P_i

PCGBTRF()

factor main partition P_i a_i = l_i u_i on each processo
lbwl, lbwu: lower and upper bandwidth of local solver

PDGBTRF()

factor main partition P_i a_i = l_i u_i on each processo
lbwl, lbwu: lower and upper bandwidth of local solver

PSGBTRF()

factor main partition P_i a_i = l_i u_i on each processo
lbwl, lbwu: lower and upper bandwidth of local solver

PZGBTRF()

factor main partition P_i a_i = l_i u_i on each processo
lbwl, lbwu: lower and upper bandwidth of local solver

Pack

PCDBTRF()

Pack params and positions into arrays for global consistency chec

PCDBTRSV()

Pack params and positions into arrays for global consistency chec

PCDTTRF()

Pack params and positions into arrays for global consistency chec

PCDTTRSV()

Pack params and positions into arrays for global consistency chec

PCGBTRF()

Pack params and positions into arrays for global consistency chec

PCPBTRF()

Pack params and positions into arrays for global consistency chec

PCPBTRSV()

Pack params and positions into arrays for global consistency chec

PCPTTRF()

Pack params and positions into arrays for global consistency chec

PCPTTRSV()

Pack params and positions into arrays for global consistency chec

PDDBTRF()

Pack params and positions into arrays for global consistency chec

PDDBTRSV()

Pack params and positions into arrays for global consistency chec

PDDTTRF()

Pack params and positions into arrays for global consistency chec

PDDTTRSV()

Pack params and positions into arrays for global consistency chec

PDGBTRF()

Pack params and positions into arrays for global consistency chec

PDPBTRF()

Pack params and positions into arrays for global consistency chec

PDPBTRSV()

Pack params and positions into arrays for global consistency chec

PDPTTRF()

Pack params and positions into arrays for global consistency chec

PDPTTRSV()

Pack params and positions into arrays for global consistency chec

PSDBTRF()

Pack params and positions into arrays for global consistency chec

PSDBTRSV()

Pack params and positions into arrays for global consistency chec

PSDTTRF()

Pack params and positions into arrays for global consistency chec

PSDTTRSV()

Pack params and positions into arrays for global consistency chec

PSGBTRF()

Pack params and positions into arrays for global consistency chec

PSPBTRF()

Pack params and positions into arrays for global consistency chec

PSPBTRSV()

Pack params and positions into arrays for global consistency chec

PSPTTRF()

Pack params and positions into arrays for global consistency chec

PSPTTRSV()

Pack params and positions into arrays for global consistency chec

PZDBTRF()

Pack params and positions into arrays for global consistency chec

PZDBTRSV()

Pack params and positions into arrays for global consistency chec

PZDTTRF()

Pack params and positions into arrays for global consistency chec

PZDTTRSV()

Pack params and positions into arrays for global consistency chec

PZGBTRF()

Pack params and positions into arrays for global consistency chec

PZPBTRF()

Pack params and positions into arrays for global consistency chec

PZPBTRSV()

Pack params and positions into arrays for global consistency chec

PZPTTRF()

Pack params and positions into arrays for global consistency chec

PZPTTRSV()

Pack params and positions into arrays for global consistency chec

packed

PCDBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PCDBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PCGBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PCGBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PCPBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PCPBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PDDBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PDDBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PDGBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PDGBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PDPBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PDPBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PSDBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PSDBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PSGBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PSGBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PSPBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PSPBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PZDBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PZDBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PZGBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PZGBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PZPBSV()

on entry, this array contains the local pieces of the
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

PZPBTRS()

l^t a(1:n, ja:ja+n-1).
this local portion is stored in the packed banded forma
scalapack manual for more detail on the format of

pair

DLASORTE()

dlasorte sorts eigenpairs so that real eigenpairs are together an
since every 2nd subdiagonal is guaranteed to be zero.

PCGBTRF()

bm1 = m for 1st block on proc pair, bm2 2nd bloc

PDGBTRF()

bm1 = m for 1st block on proc pair, bm2 2nd bloc

PSGBTRF()

bm1 = m for 1st block on proc pair, bm2 2nd bloc

PZGBTRF()

bm1 = m for 1st block on proc pair, bm2 2nd bloc

SLASORTE()

slasorte sorts eigenpairs so that real eigenpairs are together an
since every 2nd subdiagonal is guaranteed to be zero.

paired

DLASORTE()

this is set if the input matrix had an odd number of real
eigenvalues and things couldn't be paired or if the inpu
0 indicates successful completion.

SLASORTE()

this is set if the input matrix had an odd number of real
eigenvalues and things couldn't be paired or if the inpu
0 indicates successful completion.

panel

PJLAENV()

= 1: the data layout blocksize;
= 2: the panel blocking factor
= 4: execution path control;

Parallel

DLASORTE()

since every 2nd subdiagonal is guaranteed to be zero.
this routine does no Parallel work
arguments

PCDBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCDBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCDTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCDTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCGBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCGBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCGETF2()

this is the right-looking Parallel level 2 blas version of th

PCGETRF()

this is the right-looking Parallel level 3 blas version of th

PCHEEVD()

reference:  f. tisseur and j. dongarra, "a Parallel divide an
on distributed memory architectures",

PCHEEVX()

see "on the correctness of Parallel bisection in floatin

PCHEGVX()

see "on the correctness of Parallel bisection in floatin

PCHENTRD()

pchentrd is a prototype version of pchetrd which uses tailored
codes (either the serial, chetrd, or the Parallel code, pchettrd

PCLACP3()

pclacp3 is an auxiliary routine that copies from a global Parallel
the entire submatrix that is copied gets placed on one node or

PCPBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCPBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCPTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCPTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PCSTEIN()

pcstein computes the eigenvectors of a symmetric tridiagonal matrix
in Parallel, using inverse iteration. the eigenvectors foun
orthogonalize vectors that are on different processes. the extent

PCTREVC()

pctrevc computes some or all of the right and/or left eigenvectors of
a complex upper triangular matrix t in Parallel
the right eigenvector x and the left eigenvector y of t corresponding

PDDBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDDBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDDTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDDTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDGBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDGBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDGETF2()

this is the right-looking Parallel level 2 blas version of th

PDGETRF()

this is the right-looking Parallel level 3 blas version of th

PDLACP3()

pdlacp3 is an auxiliary routine that copies from a global Parallel
the entire submatrix that is copied gets placed on one node or

PDLAED1()

matrix after modification by a rank-one symmetric matrix,
in Parallel
t = q(in) ( d(in) + rho * z*z' ) q'(in) = q(out) * d(out) * q'(out)

PDPBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDPBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDPTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDPTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PDSTEBZ()

pdstebz computes the eigenvalues of a symmetric tridiagonal matrix in
Parallel. the user may ask for all eigenvalues, all eigenvalues i
static partitioning of work is done at the beginning of pdstebz which

PDSTEDC()

pdstedc computes all eigenvalues and eigenvectors of a
symmetric tridiagonal matrix in Parallel, using the divide an

PDSTEIN()

pdstein computes the eigenvectors of a symmetric tridiagonal matrix
in Parallel, using inverse iteration. the eigenvectors foun
orthogonalize vectors that are on different processes. the extent

PDSYEVD()

reference:  f. tisseur and j. dongarra, "a Parallel divide an
on distributed memory architectures",

PDSYEVX()

see "on the correctness of Parallel bisection in floatin

PDSYGVX()

see "on the correctness of Parallel bisection in floatin

PDSYNTRD()

pdsyntrd is a prototype version of pdsytrd which uses tailored
codes (either the serial, dsytrd, or the Parallel code, pdsyttrd

PSDBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSDBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSDTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSDTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSGBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSGBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSGETF2()

this is the right-looking Parallel level 2 blas version of th

PSGETRF()

this is the right-looking Parallel level 3 blas version of th

PSLACP3()

pslacp3 is an auxiliary routine that copies from a global Parallel
the entire submatrix that is copied gets placed on one node or

PSLAED1()

matrix after modification by a rank-one symmetric matrix,
in Parallel
t = q(in) ( d(in) + rho * z*z' ) q'(in) = q(out) * d(out) * q'(out)

PSPBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSPBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSPTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSPTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PSSTEBZ()

psstebz computes the eigenvalues of a symmetric tridiagonal matrix in
Parallel. the user may ask for all eigenvalues, all eigenvalues i
static partitioning of work is done at the beginning of psstebz which

PSSTEDC()

psstedc computes all eigenvalues and eigenvectors of a
symmetric tridiagonal matrix in Parallel, using the divide an

PSSTEIN()

psstein computes the eigenvectors of a symmetric tridiagonal matrix
in Parallel, using inverse iteration. the eigenvectors foun
orthogonalize vectors that are on different processes. the extent

PSSYEVD()

reference:  f. tisseur and j. dongarra, "a Parallel divide an
on distributed memory architectures",

PSSYEVX()

see "on the correctness of Parallel bisection in floatin

PSSYGVX()

see "on the correctness of Parallel bisection in floatin

PSSYNTRD()

pssyntrd is a prototype version of pssytrd which uses tailored
codes (either the serial, ssytrd, or the Parallel code, pssyttrd

PZDBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZDBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZDTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZDTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZGBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZGBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZGETF2()

this is the right-looking Parallel level 2 blas version of th

PZGETRF()

this is the right-looking Parallel level 3 blas version of th

PZHEEVD()

reference:  f. tisseur and j. dongarra, "a Parallel divide an
on distributed memory architectures",

PZHEEVX()

see "on the correctness of Parallel bisection in floatin

PZHEGVX()

see "on the correctness of Parallel bisection in floatin

PZHENTRD()

pzhentrd is a prototype version of pzhetrd which uses tailored
codes (either the serial, zhetrd, or the Parallel code, pzhettrd

PZLACP3()

pzlacp3 is an auxiliary routine that copies from a global Parallel
the entire submatrix that is copied gets placed on one node or

PZPBSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZPBTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZPTSV()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZPTTRS()

the mapping for matrices must be blocked, reflecting the nature
of the divide and conquer algorithm as a task-Parallel algorithm
chunk of the matrix.

