------------------------------------------------------------------------------ -- -- -- GNAT RUN-TIME COMPONENTS -- -- -- -- I N T E R F A C E S . F O R T R A N . B L A S -- -- -- -- S p e c -- -- -- -- Copyright (C) 2006-2009, Free Software Foundation, Inc. -- -- -- -- GNAT is free software; you can redistribute it and/or modify it under -- -- terms of the GNU General Public License as published by the Free Soft- -- -- ware Foundation; either version 3, or (at your option) any later ver- -- -- sion. GNAT is distributed in the hope that it will be useful, but WITH- -- -- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY -- -- or FITNESS FOR A PARTICULAR PURPOSE. -- -- -- -- As a special exception under Section 7 of GPL version 3, you are granted -- -- additional permissions described in the GCC Runtime Library Exception, -- -- version 3.1, as published by the Free Software Foundation. -- -- -- -- You should have received a copy of the GNU General Public License and -- -- a copy of the GCC Runtime Library Exception along with this program; -- -- see the files COPYING3 and COPYING.RUNTIME respectively. If not, see -- -- . -- -- -- -- GNAT was originally developed by the GNAT team at New York University. -- -- Extensive contributions were provided by Ada Core Technologies Inc. -- -- -- ------------------------------------------------------------------------------ -- This package provides a thin binding to the standard Fortran BLAS library. -- Documentation and a reference BLAS implementation is available from -- ftp://ftp.netlib.org. The main purpose of this package is to facilitate -- implementation of the Ada 2005 Ada.Numerics.Generic_Real_Arrays and -- Ada.Numerics.Generic_Complex_Arrays packages. Bindings to other BLAS -- routines may be added over time. -- As actual linker arguments to link with the BLAS implementation differs -- according to platform and chosen BLAS implementation, the linker arguments -- are given in the body of this package. The body may need to be modified in -- order to link with different BLAS implementations tuned to the specific -- target. package Interfaces.Fortran.BLAS is pragma Pure; pragma Elaborate_Body; No_Trans : aliased constant Character := 'N'; Trans : aliased constant Character := 'T'; Conj_Trans : aliased constant Character := 'C'; -- Vector types type Real_Vector is array (Integer range <>) of Real; type Complex_Vector is array (Integer range <>) of Complex; type Double_Precision_Vector is array (Integer range <>) of Double_Precision; type Double_Complex_Vector is array (Integer range <>) of Double_Complex; -- Matrix types type Real_Matrix is array (Integer range <>, Integer range <>) of Real; type Double_Precision_Matrix is array (Integer range <>, Integer range <>) of Double_Precision; type Complex_Matrix is array (Integer range <>, Integer range <>) of Complex; type Double_Complex_Matrix is array (Integer range <>, Integer range <>) of Double_Complex; -- BLAS Level 1 function sdot (N : Positive; X : Real_Vector; Inc_X : Integer := 1; Y : Real_Vector; Inc_Y : Integer := 1) return Real; function ddot (N : Positive; X : Double_Precision_Vector; Inc_X : Integer := 1; Y : Double_Precision_Vector; Inc_Y : Integer := 1) return Double_Precision; function cdotu (N : Positive; X : Complex_Vector; Inc_X : Integer := 1; Y : Complex_Vector; Inc_Y : Integer := 1) return Complex; function zdotu (N : Positive; X : Double_Complex_Vector; Inc_X : Integer := 1; Y : Double_Complex_Vector; Inc_Y : Integer := 1) return Double_Complex; function snrm2 (N : Natural; X : Real_Vector; Inc_X : Integer := 1) return Real; function