Fix all spelling errors reported at

Reference-LAPACK#518 (comment)
Many thanks to @jschleus

Author: weslleyspereira

BLAS/SRC/srotmg.f

@@ -191,7 +191,7 @@ SUBROUTINE SROTMG(SD1,SD2,SX1,SY1,SPARAM)
             END IF
          END IF
 
-*     PROCESURE..SCALE-CHECK
+*     PROCEDURE..SCALE-CHECK
          IF (SD1.NE.ZERO) THEN
             DO WHILE ((SD1.LE.RGAMSQ) .OR. (SD1.GE.GAMSQ))
                IF (SFLAG.EQ.ZERO) THEN

CMakeLists.txt

@@ -10,7 +10,7 @@ set(
   ${LAPACK_MAJOR_VERSION}.${LAPACK_MINOR_VERSION}.${LAPACK_PATCH_VERSION}
   )
 
-# Add the CMake directory for custon CMake modules
+# Add the CMake directory for custom CMake modules
 set(CMAKE_MODULE_PATH "${LAPACK_SOURCE_DIR}/CMAKE" ${CMAKE_MODULE_PATH})
 
 # Export all symbols on Windows when building shared libraries

SRC/cgejsv.f

@@ -368,7 +368,7 @@
 *>                    singular values might be lost.
 *>
 *>          If full SVD is needed, the following two condition numbers are
-*>          useful for the analysis of the algorithm. They are provied for
+*>          useful for the analysis of the algorithm. They are provided for
 *>          a developer/implementer who is familiar with the details of
 *>          the method.
 *>
@@ -1314,7 +1314,7 @@ SUBROUTINE CGEJSV( JOBA, JOBU, JOBV, JOBR, JOBT, JOBP,
 *     (eg speed by replacing global with restricted window pivoting, such
 *     as in xGEQPX from TOMS # 782). Good results will be obtained using
 *     xGEQPX with properly (!) chosen numerical parameters.
-*     Any improvement of CGEQP3 improves overal performance of CGEJSV.
+*     Any improvement of CGEQP3 improves overall performance of CGEJSV.
 *
 *     A * P1 = Q1 * [ R1^* 0]^*:
       DO 1963 p = 1, N

SRC/cgesvdq.f

@@ -315,7 +315,7 @@
 *>          N^(-1/4) * RWORK(1) <= ||pinv(C)||_2 <= N^(1/4) * RWORK(1).
 *>          Otherwise, RWORK(1) = -1.
 *>          2. RWORK(2) contains the number of singular values computed as
-*>          exact zeros in CGESVD applied to the upper triangular or trapeziodal
+*>          exact zeros in CGESVD applied to the upper triangular or trapezoidal
 *>          R (from the initial QR factorization). In case of early exit (no call to
 *>          CGESVD, such as in the case of zero matrix) RWORK(2) = -1.
 *>
@@ -640,7 +640,7 @@ SUBROUTINE CGESVDQ( JOBA, JOBP, JOBR, JOBU, JOBV, M, N, A, LDA,
                  IF ( WNTVA ) THEN
 *                   .. minimal workspace length for N x N/2 CGEQRF
                     LWQRF  = MAX( N/2, 1 )
-*                   .. minimal workspace lengt for N/2 x N/2 CGESVD
+*                   .. minimal workspace lengthfor N/2 x N/2 CGESVD
                     LWSVD2 = MAX( 3 * (N/2), 1 )
                     LWUNQ2 = MAX( N, 1 )
                     MINWRK2 = MAX( LWQP3, N/2+LWQRF, N/2+LWSVD2,

SRC/cgesvj.f

@@ -296,7 +296,7 @@
 *> spectral condition number. The best performance of this Jacobi SVD
 *> procedure is achieved if used in an  accelerated version of Drmac and
 *> Veselic [4,5], and it is the kernel routine in the SIGMA library [6].
-*> Some tunning parameters (marked with [TP]) are available for the
+*> Some tuning parameters (marked with [TP]) are available for the
 *> implementer.
 *> The computational range for the nonzero singular values is the  machine
 *> number interval ( UNDERFLOW , OVERFLOW ). In extreme cases, even

SRC/cgsvj1.f

@@ -40,7 +40,7 @@
 *> CGSVJ1 is called from CGESVJ as a pre-processor and that is its main
 *> purpose. It applies Jacobi rotations in the same way as CGESVJ does, but
 *> it targets only particular pivots and it does not check convergence
-*> (stopping criterion). Few tunning parameters (marked by [TP]) are
+*> (stopping criterion). Few tuning parameters (marked by [TP]) are
 *> available for the implementer.
 *>
 *> Further Details
@@ -61,7 +61,7 @@
 *> In terms of the columns of A, the first N1 columns are rotated 'against'
 *> the remaining N-N1 columns, trying to increase the angle between the
 *> corresponding subspaces. The off-diagonal block is N1-by(N-N1) and it is
-*> tiled using quadratic tiles of side KBL. Here, KBL is a tunning parameter.
+*> tiled using quadratic tiles of side KBL. Here, KBL is a tuning parameter.
 *> The number of sweeps is given in NSWEEP and the orthogonality threshold
 *> is given in TOL.
 *> \endverbatim

SRC/chetrd_hb2st.F

@@ -395,7 +395,7 @@ SUBROUTINE CHETRD_HB2ST( STAGE1, VECT, UPLO, N, KD, AB, LDAB,
 *     For that, for real precision just copy the diag and offdiag 
 *     to D and E while for the COMPLEX case the bulge chasing is  
 *     performed to convert the hermetian tridiagonal to symmetric 
-*     tridiagonal. A simpler coversion formula might be used, but then 
+*     tridiagonal. A simpler conversion formula might be used, but then 
 *     updating the Q matrix will be required and based if Q is generated
 *     or not this might complicate the story. 
 *      

SRC/chetrf_rook.f

@@ -35,7 +35,7 @@
 *>
 *> \verbatim
 *>
-*> CHETRF_ROOK computes the factorization of a comlex Hermitian matrix A
+*> CHETRF_ROOK computes the factorization of a complex Hermitian matrix A
 *> using the bounded Bunch-Kaufman ("rook") diagonal pivoting method.
 *> The form of the factorization is
 *>

SRC/cla_gbrfsx_extended.f

@@ -368,7 +368,7 @@
 *>          DZ_UB is REAL
 *>     Determines when to start considering componentwise convergence.
 *>     Componentwise convergence is only considered after each component
-*>     of the solution Y is stable, which we definte as the relative
+*>     of the solution Y is stable, which we define as the relative
 *>     change in each component being less than DZ_UB. The default value
 *>     is 0.25, requiring the first bit to be stable. See LAWN 165 for
 *>     more details.

SRC/cla_gerfsx_extended.f

@@ -356,7 +356,7 @@
 *>          DZ_UB is REAL
 *>     Determines when to start considering componentwise convergence.
 *>     Componentwise convergence is only considered after each component
-*>     of the solution Y is stable, which we definte as the relative
+*>     of the solution Y is stable, which we define as the relative
 *>     change in each component being less than DZ_UB. The default value
 *>     is 0.25, requiring the first bit to be stable. See LAWN 165 for
 *>     more details.