PZSTEIN()

pzstein computes the eigenvectors of a symmetric tridiagonal matrix
in Parallel, using inverse iteration. the eigenvectors foun
orthogonalize vectors that are on different processes. the extent

PZTREVC()

pztrevc computes some or all of the right and/or left eigenvectors of
a complex upper triangular matrix t in Parallel
the right eigenvector x and the left eigenvector y of t corresponding

SLASORTE()

since every 2nd subdiagonal is guaranteed to be zero.
this routine does no Parallel work
arguments

parallelism

PCLAHQR()

communication can be broken into several parts for
efficient parallelism
loop over all the bulges, just doing the work

PZLAHQR()

communication can be broken into several parts for
efficient parallelism
loop over all the bulges, just doing the work

parallelization

PCHETTRD()

details of the parallelization
we delay spreading v across to all processor columns (which

PCLASMSUB()

this code is basically a parallelization of the following sni

PDLASMSUB()

this code is basically a parallelization of the following sni

PDSYTTRD()

details of the parallelization
we delay spreading v across to all processor columns (which

PSLASMSUB()

this code is basically a parallelization of the following sni

PSSYTTRD()

details of the parallelization
we delay spreading v across to all processor columns (which

PZHETTRD()

details of the parallelization
we delay spreading v across to all processor columns (which

PZLASMSUB()

this code is basically a parallelization of the following sni

parameter

PCHEEVX()

workspace needs are larger for pcheevx.
liwork parameter adde
orfac, icluster() and gap() parameters added

PCSTEIN()

orthogonalize vectors that are on different processes. the extent
of orthogonalization is controlled by the input parameter lwork
process. pcstein decides on the allocation of work among the

PDSTEBZ()

point arithmetic.
cure: increase the parameter "fudge", recompile

PDSTEIN()

orthogonalize vectors that are on different processes. the extent
of orthogonalization is controlled by the input parameter lwork
process. pdstein decides on the allocation of work among the

PDSYEVX()

workspace needs are larger for pdsyevx.
liwork parameter adde
orfac, icluster() and gap() parameters added

PJLAENV()

tailored eigen-routines to choose
problem-dependent parameters for the local environment.  see ispe

PSSTEBZ()

point arithmetic.
cure: increase the parameter "fudge", recompile

PSSTEIN()

orthogonalize vectors that are on different processes. the extent
of orthogonalization is controlled by the input parameter lwork
process. psstein decides on the allocation of work among the

PSSYEVX()

workspace needs are larger for pssyevx.
liwork parameter adde
orfac, icluster() and gap() parameters added

PZHEEVX()

workspace needs are larger for pzheevx.
liwork parameter adde
orfac, icluster() and gap() parameters added

PZSTEIN()

orthogonalize vectors that are on different processes. the extent
of orthogonalization is controlled by the input parameter lwork
process. pzstein decides on the allocation of work among the

parameters

CDBTF2()

.. parameters .

CDBTRF()

test the input parameters

CLAMSH()

.. parameters .

CLANV2()

sn      (output) complex
parameters of the rotation matrix
further details

DDBTF2()

.. parameters .

DDBTRF()

test the input parameters

DLAMSH ()

.. parameters .

DLASORTE()

.. parameters .

DSTEIN2()

test the input parameters

DSTEQR2()

test the input parameters

PCDBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCDBTRF()

test the input parameters

PCDBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCDBTRSV()

test the input parameters

PCDTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCDTTRF()

test the input parameters

PCDTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCDTTRSV()

test the input parameters

PCGBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCGBTRF()

test the input parameters

PCGBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCGEBD2()

.. parameters .

PCGEBRD()

.. parameters .

PCGECON()

.. parameters .

PCGEEQU()

.. parameters .

PCGEHD2()

.. parameters .

PCGEHRD()

.. parameters .

PCGELQ2()

.. parameters .

PCGELQF()

.. parameters .

PCGELS()

.. parameters .

PCGEQL2()

.. parameters .

PCGEQLF()

.. parameters .

PCGEQPF()

.. parameters .

PCGEQR2()

.. parameters .

PCGEQRF()

.. parameters .

PCGERFS()

internal parameters

PCGERQ2()

.. parameters .

PCGERQF()

.. parameters .

PCGESV()

.. parameters .

PCGESVX()

.. parameters .

PCGETF2()

.. parameters .

PCGETRF()

.. parameters .

PCGETRI()

.. parameters .

PCGETRS()

.. parameters .

PCGGQRF()

.. parameters .

PCGGRQF()

.. parameters .

PCHEEV()

on output, work(1) returns the workspace needed to guarantee
completion.  if the input parameters are incorrect, work(1

PCHEEVD()

on output rwork(1) returns the real workspace needed to
guarantee completion.  if the input parameters are incorrect

PCHEEVX()

orfac, icluster() and gap() parameters adde

PCHEGS2()

.. parameters .

PCHEGST()

.. parameters .

PCHEGVX()

.. parameters .

PCHENGST()

.. parameters .

PCHENTRD()

.. parameters .

PCHETD2()

.. parameters .

PCHETRD()

.. parameters .

PCHETTRD()

.. parameters .

PCLABRD()

.. parameters .

PCLACGV()

.. parameters .

PCLACON()

where a' is the conjugate transpose of a, and pclacon must
be re-called with all the other parameters unchanged
ix      (global input) integer

PCLACONSB()

.. parameters .

PCLACP2()

.. parameters .

PCLACP3()

.. parameters .

PCLACPY()

.. parameters .

PCLAEVSWP()

lrwork   (local input) integer dimension of rwork
.. parameters .

PCLAHRD()

.. parameters .

PCLAMR1D()

.. parameters .

PCLANGE()

.. parameters .

PCLANHE()

get grid parameters and local indexes

PCLANHS()

get grid parameters

PCLANSY()

get grid parameters and local indexes

PCLANTR()

get grid parameters

PCLAPIV()

.. parameters .

PCLAPV2()

.. parameters .

PCLAQGE()

internal parameters

PCLAQSY()

internal parameters

PCLARF()

get grid parameters

PCLARFB()

.. parameters .

PCLARFC()

get grid parameters

PCLARFG()

.. parameters .

PCLARFT()

.. parameters .

PCLARZ()

get grid parameters

PCLARZB()

.. parameters .

PCLARZC()

get grid parameters

PCLARZT()

.. parameters .

PCLASCL()

.. parameters .

PCLASE2()

.. parameters .

PCLASET()

.. parameters .

PCLASMSUB()

.. parameters .

PCLASSQ()

.. parameters .

PCLASWP()

.. parameters .

PCLATRA()

.. parameters .

PCLATRD()

.. parameters .

PCLATRS()

get grid parameters

PCLATRZ()

.. parameters .

PCLATTRS()

test the input parameters

PCLAUU2()

.. parameters .

PCLAUUM()

.. parameters .

PCLAWIL()

.. parameters .

PCMAX1()

.. parameters .

PCPBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCPBTRF()

test the input parameters

PCPBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCPBTRSV()

test the input parameters

PCPOCON()

.. parameters .

PCPOEQU()

.. parameters .

PCPORFS()

internal parameters

PCPOSV()

.. parameters .

PCPOSVX()

.. parameters .

PCPOTF2()

.. parameters .

PCPOTRF()

.. parameters .

PCPOTRI()

.. parameters .

PCPOTRS()

.. parameters .

PCPTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCPTTRF()

test the input parameters

PCPTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PCPTTRSV()

test the input parameters

PCSRSCL()

.. parameters .

PCTRCON()

.. parameters .

PCTREVC()

.. parameters .

PCTRRFS()

.. parameters .

PCTRTI2()

.. parameters .

PCTRTRI()

.. parameters .

