Intel IPP for small matrices supports floating-point data with single and double precision.
Copyright © 2000 - 2011, Intel Corporation. All rights reserved.