combustion engine test set with 16 processors of a Cray T3E. The numbers at the top of each column indicate the RelativeCostFactor of the experiments in that column.... In PAGE 7: ... The numbers at the top of each column indicate the RelativeCostFactor of the experiments in that column. Table2 shows that across the board, URA computes partitionings with... In PAGE 8: ... These results con rm the trends shown in Figures 3 through 5. Table2 also shows that URA is able to tradeo one objective for another as the input value for RCF is changed. WF and LMSR compute the same partitioning regardless of the value for RCF.... ..."

Cited by 23

### Table 1: Results of Incremental Type Inference The man.ca program computes the Mandelbrot set using a dynamic algorithm. sim- ple.ca is the SIMPLE hydrodynamic simulation and p-i-c.ca is a particle-in-cell code. titest7.ca is a synthetic code designed to illustrate the algorithm apos;s e ectiveness and ap- pears in Appendix A. The mmult.ca program multiplies integer and oating point matrixes using a polymorphic library. poly.ca evaluates integer and oating point polynomials. The program tsp.ca solves the traveling salesman problem. quicksort.ca implements the quick- sort algorithm. queens.ca solves the N-queens problem, and t.ca computes a Fast Fourier Transform using a butter y network. All test cases were compiled with the standard CA prologue (240 lines of code) and are available along with the language manual [8] and the

1993

"... In PAGE 16: ... We have tested the type inference system on more than 20,000 lines of CA code. The results on a variety of real and synthetic programs appear in Table1 . precise refers to our incremental inference algorithm, palsberg refers to the inference algorithm in [14], and static refers to a basic constraint based inference which allocates exactly one type variable... ..."

Cited by 6

### Table 1: Results of Incremental Type Inference The man.ca program computes the Mandelbrot set using a dynamic algorithm. sim- ple.ca is the SIMPLE hydrodynamic simulation and p-i-c.ca is a particle-in-cell code. titest7.ca is a synthetic code designed to illustrate the algorithm apos;s e ectiveness and ap- pears in Appendix A. The mmult.ca program multiplies integer and oating point matrixes using a polymorphic library. poly.ca evaluates integer and oating point polynomials. The program tsp.ca solves the traveling salesman problem. quicksort.ca implements the quick- sort algorithm. queens.ca solves the N-queens problem, and t.ca computes a Fast Fourier Transform using a butter y network. All test cases were compiled with the standard CA prologue (240 lines of code) and are available along with the language manual [8] and the

1993

"... In PAGE 16: ... We have tested the type inference system on more than 20,000 lines of CA code. The results on a variety of real and synthetic programs appear in Table1 . precise refers to our incremental inference algorithm, palsberg refers to the inference algorithm in [14], and static refers to a basic constraint based inference which allocates exactly one type variable... ..."

Cited by 6

### Table 1. a.) A classification of core selection algorithms. M: multi-core (vs. single-core), C: constrained (vs. unconstrained), A: asymmetric (vs. symmetric), D: distributed (vs. centralized). b.) Computational and message exchange complexity of core selection algorithms. M, C and V as set handles denote respectively the sets of multicast group members, candidate cores and the entire nodes in the domain. hdiam and e are respectively the maximum hop-distance between any two domain nodes and the maximum node degree in the domain.

"... In PAGE 1: ... Latter research shifted to multi-core selection which, allowing multiple shared trees for distinct receiver partitions, further broadened the solution space not only for potential improvements on the efficiency, but also for the range of successful solutions for the delay-constrained case. In Table1 -a, we present a classification of core selection algorithms along with the complexity of each algorithm in the distributed, distance-vector routing environment. We primarily distinguish whether the algorithm is single-core or multi-core algorithm.... In PAGE 2: ... We classify the algorithm as distributed if its deployment with no reliance on an external message exchange protocol in the distributed platform is feasible, centralized otherwise. The entry of column Coordinator in Table1 -b is the computational complexity of the process operating on the designated node coordinating the core selection process. For distributed algorithms, we also analyze the proportional growth of message exchange throughout the entire algorithm and present our results in respective columns of the table.... ..."

### Table 3: Performance for particle-in-

2005

"... In PAGE 7: ... Therefore, we use point sprite textures of size 322 for all applications as an optimal compromise between texture resolution and speed. Table3 shows performance measurements for particle-in-cell ad- vection. The size of the property texture is identical to the number of particles.... ..."

Cited by 3

### Table 1. Efficiency of the Partition evaluate heuristic.

2002

"... In PAGE 3: ... This a59 a10a62a61 a63 a14 algorithm significantly reduces the computation performed. Table1 presents some experimental data on the efficiency of Par- tition evaluate for partition-space pruning. The number of possible unique partitions a0 a1 a10 a4a3 a14 (estimated using a0 a1 a10 a4a3 a14 a53a24 a6 a7a9a8a11a10 a1a13a12 a14a116a1a13a16 a6 a18a17 a12 ) is presented for several values of a3 and a5 .... In PAGE 3: ... We choose p21241 to illustrate the efficiency of our heuristics, because the exhaustive method [8] was found to be inadequate for wrapper/TAM co-design for p21241; the method did not complete even for a5a54a24a31a32 . From Table1 , it can be seen that Partition evaluate evaluates on average only 2% of the unique partitions. Thus there is a significant reduction in the execu-... ..."

Cited by 20

### Table 3. Comparison of efficiency and bandwidth properties.

2006

"... In PAGE 7: ... A full discussion of these issues is beyond the scope of the present paper. In Table3 we answer the following questions, at the 80-bit and 256-bit secu- rity levels. (Notes to the table can be found in the appendix.... ..."

Cited by 3

### Table 3. Comparison of efficiency and bandwidth properties.

2006

"... In PAGE 7: ... A full discussion of these issues is beyond the scope of the present paper. In Table3 we answer the following questions, at the 80-bit and 256-bit secu- rity levels. (Notes to the table can be found in the appendix.... ..."

Cited by 3