PCTRTRS()

.. parameters .

PCTZRZF()

.. parameters .

PCUNG2L()

.. parameters .

PCUNG2R()

.. parameters .

PCUNGL2()

.. parameters .

PCUNGLQ()

.. parameters .

PCUNGQL()

.. parameters .

PCUNGQR()

.. parameters .

PCUNGR2()

.. parameters .

PCUNGRQ()

.. parameters .

PCUNM2L()

.. parameters .

PCUNM2R()

.. parameters .

PCUNMBR()

.. parameters .

PCUNMHR()

.. parameters .

PCUNML2()

.. parameters .

PCUNMLQ()

.. parameters .

PCUNMQL()

.. parameters .

PCUNMQR()

.. parameters .

PCUNMR2()

.. parameters .

PCUNMR3()

.. parameters .

PCUNMRQ()

.. parameters .

PCUNMRZ()

.. parameters .

PCUNMTR()

.. parameters .

PDDBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDDBTRF()

test the input parameters

PDDBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDDBTRSV()

test the input parameters

PDDTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDDTTRF()

test the input parameters

PDDTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDDTTRSV()

test the input parameters

PDGBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDGBTRF()

test the input parameters

PDGBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDGEBD2()

.. parameters .

PDGEBRD()

.. parameters .

PDGECON()

.. parameters .

PDGEEQU()

.. parameters .

PDGEHD2()

.. parameters .

PDGEHRD()

.. parameters .

PDGELQ2()

.. parameters .

PDGELQF()

.. parameters .

PDGELS()

.. parameters .

PDGEQL2()

.. parameters .

PDGEQLF()

.. parameters .

PDGEQPF()

.. parameters .

PDGEQR2()

.. parameters .

PDGEQRF()

.. parameters .

PDGERFS()

internal parameters

PDGERQ2()

.. parameters .

PDGERQF()

.. parameters .

PDGESV()

.. parameters .

PDGESVX()

.. parameters .

PDGETF2()

.. parameters .

PDGETRF()

.. parameters .

PDGETRI()

.. parameters .

PDGETRS()

.. parameters .

PDGGQRF()

.. parameters .

PDGGRQF()

.. parameters .

PDLABRD()

.. parameters .

PDLACON()

a' * x,  if kase=2,
pdlacon must be re-called with all the other parameters

PDLACONSB()

.. parameters .

PDLACP2()

.. parameters .

PDLACP3()

.. parameters .

PDLACPY()

.. parameters .

PDLAED0()

.. parameters .

PDLAED1()

.. parameters .

PDLAED2()

.. parameters .

PDLAED3()

.. parameters .

PDLAEVSWP()

lwork   (local input) integer dimension of work
.. parameters .

PDLAHRD()

.. parameters .

PDLAMCH()

pdlamch determines double precision machine parameters
arguments

PDLAMR1D()

.. parameters .

PDLANGE()

.. parameters .

PDLANHS()

get grid parameters

PDLANSY()

get grid parameters and local indexes

PDLANTR()

get grid parameters

PDLAPIV()

.. parameters .

PDLAPV2()

.. parameters .

PDLAQGE()

internal parameters

PDLAQSY()

internal parameters

PDLARED1D()

.. parameters .

PDLARED2D()

.. parameters .

PDLARF()

get grid parameters

PDLARFB()

.. parameters .

PDLARFG()

.. parameters .

PDLARFT()

.. parameters .

PDLARZ()

get grid parameters

PDLARZB()

.. parameters .

PDLARZT()

.. parameters .

PDLASCL()

.. parameters .

PDLASE2()

.. parameters .

PDLASET()

.. parameters .

PDLASMSUB()

.. parameters .

PDLASRT()

.. parameters .

PDLASSQ()

.. parameters .

PDLASWP()

.. parameters .

PDLATRA()

.. parameters .

PDLATRD()

.. parameters .

PDLATRS()

get grid parameters

PDLATRZ()

.. parameters .

PDLAUU2()

.. parameters .

PDLAUUM()

.. parameters .

PDLAWIL()

.. parameters .

PDORG2L()

.. parameters .

PDORG2R()

.. parameters .

PDORGL2()

.. parameters .

PDORGLQ()

.. parameters .

PDORGQL()

.. parameters .

PDORGQR()

.. parameters .

PDORGR2()

.. parameters .

PDORGRQ()

.. parameters .

PDORM2L()

.. parameters .

PDORM2R()

.. parameters .

PDORMBR()

.. parameters .

PDORMHR()

.. parameters .

PDORML2()

.. parameters .

PDORMLQ()

.. parameters .

PDORMQL()

.. parameters .

PDORMQR()

.. parameters .

PDORMR2()

.. parameters .

PDORMR3()

.. parameters .

PDORMRQ()

.. parameters .

PDORMRZ()

.. parameters .

PDORMTR()

.. parameters .

PDPBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDPBTRF()

test the input parameters

PDPBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDPBTRSV()

test the input parameters

PDPOCON()

.. parameters .

PDPOEQU()

.. parameters .

PDPORFS()

internal parameters

PDPOSV()

.. parameters .

PDPOSVX()

.. parameters .

PDPOTF2()

.. parameters .

PDPOTRF()

.. parameters .

PDPOTRI()

.. parameters .

PDPOTRS()

.. parameters .

PDPTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDPTTRF()

test the input parameters

PDPTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PDPTTRSV()

test the input parameters

PDRSCL()

.. parameters .

PDSTEBZ()

internal parameters

PDSTEDC()

.. parameters .

PDSYEV()

needed to guarantee completion.
if the input parameters are incorrect, work(1) may also b

PDSYEVD()

.. parameters .

PDSYEVX()

orfac, icluster() and gap() parameters adde

PDSYGS2()

.. parameters .

PDSYGST()

.. parameters .

PDSYGVX()

.. parameters .

PDSYNGST()

.. parameters .

PDSYNTRD()

.. parameters .

PDSYTD2()

.. parameters .

PDSYTRD()

.. parameters .

PDSYTTRD()

.. parameters .

PDTRCON()

.. parameters .

PDTRRFS()

.. parameters .

PDTRTI2()

.. parameters .

PDTRTRI()

.. parameters .

PDTRTRS()

.. parameters .

PDTZRZF()

.. parameters .

PDZSUM1()

parameters

PJLAENV()

tailored eigen-routines to choose
problem-dependent parameters for the local environment.  see ispe

PSCSUM1()

parameters

PSDBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSDBTRF()

test the input parameters

PSDBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSDBTRSV()

test the input parameters

PSDTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSDTTRF()

test the input parameters

PSDTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSDTTRSV()

test the input parameters

PSGBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSGBTRF()

test the input parameters

PSGBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSGEBD2()

.. parameters .

PSGEBRD()

.. parameters .

PSGECON()

.. parameters .

PSGEEQU()

.. parameters .

PSGEHD2()

.. parameters .

PSGEHRD()

.. parameters .

PSGELQ2()

.. parameters .

PSGELQF()

.. parameters .

PSGELS()

.. parameters .

PSGEQL2()

.. parameters .

PSGEQLF()

.. parameters .

PSGEQPF()

.. parameters .

PSGEQR2()

.. parameters .

PSGEQRF()

.. parameters .

PSGERFS()

internal parameters

PSGERQ2()

.. parameters .

PSGERQF()

.. parameters .

PSGESV()

.. parameters .

PSGESVX()

.. parameters .

PSGETF2()

.. parameters .

PSGETRF()

.. parameters .

PSGETRI()

.. parameters .

PSGETRS()

.. parameters .

PSGGQRF()

.. parameters .

PSGGRQF()

.. parameters .

PSLABRD()

.. parameters .

PSLACON()

a' * x,  if kase=2,
pslacon must be re-called with all the other parameters

PSLACONSB()

.. parameters .

PSLACP2()

.. parameters .

PSLACP3()

.. parameters .

PSLACPY()

.. parameters .

PSLAED0()

.. parameters .

PSLAED1()

.. parameters .

PSLAED2()

.. parameters .

PSLAED3()

.. parameters .

PSLAEVSWP()

lwork   (local input) integer dimension of work
.. parameters .

PSLAHRD()

.. parameters .

PSLAMCH()

pslamch determines single precision machine parameters
arguments

PSLAMR1D()

.. parameters .

PSLANGE()

.. parameters .

PSLANHS()

get grid parameters

PSLANSY()

get grid parameters and local indexes

PSLANTR()

get grid parameters

PSLAPIV()

.. parameters .

PSLAPV2()

.. parameters .

PSLAQGE()

internal parameters

PSLAQSY()

internal parameters

PSLARED1D()

.. parameters .

PSLARED2D()

.. parameters .

PSLARF()

get grid parameters

PSLARFB()

.. parameters .

PSLARFG()

.. parameters .

PSLARFT()

.. parameters .

PSLARZ()

get grid parameters

PSLARZB()

.. parameters .

PSLARZT()

.. parameters .

