CUDA Pro Tip: Fast and Robust Computation of Givens Rotations

Originally published at:

A Givens rotation [1] represents a rotation in a plane represented by a matrix of the form , where the intersections of the th and th columns contain the values and . Multiplying a vector by a Givens rotation matrix represents a rotation of the vector in the plane by radians. According to Wikipedia, the…

One of the first investigations of the performance of a Fast Givens Rotation algorithm to calculate a QR factorization can be found in, "Benchmarking the NVIDIA 8800GTX with the CUDA Development Platform" (McGraw-Herdeg, et al, 2007 MIT/Lincoln Labs HPEC Workshop). QR factorization is part of the HPEC Challenge Benchmark Suite. Having a reciprocal hypotenuse function will be of great benefit to a lot of signal and image processing applications. The HPEC abstract can be found here:

Cool stuff. Any plans to make a complex given rotation implementation?