Gerben J. Hekstra

Learn More
Within Philips Research Labs, we are investigating the 64-bit VLIW core for future TriMedia processors. We have performed an extensive Design Space Exploration (DSE) on this core using quantitative analysis, using a benchmark suite of applications which are representative for multi-media processing. We have explored, among others, the configurations of the(More)
In this paper an algorithm and architecture for computing the eigenvalue decomposition (EVD) of a symmetric matrix is presented. The EVD is computed using a Jacobi-type method, where the angle of the rotations is approximated by an angle k , corresponding to an orthonormal {rotation. These orthonormal {rotations are based on the idea of CORDIC and share the(More)
At Philips Research Labs, we are investigating the 64-bit VLIW core (also called CPU64) for future TriMedia processors. This processor is targeted towards embedded mul-timedia applications. In order to be able to perform a quantitative design space exploration, a set of benchmark applications has been developed which is representative of the application(More)
In this paper, we present a full precision floating-point Cordic algorithm and a correponding word-serial Cordic architecture. The extension to existing block floating-point Cordic algorithms is in a floating-point representation for the angle. The angle is represented as a combination of exponent , micro-rotation bits and two bits to indicate pre-rotations(More)
CORDIC based IIR digital lters are orthogonal lters whose internal computations consist of orthogonal transformations. These lters possess desirable properties for VLSI implementations such as regularity, local connection, low sensitivity to nite word-length implementation, and elimination of limit cycles. Recently, ne-grain pipelined CORDIC based IIR(More)
In this paper the implementation of the SVD{updating algorithm using orthonormal { rotations is presented. An orthonormal {rotation is a rotation by an angle of a given set of {rotation angles (e.g. the angles i = arctan2 ?i) which are choosen such that the rotation can be implemented by a small amount of shift{add operations. A version of the SVD{updating(More)
| One can nd ample examples in the literature of implementations of image transforms such as the discrete cosine transform and the lapped orthogonal transform. The objective is invariantly the minimization of the number of multiplies and adds. Of course, a reduction of operations from O(N 2)to O(NlogN) is a great achievement, yet the cost resulting from(More)