PSLASCL()

.. parameters .

PSLASE2()

.. parameters .

PSLASET()

.. parameters .

PSLASMSUB()

.. parameters .

PSLASRT()

.. parameters .

PSLASSQ()

.. parameters .

PSLASWP()

.. parameters .

PSLATRA()

.. parameters .

PSLATRD()

.. parameters .

PSLATRS()

get grid parameters

PSLATRZ()

.. parameters .

PSLAUU2()

.. parameters .

PSLAUUM()

.. parameters .

PSLAWIL()

.. parameters .

PSORG2L()

.. parameters .

PSORG2R()

.. parameters .

PSORGL2()

.. parameters .

PSORGLQ()

.. parameters .

PSORGQL()

.. parameters .

PSORGQR()

.. parameters .

PSORGR2()

.. parameters .

PSORGRQ()

.. parameters .

PSORM2L()

.. parameters .

PSORM2R()

.. parameters .

PSORMBR()

.. parameters .

PSORMHR()

.. parameters .

PSORML2()

.. parameters .

PSORMLQ()

.. parameters .

PSORMQL()

.. parameters .

PSORMQR()

.. parameters .

PSORMR2()

.. parameters .

PSORMR3()

.. parameters .

PSORMRQ()

.. parameters .

PSORMRZ()

.. parameters .

PSORMTR()

.. parameters .

PSPBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSPBTRF()

test the input parameters

PSPBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSPBTRSV()

test the input parameters

PSPOCON()

.. parameters .

PSPOEQU()

.. parameters .

PSPORFS()

internal parameters

PSPOSV()

.. parameters .

PSPOSVX()

.. parameters .

PSPOTF2()

.. parameters .

PSPOTRF()

.. parameters .

PSPOTRI()

.. parameters .

PSPOTRS()

.. parameters .

PSPTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSPTTRF()

test the input parameters

PSPTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PSPTTRSV()

test the input parameters

PSRSCL()

.. parameters .

PSSTEBZ()

internal parameters

PSSTEDC()

.. parameters .

PSSYEV()

needed to guarantee completion.
if the input parameters are incorrect, work(1) may also b

PSSYEVD()

.. parameters .

PSSYEVX()

orfac, icluster() and gap() parameters adde

PSSYGS2()

.. parameters .

PSSYGST()

.. parameters .

PSSYGVX()

.. parameters .

PSSYNGST()

.. parameters .

PSSYNTRD()

.. parameters .

PSSYTD2()

.. parameters .

PSSYTRD()

.. parameters .

PSSYTTRD()

.. parameters .

PSTRCON()

.. parameters .

PSTRRFS()

.. parameters .

PSTRTI2()

.. parameters .

PSTRTRI()

.. parameters .

PSTRTRS()

.. parameters .

PSTZRZF()

.. parameters .

PZDBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZDBTRF()

test the input parameters

PZDBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZDBTRSV()

test the input parameters

PZDRSCL()

.. parameters .

PZDTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZDTTRF()

test the input parameters

PZDTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZDTTRSV()

test the input parameters

PZGBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZGBTRF()

test the input parameters

PZGBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZGEBD2()

.. parameters .

PZGEBRD()

.. parameters .

PZGECON()

.. parameters .

PZGEEQU()

.. parameters .

PZGEHD2()

.. parameters .

PZGEHRD()

.. parameters .

PZGELQ2()

.. parameters .

PZGELQF()

.. parameters .

PZGELS()

.. parameters .

PZGEQL2()

.. parameters .

PZGEQLF()

.. parameters .

PZGEQPF()

.. parameters .

PZGEQR2()

.. parameters .

PZGEQRF()

.. parameters .

PZGERFS()

internal parameters

PZGERQ2()

.. parameters .

PZGERQF()

.. parameters .

PZGESV()

.. parameters .

PZGESVX()

.. parameters .

PZGETF2()

.. parameters .

PZGETRF()

.. parameters .

PZGETRI()

.. parameters .

PZGETRS()

.. parameters .

PZGGQRF()

.. parameters .

PZGGRQF()

.. parameters .

PZHEEV()

on output, work(1) returns the workspace needed to guarantee
completion.  if the input parameters are incorrect, work(1

PZHEEVD()

on output rwork(1) returns the real workspace needed to
guarantee completion.  if the input parameters are incorrect

PZHEEVX()

orfac, icluster() and gap() parameters adde

PZHEGS2()

.. parameters .

PZHEGST()

.. parameters .

PZHEGVX()

.. parameters .

PZHENGST()

.. parameters .

PZHENTRD()

.. parameters .

PZHETD2()

.. parameters .

PZHETRD()

.. parameters .

PZHETTRD()

.. parameters .

PZLABRD()

.. parameters .

PZLACGV()

.. parameters .

PZLACON()

where a' is the conjugate transpose of a, and pzlacon must
be re-called with all the other parameters unchanged
ix      (global input) integer

PZLACONSB()

.. parameters .

PZLACP2()

.. parameters .

PZLACP3()

.. parameters .

PZLACPY()

.. parameters .

PZLAEVSWP()

lrwork   (local input) integer dimension of rwork
.. parameters .

PZLAHRD()

.. parameters .

PZLAMR1D()

.. parameters .

PZLANGE()

.. parameters .

PZLANHE()

get grid parameters and local indexes

PZLANHS()

get grid parameters

PZLANSY()

get grid parameters and local indexes

PZLANTR()

get grid parameters

PZLAPIV()

.. parameters .

PZLAPV2()

.. parameters .

PZLAQGE()

internal parameters

PZLAQSY()

internal parameters

PZLARF()

get grid parameters

PZLARFB()

.. parameters .

PZLARFC()

get grid parameters

PZLARFG()

.. parameters .

PZLARFT()

.. parameters .

PZLARZ()

get grid parameters

PZLARZB()

.. parameters .

PZLARZC()

get grid parameters

PZLARZT()

.. parameters .

PZLASCL()

.. parameters .

PZLASE2()

.. parameters .

PZLASET()

.. parameters .

PZLASMSUB()

.. parameters .

PZLASSQ()

.. parameters .

PZLASWP()

.. parameters .

PZLATRA()

.. parameters .

PZLATRD()

.. parameters .

PZLATRS()

get grid parameters

PZLATRZ()

.. parameters .

PZLATTRS()

test the input parameters

PZLAUU2()

.. parameters .

PZLAUUM()

.. parameters .

PZLAWIL()

.. parameters .

PZMAX1()

.. parameters .

PZPBSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZPBTRF()

test the input parameters

PZPBTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZPBTRSV()

test the input parameters

PZPOCON()

.. parameters .

PZPOEQU()

.. parameters .

PZPORFS()

internal parameters

PZPOSV()

.. parameters .

PZPOSVX()

.. parameters .

PZPOTF2()

.. parameters .

PZPOTRF()

.. parameters .

PZPOTRI()

.. parameters .

PZPOTRS()

.. parameters .

PZPTSV()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZPTTRF()

test the input parameters

PZPTTRS()

the following are restrictions on the input parameters. some of thes
may reflect fundamental technical limitations.

PZPTTRSV()

test the input parameters

PZTRCON()

.. parameters .

PZTREVC()

.. parameters .

PZTRRFS()

.. parameters .

PZTRTI2()

.. parameters .

PZTRTRI()

.. parameters .

PZTRTRS()

.. parameters .

PZTZRZF()

.. parameters .

PZUNG2L()

.. parameters .

PZUNG2R()

.. parameters .

PZUNGL2()

.. parameters .

PZUNGLQ()

.. parameters .

PZUNGQL()

.. parameters .

PZUNGQR()

.. parameters .

PZUNGR2()

.. parameters .

PZUNGRQ()

.. parameters .

PZUNM2L()

.. parameters .

PZUNM2R()

.. parameters .

PZUNMBR()

.. parameters .

PZUNMHR()

.. parameters .

PZUNML2()

.. parameters .

PZUNMLQ()

.. parameters .

PZUNMQL()

.. parameters .

PZUNMQR()

.. parameters .

PZUNMR2()

.. parameters .

PZUNMR3()

.. parameters .

PZUNMRQ()

.. parameters .

PZUNMRZ()

.. parameters .

PZUNMTR()

.. parameters .

SDBTF2()

.. parameters .

SDBTRF()

test the input parameters

SLAMSH ()

.. parameters .

SLASORTE()

.. parameters .

SSTEIN2()

test the input parameters

SSTEQR2()

test the input parameters

ZDBTF2()

.. parameters .

ZDBTRF()

test the input parameters

ZLAMSH()

.. parameters .

