Select which methods should be included in the table (and the graph).
If the type of the data is not selected (VAR), the table will be comprehensive, but the graph will be useless.
In order to get a graph, select either the degree or the bitsize to constrain the sets used.
Graph color (for the eps)
Mono
Color

Select methods:
Sturm
Sleeve
Symbolic Numeric
CF
NCFF
NCF
Descartes
Bitstream Descartes
RS
Select data attributes:


Type:
Degree:
Bitsize:

Plot Range (vertical)
From:
To:

Plot Range (horizontal)
From:
To:
The following combinations are feasible:
Type Degree Bitsize
3D Arrangments 12 200,400,600,800,1000,
1200,1400,1600,1800,2000
Voronoi of ellipses 184 3500
Random 3,10,30,50,100 10,20,30,40,50,2000,4000,6000,8000
Random 5 10,20,30,40,50
Random 70 2000,4000,6000,8000
Rational 3,5,10,30,50,100 10,20,30,40,50
Integer 3,5,10,30,50,100 10,20,30,40,50
Mignotte 3,5,10,30,50,100 10,20,30,40,50


An analysis of the results and full description of the benchmarks is available in
"Experimental evaluation and cross-benchmarking of univariate real solvers",
Hemmer M., Tsigaridas E.P., Zafeirakopoulos Z., Emiris I.Z., Karavelas M. and Mourrain B.
Final version to appear in Workshop Symb. Numeric Comput. '09

The datasets used during the benchmarks are available from the following links:
random
ratroots
introots
Mignotte
3D Arragnements
Voronoi diagram of ellipses
Random with big bitsize



Author: Zafeirakopoulos Zafeirakis