Matrix: Stevenson/LargeRegFile

Description: large register file, digital circuit optimization

 (bipartite graph drawing)

• Home page of the UF Sparse Matrix Collection
• Matrix group: Stevenson
• Click here for a description of the Stevenson group.
• Click here for a list of all matrices
• Click here for a list of all matrix groups
• download as a MATLAB mat-file, file size: 8 MB. Use UFget(2273) or UFget('Stevenson/LargeRegFile') in MATLAB.
• download in Matrix Market format, file size: 16 MB.
• download in Rutherford/Boeing format, file size: 12 MB.

 Matrix properties number of rows 2,111,154 number of columns 801,374 nonzeros 4,944,201 structural full rank? yes structural rank 801,374 # of blocks from dmperm 32,778 # strongly connected comp. 2 explicit zero entries 0 nonzero pattern symmetry 0% numeric value symmetry 0% type real structure rectangular Cholesky candidate? no positive definite? no

 author J. Stevenson editor T. Davis date 2010 kind circuit simulation problem 2D/3D problem? no

Notes:

```Circuit optimization for digital circuits.  John Peter Stevenson, Stanford.

LargeRegFile is a matrix representing an optimization problem.
Specifically, it is a geometric program, a kind of convex optimization
problem.  The matrix entries are the exponents found in the geometric
program constraints; specifically, this matrix is a portion of the
input required by mskexpopt:
http://docs.mosek.com/6.0/capi/node008.html

This matrix represents a digital circuit.  The circuit is a large
register file: 256 registers, 64 bits each, 2 read ports, 1 write port.
The optimization problem is to minimize circuit delay, subject to
constraints on energy cost and area cost. The optimization variables
are the transistor size, the supply voltage, and the threshold voltage.
The circuit optimization method is described by Patil, et. al.: Robust
energy-efficient adder topologies. In ARITH07: Proceedings of the 18th
IEEE Symposium on Computer Arithmetic, pages 16-28, Washington, DC,
USA, 2007. IEEE Computer Society.
```

 Ordering statistics: result nnz(V) for QR, upper bound nnz(L) for LU, with COLAMD 916,786,879 nnz(R) for QR, upper bound nnz(U) for LU, with COLAMD 4,902,416

For a description of the statistics displayed above, click here.

Maintained by Tim Davis, last updated 12-Mar-2014.
Matrix pictures by cspy, a MATLAB function in the CSparse package.
Matrix graphs by Yifan Hu, AT&T Labs Visualization Group.