ZLANV2()

sn      (output) complex*16
parameters of the rotation matrix
further details

params

PCDBTRF()

pack params and positions into arrays for global consistency chec

PCDBTRSV()

pack params and positions into arrays for global consistency chec

PCDTTRF()

pack params and positions into arrays for global consistency chec

PCDTTRSV()

pack params and positions into arrays for global consistency chec

PCGBTRF()

pack params and positions into arrays for global consistency chec

PCPBTRF()

pack params and positions into arrays for global consistency chec

PCPBTRSV()

pack params and positions into arrays for global consistency chec

PCPTTRF()

pack params and positions into arrays for global consistency chec

PCPTTRSV()

pack params and positions into arrays for global consistency chec

PDDBTRF()

pack params and positions into arrays for global consistency chec

PDDBTRSV()

pack params and positions into arrays for global consistency chec

PDDTTRF()

pack params and positions into arrays for global consistency chec

PDDTTRSV()

pack params and positions into arrays for global consistency chec

PDGBTRF()

pack params and positions into arrays for global consistency chec

PDPBTRF()

pack params and positions into arrays for global consistency chec

PDPBTRSV()

pack params and positions into arrays for global consistency chec

PDPTTRF()

pack params and positions into arrays for global consistency chec

PDPTTRSV()

pack params and positions into arrays for global consistency chec

PSDBTRF()

pack params and positions into arrays for global consistency chec

PSDBTRSV()

pack params and positions into arrays for global consistency chec

PSDTTRF()

pack params and positions into arrays for global consistency chec

PSDTTRSV()

pack params and positions into arrays for global consistency chec

PSGBTRF()

pack params and positions into arrays for global consistency chec

PSPBTRF()

pack params and positions into arrays for global consistency chec

PSPBTRSV()

pack params and positions into arrays for global consistency chec

PSPTTRF()

pack params and positions into arrays for global consistency chec

PSPTTRSV()

pack params and positions into arrays for global consistency chec

PZDBTRF()

pack params and positions into arrays for global consistency chec

PZDBTRSV()

pack params and positions into arrays for global consistency chec

PZDTTRF()

pack params and positions into arrays for global consistency chec

PZDTTRSV()

pack params and positions into arrays for global consistency chec

PZGBTRF()

pack params and positions into arrays for global consistency chec

PZPBTRF()

pack params and positions into arrays for global consistency chec

PZPBTRSV()

pack params and positions into arrays for global consistency chec

PZPTTRF()

pack params and positions into arrays for global consistency chec

PZPTTRSV()

pack params and positions into arrays for global consistency chec

paramters

DLAPST()

test the input paramters

DLASRT2()

test the input paramters

SLAPST()

test the input paramters

SLASRT2()

test the input paramters

paranoid

PDLAEBZ()

pivmin  (input) double precision
the minimum absolute of a "pivot" in the "paranoid
least max_j |e(j)^2| *safe_min, and at least safe_min, where

PDLAPDCT()

the innermost loop to avoid overflow and determine the sign of a
floating point number. pdlapdct will be referred to as the "paranoid

PSLAEBZ()

pivmin  (input) real
the minimum absolute of a "pivot" in the "paranoid
least max_j |e(j)^2| *safe_min, and at least safe_min, where

PSLAPDCT()

the innermost loop to avoid overflow and determine the sign of a
floating point number. pslapdct will be referred to as the "paranoid

part

CCOMBAMAX1 ()

ccombamax1 finds the element having maximum real part absolut

CDBTRF()

the active part of the matrix is partitione
a11   a12   a13

CTRMVT()

before entry with  uplo = 'u' or 'u', the leading n by n
upper triangular part of the array t must contain the uppe
t is not referenced.

DDBTRF()

the active part of the matrix is partitione
a11   a12   a13

DTRMVT()

before entry with  uplo = 'u' or 'u', the leading n by n
upper triangular part of the array t must contain the uppe
t is not referenced.

PCDBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCDBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCDTSV()

dl      (local input/local output) complex pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PCDTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCDTTRS()

dl      (local input/local output) complex pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PCDTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCGBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCHEEV()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCHEEVD()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCHEEVX()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCHEGS2()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PCHEGST()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PCHEGVX()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
leading n-by-n lower triangular part of sub( a ) contains

PCHENGST()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PCHENTRD()

pchentrd is faster than pchetrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PCHETD2()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCHETRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCHETTRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCLABRD()

returns the matrices x and y which are needed to apply the transfor-
mation to the unreduced part of sub( a )
if m >= n, sub( a ) is reduced to upper bidiagonal form; if m < n, to

PCLACONSB()

(desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PCLACP2()

pclacp2 copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PCLACPY()

pclacpy copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PCLAHQR()

more clever.  we break each transformation down into 3
parts
a group of rotn transformations (this is on

PCLAHRD()

v which is needed, with t and y, to apply the transformation to the
unreduced part of the matrix, using an update of the form

PCLANHE()

icurcol : process column containing diagonal block
irsc0   : pointer to part of work used to store the rowsums whil
irsr0   : pointer to part of work used to store the rowsums after

PCLANSY()

icurcol : process column containing diagonal block
irsc0   : pointer to part of work used to store the rowsums whil
irsr0   : pointer to part of work used to store the rowsums after

PCLAQSY()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCLARZB()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PCLASE2()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PCLASET()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PCLASMSUB()

a       (global input) complex array, dimension (desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PCLATRD()

q' * sub( a ) * q, and returns the matrices v and w which are
needed to apply the transformation to the unreduced part of sub( a )
if uplo = 'u', pclatrd reduces the last nb rows and columns of a

PCLATRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors. l > 0
a       (local input/local output) complex pointer into the

PCLAUU2()

pclauu2 computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PCLAUUM()

pclauum computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PCMAX1()

if incx = m_x, then sub( x ) is a vector distributed over a process
row. each process part of this row receives the result
if incx = 1, then sub( x ) is a vector distributed over a process

PCPBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCPBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCPORFS()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PCPOSV()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PCPOSVX()

diag(sr)*a*diag(sc).  if uplo = 'u', the leading
n-by-n upper triangular part of a contains the uppe
triangular part of a is not referenced.  if uplo = 'l', the

PCPOTF2()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PCPOTRF()

n-by-n hermitian distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PCPTSV()

d       (local input/local output) complex pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PCPTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCPTTRS()

d       (local input/local output) complex pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PCPTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PCTRCON()

matrix a(ia:ia+n-1,ja:ja+n-1). if uplo = 'u', the leading
n-by-n upper triangular part of this distributed matrix con
triangular part is not referenced.  if uplo = 'l', the

PCTRRFS()

distributed matrix sub( a ).
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PCTRTI2()

sub( a ). if uplo = 'u', the leading n-by-n upper triangular
part of the matrix sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PCTRTRI()

triangular matrix sub( a ).  if uplo = 'u', the leading
n-by-n upper triangular part of the matrix sub( a ) contain
lower triangular part of sub( a ) is not referenced.

PCTRTRS()

matrix sub( a ).  if uplo = 'u', the leading n-by-n upper
triangular part of sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PCTZRZF()

sub( a ) which is to be factored. on exit, the leading m-by-m
upper triangular part of sub( a ) contains the upper trian
sub( a ), with the array tau, represent the unitary matrix z

PCUNMR3()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PCUNMRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PDDBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDDBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDDTSV()

dl      (local input/local output) double precision pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PDDTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDDTTRS()

dl      (local input/local output) double precision pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PDDTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDGBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDLABRD()

and returns the matrices x and y which are needed to apply the
transformation to the unreduced part of sub( a )
if m >= n, sub( a ) is reduced to upper bidiagonal form; if m < n, to

PDLACONSB()

(desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PDLACP2()

pdlacp2 copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PDLACPY()

pdlacpy copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PDLAHRD()

v which is needed, with t and y, to apply the transformation to the
unreduced part of the matrix, using an update of the form

PDLANSY()

icurcol : process column containing diagonal block
irsc0   : pointer to part of work used to store the rowsums whil
irsr0   : pointer to part of work used to store the rowsums after

PDLAQSY()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDLARZB()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PDLASE2()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PDLASET()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PDLASMSUB()

(desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PDLATRD()

and returns the matrices v and w which are needed to apply the
transformation to the unreduced part of sub( a )
if uplo = 'u', pdlatrd reduces the last nb rows and columns of a

PDLATRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors. l > 0
a       (local input/local output) double precision pointer into the

PDLAUU2()

pdlauu2 computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PDLAUUM()

pdlauum computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PDORMR3()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PDORMRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PDPBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDPBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDPORFS()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDPOSV()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PDPOSVX()

diag(sr)*a*diag(sc).  if uplo = 'u', the leading
n-by-n upper triangular part of a contains the uppe
triangular part of a is not referenced.  if uplo = 'l', the

PDPOTF2()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PDPOTRF()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PDPTSV()

d       (local input/local output) double precision pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PDPTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDPTTRS()

d       (local input/local output) double precision pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PDPTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PDSYEV()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDSYEVD()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDSYEVX()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDSYGS2()

n-by-n symmetric distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PDSYGST()

n-by-n symmetric distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PDSYGVX()

n-by-n symmetric distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
leading n-by-n lower triangular part of sub( a ) contains

PDSYNGST()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PDSYNTRD()

pdsyntrd is faster than pdsytrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PDSYTD2()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDSYTRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDSYTTRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PDTRCON()

matrix a(ia:ia+n-1,ja:ja+n-1). if uplo = 'u', the leading
n-by-n upper triangular part of this distributed matrix con
triangular part is not referenced.  if uplo = 'l', the

PDTRRFS()

distributed matrix sub( a ).
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PDTRTI2()

sub( a ). if uplo = 'u', the leading n-by-n upper triangular
part of the matrix sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PDTRTRI()

triangular matrix sub( a ).  if uplo = 'u', the leading
n-by-n upper triangular part of the matrix sub( a ) contain
lower triangular part of sub( a ) is not referenced.

