11.07.2015 Views

ALGORITHMS FOR SOLVING LINEAR AND POLYNOMIAL ...

ALGORITHMS FOR SOLVING LINEAR AND POLYNOMIAL ...

ALGORITHMS FOR SOLVING LINEAR AND POLYNOMIAL ...

SHOW MORE
SHOW LESS
  • No tags were found...

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

5.2 Rapid Subspace Enumeration . . . . . . . . . . . . . . . . . . . . . . 785.3 The Four Russians Matrix Multiplication Algorithm . . . . . . . . . . 805.3.1 Role of the Gray Code . . . . . . . . . . . . . . . . . . . . . . 815.3.2 Transposing the Matrix Product . . . . . . . . . . . . . . . . . 825.3.3 Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . 825.3.4 A Quick Computation . . . . . . . . . . . . . . . . . . . . . . 825.3.5 M4RM Experiments Performed by SAGE Staff . . . . . . . . . 835.4 The Four Russians Matrix Inversion Algorithm . . . . . . . . . . . . . 845.4.1 Stage 1: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 865.4.2 Stage 2: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 865.4.3 Stage 3: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 865.4.4 A Curious Note on Stage 1 of M4RI . . . . . . . . . . . . . . . 875.4.5 Triangulation or Inversion? . . . . . . . . . . . . . . . . . . . . 905.5 Experimental and Numerical Results . . . . . . . . . . . . . . . . . . 915.6 Exact Analysis of Complexity . . . . . . . . . . . . . . . . . . . . . . 965.6.1 An Alternative Computation . . . . . . . . . . . . . . . . . . . 975.6.2 Full Elimination, not Triangular . . . . . . . . . . . . . . . . . 985.6.3 The Rank of 3k Rows, or Why k + ɛ is not Enough . . . . . . 995.6.4 Using Bulk Logical Operations . . . . . . . . . . . . . . . . . . 1015.6.5 M4RI Experiments Performed by SAGE Staff . . . . . . . . . 1025.6.5.1 Determination of k . . . . . . . . . . . . . . . . . . . 1025.6.5.2 Comparison to Magma . . . . . . . . . . . . . . . . . 1025.6.5.3 The Transpose Experiment . . . . . . . . . . . . . . 1035.7 Pairing With Strassen’s Algorithm for Matrix Multiplication . . . . . 1035.8 The Unsuitability of Strassen’s Algorithm for Inversion . . . . . . . . 1055.8.1 Bunch and Hopcroft’s Solution . . . . . . . . . . . . . . . . . 1075.8.2 Ibara, Moran, and Hui’s Solution . . . . . . . . . . . . . . . . 1086 An Impractical Method of Accelerating Matrix Operations in Rings of FiniteSize 1136.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1136.1.1 Feasibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1146.2 The Algorithm over a Finite Ring . . . . . . . . . . . . . . . . . . . . 1156.2.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1166.2.2 Complexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1166.2.3 Taking Advantage of z ≠ 1 . . . . . . . . . . . . . . . . . . . . 1186.2.4 The Transpose of Matrix Multiplication . . . . . . . . . . . . 1186.3 Choosing Values of b . . . . . . . . . . . . . . . . . . . . . . . . . . . 1196.3.1 The “Conservative” Algorithm . . . . . . . . . . . . . . . . . . 1196.3.2 The “Liberal” Algorithm . . . . . . . . . . . . . . . . . . . . . 1206.3.3 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1226.4 Over Finite Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1226.4.1 Complexity Penalty . . . . . . . . . . . . . . . . . . . . . . . . 1236.4.2 Memory Requirements . . . . . . . . . . . . . . . . . . . . . . 1236.4.3 Time Requirements . . . . . . . . . . . . . . . . . . . . . . . . 124xiii

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!