dnrm2 (N : Natural; X : Double_Precision_Vector; Inc_X : Integer := 1) return Double_Precision; function scnrm2 (N : Natural; X : Complex_Vector; Inc_X : Integer := 1) return Real; function dznrm2 (N : Natural; X : Double_Complex_Vector; Inc_X : Integer := 1) return Double_Precision; -- BLAS Level 2 procedure sgemv (Trans : access constant Character; M : Natural := 0; N : Natural := 0; Alpha : Real := 1.0; A : Real_Matrix; Ld_A : Positive; X : Real_Vector; Inc_X : Integer := 1; -- must be non-zero Beta : Real := 0.0; Y : in out Real_Vector; Inc_Y : Integer := 1); -- must be non-zero procedure dgemv (Trans : access constant Character; M : Natural := 0; N : Natural := 0; Alpha : Double_Precision := 1.0; A : Double_Precision_Matrix; Ld_A : Positive; X : Double_Precision_Vector; Inc_X : Integer := 1; -- must be non-zero Beta : Double_Precision := 0.0; Y : in out Double_Precision_Vector; Inc_Y : Integer := 1); -- must be non-zero procedure cgemv (Trans : access constant Character; M : Natural := 0; N : Natural := 0; Alpha : Complex := (1.0, 1.0); A : Complex_Matrix; Ld_A : Positive; X : Complex_Vector; Inc_X : Integer := 1; -- must be non-zero Beta : Complex := (0.0, 0.0); Y : in out Complex_Vector; Inc_Y : Integer := 1); -- must be non-zero procedure zgemv (Trans : access constant Character; M : Natural := 0; N : Natural := 0; Alpha : Double_Complex := (1.0, 1.0); A : Double_Complex_Matrix; Ld_A : Positive; X : Double_Complex_Vector; Inc_X : Integer := 1; -- must be non-zero Beta : Double_Complex := (0.0, 0.0); Y : in out Double_Complex_Vector; Inc_Y : Integer := 1); -- must be non-zero -- BLAS Level 3 procedure sgemm (Trans_A : access constant Character; Trans_B : access constant Character; M : Positive; N : Positive; K : Positive; Alpha : Real := 1.0; A : Real_Matrix; Ld_A : Integer; B : Real_Matrix; Ld_B : Integer; Beta : Real := 0.0; C : in out Real_Matrix; Ld_C : Integer); procedure dgemm (Trans_A : access constant Character; Trans_B : access constant Character; M : Positive; N : Positive; K : Positive; Alpha : Double_Precision := 1.0; A : Double_Precision_Matrix; Ld_A : Integer; B : Double_Precision_Matrix; Ld_B : Integer; Beta : Double_Precision := 0.0; C : in out Double_Precision_Matrix; Ld_C : Integer); procedure cgemm (Trans_A : access constant Character; Trans_B : access constant Character; M : Positive; N : Positive; K : Positive; Alpha : Complex := (1.0, 1.0); A : Complex_Matrix; Ld_A : Integer; B : Complex_Matrix; Ld_B : Integer; Beta : Complex := (0.0, 0.0); C : in out Complex_Matrix; Ld_C : Integer); procedure zgemm (Trans_A : access constant Character; Trans_B : access constant Character; M : Positive; N : Positive; K : Positive; Alpha : Double_Complex := (1.0, 1.0); A : Double_Complex_Matrix; Ld_A : Integer; B : Double_Complex_Matrix; Ld_B : Integer; Beta : Double_Complex := (0.0, 0.0); C : in out Double_Complex_Matrix; Ld_C : Integer); private pragma Import (Fortran, cdotu, "cdotu_"); pragma Import (Fortran, cgemm, "cgemm_"); pragma Import (Fortran, cgemv, "cgemv_"); pragma Import (Fortran, ddot, "ddot_"); pragma Import (Fortran, dgemm, "dgemm_"); pragma Import (Fortran, dgemv, "dgemv_"); pragma Import (Fortran, dnrm2, "dnrm2_"); pragma Import (Fortran, dznrm2, "dznrm2_"); pragma Import (Fortran, scnrm2, "scnrm2_"); pragma Import (Fortran, sdot, "sdot_"); pragma Import (Fortran, sgemm, "sgemm_"); pragma Import (Fortran, sgemv, "sgemv_"); pragma Import (Fortran, snrm2, "snrm2_"); pragma Import (Fortran, zdotu, "zdotu_"); pragma Import (Fortran, zgemm, "zgemm_"); pragma Import (Fortran, zgemv, "zgemv_"); end Interfaces.Fortran.BLAS;