PDTRTRS()

matrix sub( a ).  if uplo = 'u', the leading n-by-n upper
triangular part of sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PDTZRZF()

sub( a ) which is to be factored. on exit, the leading m-by-m
upper triangular part of sub( a ) contains the upper trian
sub( a ), with the array tau, represent the orthogonal matrix

PDZSUM1()

if incx = m_x, then sub( x ) is a vector distributed over a process
row. each process part of this row receives the result
if incx = 1, then sub( x ) is a vector distributed over a process

PSCSUM1()

if incx = m_x, then sub( x ) is a vector distributed over a process
row. each process part of this row receives the result
if incx = 1, then sub( x ) is a vector distributed over a process

PSDBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSDBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSDTSV()

dl      (local input/local output) real pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PSDTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSDTTRS()

dl      (local input/local output) real pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PSDTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSGBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSLABRD()

and returns the matrices x and y which are needed to apply the
transformation to the unreduced part of sub( a )
if m >= n, sub( a ) is reduced to upper bidiagonal form; if m < n, to

PSLACONSB()

(desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PSLACP2()

pslacp2 copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PSLACPY()

pslacpy copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PSLAHRD()

v which is needed, with t and y, to apply the transformation to the
unreduced part of the matrix, using an update of the form

PSLANSY()

icurcol : process column containing diagonal block
irsc0   : pointer to part of work used to store the rowsums whil
irsr0   : pointer to part of work used to store the rowsums after

PSLAQSY()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSLARZB()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PSLASE2()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PSLASET()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PSLASMSUB()

(desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PSLATRD()

and returns the matrices v and w which are needed to apply the
transformation to the unreduced part of sub( a )
if uplo = 'u', pslatrd reduces the last nb rows and columns of a

PSLATRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors. l > 0
a       (local input/local output) real pointer into the

PSLAUU2()

pslauu2 computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PSLAUUM()

pslauum computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PSORMR3()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PSORMRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PSPBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSPBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSPORFS()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSPOSV()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PSPOSVX()

diag(sr)*a*diag(sc).  if uplo = 'u', the leading
n-by-n upper triangular part of a contains the uppe
triangular part of a is not referenced.  if uplo = 'l', the

PSPOTF2()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PSPOTRF()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PSPTSV()

d       (local input/local output) real pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PSPTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSPTTRS()

d       (local input/local output) real pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PSPTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PSSYEV()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSSYEVD()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSSYEVX()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSSYGS2()

n-by-n symmetric distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PSSYGST()

n-by-n symmetric distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PSSYGVX()

n-by-n symmetric distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
leading n-by-n lower triangular part of sub( a ) contains

PSSYNGST()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PSSYNTRD()

pssyntrd is faster than pssytrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PSSYTD2()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSSYTRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSSYTTRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PSTRCON()

matrix a(ia:ia+n-1,ja:ja+n-1). if uplo = 'u', the leading
n-by-n upper triangular part of this distributed matrix con
triangular part is not referenced.  if uplo = 'l', the

PSTRRFS()

distributed matrix sub( a ).
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PSTRTI2()

sub( a ). if uplo = 'u', the leading n-by-n upper triangular
part of the matrix sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PSTRTRI()

triangular matrix sub( a ).  if uplo = 'u', the leading
n-by-n upper triangular part of the matrix sub( a ) contain
lower triangular part of sub( a ) is not referenced.

PSTRTRS()

matrix sub( a ).  if uplo = 'u', the leading n-by-n upper
triangular part of sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PSTZRZF()

sub( a ) which is to be factored. on exit, the leading m-by-m
upper triangular part of sub( a ) contains the upper trian
sub( a ), with the array tau, represent the orthogonal matrix

PZDBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZDBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZDTSV()

dl      (local input/local output) complex*16 pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PZDTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZDTTRS()

dl      (local input/local output) complex*16 pointer to local
part of global vector storing the lower diagonal of th
aligned with d.

PZDTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZGBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZHEEV()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZHEEVD()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZHEEVX()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZHEGS2()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PZHEGST()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PZHEGVX()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
leading n-by-n lower triangular part of sub( a ) contains

PZHENGST()

n-by-n hermitian distributed matrix sub( a ). if uplo = 'u',
the leading n-by-n upper triangular part of sub( a ) contain
lower triangular part is not referenced.  if uplo = 'l', the

PZHENTRD()

pzhentrd is faster than pzhetrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PZHETD2()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZHETRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZHETTRD()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZLABRD()

returns the matrices x and y which are needed to apply the transfor-
mation to the unreduced part of sub( a )
if m >= n, sub( a ) is reduced to upper bidiagonal form; if m < n, to

PZLACONSB()

(desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PZLACP2()

pzlacp2 copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PZLACPY()

pzlacpy copies all or part of a distributed matrix a to anothe
performs a local copy sub( a ) := sub( b ), where sub( a ) denotes

PZLAHQR()

more clever.  we break each transformation down into 3
parts
a group of rotn transformations (this is on

PZLAHRD()

v which is needed, with t and y, to apply the transformation to the
unreduced part of the matrix, using an update of the form

PZLANHE()

icurcol : process column containing diagonal block
irsc0   : pointer to part of work used to store the rowsums whil
irsr0   : pointer to part of work used to store the rowsums after

PZLANSY()

icurcol : process column containing diagonal block
irsc0   : pointer to part of work used to store the rowsums whil
irsr0   : pointer to part of work used to store the rowsums after

PZLAQSY()

uplo    (global input) character
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZLARZB()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PZLASE2()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PZLASET()

uplo    (global input) character
specifies the part of the distributed matrix sub( a ) to b
= 'u':      upper triangular part is set; the strictly lower

PZLASMSUB()

a       (global input) complex*16 array, dimension (desca(lld_),*)
on entry, the hessenberg matrix whose tridiagonal part i
unchanged on exit.

PZLATRD()

q' * sub( a ) * q, and returns the matrices v and w which are
needed to apply the transformation to the unreduced part of sub( a )
if uplo = 'u', pzlatrd reduces the last nb rows and columns of a

PZLATRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors. l > 0
a       (local input/local output) complex*16 pointer into the

PZLAUU2()

pzlauu2 computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PZLAUUM()

pzlauum computes the product u * u' or l' * l, where the triangular
factor u or l is stored in the upper or lower triangular part o

PZMAX1()

if incx = m_x, then sub( x ) is a vector distributed over a process
row. each process part of this row receives the result
if incx = 1, then sub( x ) is a vector distributed over a process

PZPBTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZPBTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZPORFS()

uplo    (global input) character*1
specifies whether the upper or lower triangular part of th
= 'u':  upper triangular

PZPOSV()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PZPOSVX()

diag(sr)*a*diag(sc).  if uplo = 'u', the leading
n-by-n upper triangular part of a contains the uppe
triangular part of a is not referenced.  if uplo = 'l', the

PZPOTF2()

n-by-n symmetric distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PZPOTRF()

n-by-n hermitian distributed matrix sub( a ) to be factored.
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PZPTSV()

d       (local input/local output) complex*16 pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PZPTTRF()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZPTTRS()

d       (local input/local output) complex*16 pointer to local
part of global vector storing the main diagonal of th
on exit, this array contains information containing the

PZPTTRSV()

adjust addressing into matrix space to properly get into
the beginning part of the relevant dat

PZTRCON()

matrix a(ia:ia+n-1,ja:ja+n-1). if uplo = 'u', the leading
n-by-n upper triangular part of this distributed matrix con
triangular part is not referenced.  if uplo = 'l', the

PZTRRFS()

distributed matrix sub( a ).
if uplo = 'u', the leading n-by-n upper triangular part o
and its strictly lower triangular part is not referenced.

PZTRTI2()

sub( a ). if uplo = 'u', the leading n-by-n upper triangular
part of the matrix sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PZTRTRI()

triangular matrix sub( a ).  if uplo = 'u', the leading
n-by-n upper triangular part of the matrix sub( a ) contain
lower triangular part of sub( a ) is not referenced.

PZTRTRS()

matrix sub( a ).  if uplo = 'u', the leading n-by-n upper
triangular part of sub( a ) contains the upper triangula
is not referenced.  if uplo = 'l', the leading n-by-n lower

PZTZRZF()

sub( a ) which is to be factored. on exit, the leading m-by-m
upper triangular part of sub( a ) contains the upper trian
sub( a ), with the array tau, represent the unitary matrix z

PZUNMR3()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

PZUNMRZ()

the columns of the distributed submatrix sub( a ) containing
the meaningful part of the householder reflectors

SDBTRF()

the active part of the matrix is partitione
a11   a12   a13

STRMVT()

before entry with  uplo = 'u' or 'u', the leading n by n
upper triangular part of the array t must contain the uppe
t is not referenced.

ZCOMBAMAX1 ()

zcombamax1 finds the element having maximum real part absolut

ZDBTRF()

the active part of the matrix is partitione
a11   a12   a13

ZTRMVT()

before entry with  uplo = 'u' or 'u', the leading n by n
upper triangular part of the array t must contain the uppe
t is not referenced.

partial

CDBTF2()

cdbtrf computes an lu factorization of a real m-by-n band matrix a
without using partial pivoting with row interchanges
this is the unblocked version of the algorithm, calling level 2 blas.

CDTTRF()

cdttrf computes an lu factorization of a complex tridiagonal matrix a
using elimination without partial pivoting
the factorization has the form

DDBTF2()

ddbtrf computes an lu factorization of a real m-by-n band matrix a
without using partial pivoting with row interchanges
this is the unblocked version of the algorithm, calling level 2 blas.

DDTTRF()

ddttrf computes an lu factorization of a complex tridiagonal matrix a
using elimination without partial pivoting
the factorization has the form

DSTEIN2()

compute lu factors with partial pivoting  ( pt = lu 

PCGBTRF()

lu factorization with partial pivotin

PCGESV()

the lu decomposition with partial pivoting and row interchanges i
tation matrix, l is unit lower triangular, and u is upper triangular.

PCGETF2()

distributed matrix sub( a ) = a(ia:ia+m-1,ja:ja+n-1) using
partial pivoting with row interchanges
the factorization has the form sub( a ) = p * l * u, where p is a

PCGETRF()

pcgetrf computes an lu factorization of a general m-by-n distributed
matrix sub( a ) = (ia:ia+m-1,ja:ja+n-1) using partial pivoting wit

PDGBTRF()

lu factorization with partial pivotin

PDGESV()

the lu decomposition with partial pivoting and row interchanges i
tation matrix, l is unit lower triangular, and u is upper triangular.

PDGETF2()

distributed matrix sub( a ) = a(ia:ia+m-1,ja:ja+n-1) using
partial pivoting with row interchanges
the factorization has the form sub( a ) = p * l * u, where p is a

PDGETRF()

pdgetrf computes an lu factorization of a general m-by-n distributed
matrix sub( a ) = (ia:ia+m-1,ja:ja+n-1) using partial pivoting wit

PSGBTRF()

lu factorization with partial pivotin

PSGESV()

the lu decomposition with partial pivoting and row interchanges i
tation matrix, l is unit lower triangular, and u is upper triangular.

PSGETF2()

distributed matrix sub( a ) = a(ia:ia+m-1,ja:ja+n-1) using
partial pivoting with row interchanges
the factorization has the form sub( a ) = p * l * u, where p is a

PSGETRF()

psgetrf computes an lu factorization of a general m-by-n distributed
matrix sub( a ) = (ia:ia+m-1,ja:ja+n-1) using partial pivoting wit

PZGBTRF()

lu factorization with partial pivotin

PZGESV()

the lu decomposition with partial pivoting and row interchanges i
tation matrix, l is unit lower triangular, and u is upper triangular.

PZGETF2()

distributed matrix sub( a ) = a(ia:ia+m-1,ja:ja+n-1) using
partial pivoting with row interchanges
the factorization has the form sub( a ) = p * l * u, where p is a

PZGETRF()

pzgetrf computes an lu factorization of a general m-by-n distributed
matrix sub( a ) = (ia:ia+m-1,ja:ja+n-1) using partial pivoting wit

SDBTF2()

sdbtrf computes an lu factorization of a real m-by-n band matrix a
without using partial pivoting with row interchanges
this is the unblocked version of the algorithm, calling level 2 blas.

SDTTRF()

sdttrf computes an lu factorization of a complex tridiagonal matrix a
using elimination without partial pivoting
the factorization has the form

SSTEIN2()

compute lu factors with partial pivoting  ( pt = lu 

ZDBTF2()

zdbtrf computes an lu factorization of a real m-by-n band matrix a
without using partial pivoting with row interchanges
this is the unblocked version of the algorithm, calling level 2 blas.

ZDTTRF()

zdttrf computes an lu factorization of a complex tridiagonal matrix a
using elimination without partial pivoting
the factorization has the form

participate

PCDBTRF()

the last processor does not participate in the factorization o

PCDBTRSV()

the last processor does not participate in the solution of th

PCDTTRF()

the last processor does not participate in the factorization o

PCDTTRSV()

the last processor does not participate in the solution of th

PCPBTRF()

the last processor does not participate in the factorization o

PCPBTRSV()

the last processor does not participate in the solution of th

PCPTTRF()

the last processor does not participate in the factorization o

PCPTTRSV()

the last processor does not participate in the solution of th

PDDBTRF()

the last processor does not participate in the factorization o

PDDBTRSV()

the last processor does not participate in the solution of th

PDDTTRF()

the last processor does not participate in the factorization o

PDDTTRSV()

the last processor does not participate in the solution of th

PDPBTRF()

the last processor does not participate in the factorization o

PDPBTRSV()

the last processor does not participate in the solution of th

PDPTTRF()

the last processor does not participate in the factorization o

PDPTTRSV()

the last processor does not participate in the solution of th

PSDBTRF()

the last processor does not participate in the factorization o

PSDBTRSV()

the last processor does not participate in the solution of th

PSDTTRF()

the last processor does not participate in the factorization o

PSDTTRSV()

the last processor does not participate in the solution of th

PSPBTRF()

the last processor does not participate in the factorization o

PSPBTRSV()

the last processor does not participate in the solution of th

PSPTTRF()

the last processor does not participate in the factorization o

PSPTTRSV()

the last processor does not participate in the solution of th

PZDBTRF()

the last processor does not participate in the factorization o

PZDBTRSV()

the last processor does not participate in the solution of th

PZDTTRF()

the last processor does not participate in the factorization o

PZDTTRSV()

the last processor does not participate in the solution of th

PZPBTRF()

the last processor does not participate in the factorization o

PZPBTRSV()

the last processor does not participate in the solution of th

PZPTTRF()

the last processor does not participate in the factorization o

PZPTTRSV()

the last processor does not participate in the solution of th

particular

PCGESVX()

a(ia:ia+n-1,ja:ja+n-1) after equilibration (if done).  if
rcond is less than the machine precision (in particular, i
this condition is indicated by a return code of info > 0.

PCGGQRF()

in particular, if sub( b ) is square and nonsingular, the gq
factorization of inv( sub( b ) )* sub( a ):

PCGGRQF()

in particular, if sub( b ) is square and nonsingular, the gr
factorization of sub( a )*inv( sub( b ) ):

PCPOSVX()

a after equilibration (if done).  if rcond is less than the
machine precision (in particular, if rcond = 0), the matri
indicated by a return code of info > 0, and the solution and

PCSRSCL()

because vectors may be seen as particular matrices, a distribute

PDGESVX()

a(ia:ia+n-1,ja:ja+n-1) after equilibration (if done).  if
rcond is less than the machine precision (in particular, i
this condition is indicated by a return code of info > 0.

PDGGQRF()

in particular, if sub( b ) is square and nonsingular, the gq
factorization of inv( sub( b ) )* sub( a ):

PDGGRQF()

in particular, if sub( b ) is square and nonsingular, the gr
factorization of sub( a )*inv( sub( b ) ):

PDPOSVX()

a after equilibration (if done).  if rcond is less than the
machine precision (in particular, if rcond = 0), the matri
indicated by a return code of info > 0, and the solution and

PDRSCL()

because vectors may be seen as particular matrices, a distribute

PJLAENV()

computers.  users are encouraged to modify this subroutine to set
the tuning parameters for their particular machine using the optio

PSGESVX()

a(ia:ia+n-1,ja:ja+n-1) after equilibration (if done).  if
rcond is less than the machine precision (in particular, i
this condition is indicated by a return code of info > 0.

PSGGQRF()

in particular, if sub( b ) is square and nonsingular, the gq
factorization of inv( sub( b ) )* sub( a ):

PSGGRQF()

in particular, if sub( b ) is square and nonsingular, the gr
factorization of sub( a )*inv( sub( b ) ):

PSPOSVX()

a after equilibration (if done).  if rcond is less than the
machine precision (in particular, if rcond = 0), the matri
indicated by a return code of info > 0, and the solution and

PSRSCL()

because vectors may be seen as particular matrices, a distribute

PZDRSCL()

because vectors may be seen as particular matrices, a distribute

PZGESVX()

a(ia:ia+n-1,ja:ja+n-1) after equilibration (if done).  if
rcond is less than the machine precision (in particular, i
this condition is indicated by a return code of info > 0.

PZGGQRF()

in particular, if sub( b ) is square and nonsingular, the gq
factorization of inv( sub( b ) )* sub( a ):

PZGGRQF()

in particular, if sub( b ) is square and nonsingular, the gr
factorization of sub( a )*inv( sub( b ) ):

PZPOSVX()

a after equilibration (if done).  if rcond is less than the
machine precision (in particular, if rcond = 0), the matri
indicated by a return code of info > 0, and the solution and

particularly

PCHENTRD()

pchentrd is faster than pchetrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PDSYNTRD()

pdsyntrd is faster than pdsytrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PSSYNTRD()

pssyntrd is faster than pssytrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

PZHENTRD()

pzhentrd is faster than pzhetrd on almost all matrices,
particularly small ones (i.e. n < 500 * sqrt(p) ), provided tha

partition

DLAPST()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

DLASRT2()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

PCDBTRF()

user-input value of partition siz

PCDBTRSV()

user-input value of partition siz

PCDTTRF()

user-input value of partition siz

PCDTTRSV()

user-input value of partition siz

PCGBTRF()

user-input value of partition siz

PCPBTRF()

user-input value of partition siz

PCPBTRSV()

user-input value of partition siz

PCPTTRF()

user-input value of partition siz

PCPTTRSV()

user-input value of partition siz

PDDBTRF()

user-input value of partition siz

PDDBTRSV()

user-input value of partition siz

PDDTTRF()

user-input value of partition siz

PDDTTRSV()

user-input value of partition siz

PDGBTRF()

user-input value of partition siz

PDPBTRF()

user-input value of partition siz

PDPBTRSV()

user-input value of partition siz

PDPTTRF()

user-input value of partition siz

PDPTTRSV()

user-input value of partition siz

PSDBTRF()

user-input value of partition siz

PSDBTRSV()

user-input value of partition siz

PSDTTRF()

user-input value of partition siz

PSDTTRSV()

user-input value of partition siz

PSGBTRF()

user-input value of partition siz

PSPBTRF()

user-input value of partition siz

PSPBTRSV()

user-input value of partition siz

PSPTTRF()

user-input value of partition siz

PSPTTRSV()

user-input value of partition siz

PZDBTRF()

user-input value of partition siz

PZDBTRSV()

user-input value of partition siz

PZDTTRF()

user-input value of partition siz

PZDTTRSV()

user-input value of partition siz

PZGBTRF()

user-input value of partition siz

PZPBTRF()

user-input value of partition siz

PZPBTRSV()

user-input value of partition siz

PZPTTRF()

user-input value of partition siz

PZPTTRSV()

user-input value of partition siz

SLAPST()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

SLASRT2()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

partitioned

CDBTRF()

the active part of the matrix is partitioned
a11   a12   a13

DDBTRF()

the active part of the matrix is partitioned
a11   a12   a13

SDBTRF()

the active part of the matrix is partitioned
a11   a12   a13

ZDBTRF()

the active part of the matrix is partitioned
a11   a12   a13

partitioning

CDBTRF()

which is about to be factorized. the number of rows in the
partitioning are jb, i2, i3 respectively, and the number
and the subdiagonal elements of a31 lie outside the band.

DDBTRF()

which is about to be factorized. the number of rows in the
partitioning are jb, i2, i3 respectively, and the number
and the subdiagonal elements of a31 lie outside the band.

PCDBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PCDBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PCDTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PCDTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PCGBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PCGBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PCPBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PCPBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PCPTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PCPTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PDDBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PDDBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PDDTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PDDTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PDGBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PDGBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PDPBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PDPBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PDPTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PDPTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PDSTEBZ()

the interval [vl, vu], or the eigenvalues indexed il through iu. a
static partitioning of work is done at the beginning of pdstebz whic
eigenvalues.

PSDBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PSDBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PSDTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PSDTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PSGBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PSGBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PSPBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PSPBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PSPTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PSPTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PSSTEBZ()

the interval [vl, vu], or the eigenvalues indexed il through iu. a
static partitioning of work is done at the beginning of psstebz whic
eigenvalues.

PZDBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PZDBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PZDTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PZDTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PZGBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PZGBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PZPBSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PZPBTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithm description: divide and conquer

PZPTSV()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

PZPTTRS()

implemented. these go by many names, including divide and conquer,
partitioning, domain decomposition-type, etc
algorithms are the appropriate choice.

SDBTRF()

which is about to be factorized. the number of rows in the
partitioning are jb, i2, i3 respectively, and the number
and the subdiagonal elements of a31 lie outside the band.

ZDBTRF()

which is about to be factorized. the number of rows in the
partitioning are jb, i2, i3 respectively, and the number
and the subdiagonal elements of a31 lie outside the band.

parts

DLAPST()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

DLASRT2()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

PCDBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCDBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCDTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCDTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCGBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCGBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCLAHQR()

more clever.  we break each transformation down into 3
parts
a group of rotn transformations (this is on

PCPBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCPBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCPTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PCPTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDDBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDDBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDDTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDDTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDGBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDGBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDLAEDZ()

proc (iqrow, iqcol) receive the parts of z

PDLAHQR()

more clever.  we break each transformation down into 3
parts
a group of rotn transformations (this is on

PDPBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDPBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDPTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PDPTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSDBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSDBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSDTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSDTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSGBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSGBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSLAEDZ()

proc (iqrow, iqcol) receive the parts of z

PSLAHQR()

more clever.  we break each transformation down into 3
parts
a group of rotn transformations (this is on

PSPBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSPBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSPTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PSPTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZDBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZDBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZDTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZDTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZGBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZGBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZLAHQR()

more clever.  we break each transformation down into 3
parts
a group of rotn transformations (this is on

PZPBSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZPBTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZPTSV()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

PZPTTRS()

the matrices factored on each processor. the factors of
these submatrices overwrite the corresponding parts of 
2) reduced system phase:

SLAPST()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

SLASRT2()

partition d( start:endd ) and stack parts, largest one firs
choose partition entry as median of 3

pass

PCLAPIV()

for example if the row pivots should be applied to the columns of
sub( a ), pass rowcol='c' and pivroc='c'
notes

PCLASSQ()

the routine makes only one pass through the vector sub( x )
notes

PDLAPIV()

for example if the row pivots should be applied to the columns of
sub( a ), pass rowcol='c' and pivroc='c'
notes

PDLASSQ()

the routine makes only one pass through the vector sub( x )
notes

PSLAPIV()

for example if the row pivots should be applied to the columns of
sub( a ), pass rowcol='c' and pivroc='c'
notes

PSLASSQ()

the routine makes only one pass through the vector sub( x )
notes

PZLAPIV()

for example if the row pivots should be applied to the columns of
sub( a ), pass rowcol='c' and pivroc='c'
notes

PZLASSQ()

the routine makes only one pass through the vector sub( x )
notes

passed

PCDBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCDBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCDTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCDTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCGBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCGBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCLAHQR()

work is done on that.  at the end of the border, the data is
passed back and everything stays a lot simpler

PCPBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCPBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCPTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PCPTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDDBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDDBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDDTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDDTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDGBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDGBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDLAED2()

the first k values of the final deflation-altered z-vector
which will be passed to slaed3
q2     (output) double precision array, dimension (ldq2, nq)

PDLAED3()

the first k values of the final deflation-altered z-vector
which will be passed to slaed3
z      (global input) double precision array, dimension (n)

PDLAHQR()

work is done on that.  at the end of the border, the data is
passed back and everything stays a lot simpler

PDPBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDPBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDPTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PDPTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PJLAENV()

first, n2 second, and so on, and unused problem dimensions are
passed a value of -1
in the calling subroutine.  for example, pjlaenv is used to

PSDBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSDBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSDTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSDTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSGBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSGBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSLAED2()

the first k values of the final deflation-altered z-vector
which will be passed to slaed3
q2     (output) real array, dimension (ldq2, nq)

PSLAED3()

the first k values of the final deflation-altered z-vector
which will be passed to slaed3
z      (global input) real array, dimension (n)

PSLAHQR()

work is done on that.  at the end of the border, the data is
passed back and everything stays a lot simpler

PSPBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSPBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSPTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PSPTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZDBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZDBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZDTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZDTTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZGBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZGBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZLAHQR()

work is done on that.  at the end of the border, the data is
passed back and everything stays a lot simpler

PZPBSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZPBTRS()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZPTSV()

scalapack routines *do not support intercontext operations* so that
the grid passed to a single scalapack routine *must be the same

PZPTTRS()