Skip to content

Improve variants of Cholesky and QR #847

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Jun 3, 2023
Merged

Improve variants of Cholesky and QR #847

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
de83ba6
Fix typos in docs of variants of Cholesky and QR
angsch Jun 1, 2023
65df15e
Remove SCEIL, replace calls with intrinsic CEILING
angsch Jun 2, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion SRC/VARIANTS/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@ LULL = lu/LL/cgetrf.o lu/LL/dgetrf.o lu/LL/sgetrf.o lu/LL/zgetrf.o

LUREC = lu/REC/cgetrf.o lu/REC/dgetrf.o lu/REC/sgetrf.o lu/REC/zgetrf.o

QRLL = qr/LL/cgeqrf.o qr/LL/dgeqrf.o qr/LL/sgeqrf.o qr/LL/zgeqrf.o qr/LL/sceil.o
QRLL = qr/LL/cgeqrf.o qr/LL/dgeqrf.o qr/LL/sgeqrf.o qr/LL/zgeqrf.o


.PHONY: all
Expand Down
2 changes: 1 addition & 1 deletion SRC/VARIANTS/cholesky/RL/cpotrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@
C>\details \b Purpose:
C>\verbatim
C>
C> CPOTRF computes the Cholesky factorization of a real Hermitian
C> CPOTRF computes the Cholesky factorization of a complex Hermitian
C> positive definite matrix A.
C>
C> The factorization has the form
Expand Down
2 changes: 1 addition & 1 deletion SRC/VARIANTS/cholesky/RL/zpotrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@
C>\details \b Purpose:
C>\verbatim
C>
C> ZPOTRF computes the Cholesky factorization of a real Hermitian
C> ZPOTRF computes the Cholesky factorization of a complex Hermitian
C> positive definite matrix A.
C>
C> The factorization has the form
Expand Down
4 changes: 2 additions & 2 deletions SRC/VARIANTS/cholesky/TOP/cpotrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@
C>\details \b Purpose:
C>\verbatim
C>
C> CPOTRF computes the Cholesky factorization of a real symmetric
C> CPOTRF computes the Cholesky factorization of a complex Hermitian
C> positive definite matrix A.
C>
C> The factorization has the form
Expand Down Expand Up @@ -55,7 +55,7 @@
C> \param[in,out] A
C> \verbatim
C> A is COMPLEX array, dimension (LDA,N)
C> On entry, the symmetric matrix A. If UPLO = 'U', the leading
C> On entry, the Hermitian matrix A. If UPLO = 'U', the leading
C> N-by-N upper triangular part of A contains the upper
C> triangular part of the matrix A, and the strictly lower
C> triangular part of A is not referenced. If UPLO = 'L', the
Expand Down
4 changes: 2 additions & 2 deletions SRC/VARIANTS/cholesky/TOP/zpotrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@
C>\details \b Purpose:
C>\verbatim
C>
C> ZPOTRF computes the Cholesky factorization of a real symmetric
C> ZPOTRF computes the Cholesky factorization of a complex Hermitian
C> positive definite matrix A.
C>
C> The factorization has the form
Expand Down Expand Up @@ -55,7 +55,7 @@
C> \param[in,out] A
C> \verbatim
C> A is COMPLEX*16 array, dimension (LDA,N)
C> On entry, the symmetric matrix A. If UPLO = 'U', the leading
C> On entry, the Hermitian matrix A. If UPLO = 'U', the leading
C> N-by-N upper triangular part of A contains the upper
C> triangular part of the matrix A, and the strictly lower
C> triangular part of A is not referenced. If UPLO = 'L', the
Expand Down
13 changes: 6 additions & 7 deletions SRC/VARIANTS/qr/LL/cgeqrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@
C>\details \b Purpose:
C>\verbatim
C>
C> CGEQRF computes a QR factorization of a real M-by-N matrix A:
C> CGEQRF computes a QR factorization of a complex M-by-N matrix A:
C> A = Q * R.
C>
C> This is the left-looking Level 3 BLAS version of the algorithm.
Expand Down Expand Up @@ -172,12 +172,11 @@ SUBROUTINE CGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
EXTERNAL CGEQR2, CLARFB, CLARFT, XERBLA
* ..
* .. Intrinsic Functions ..
INTRINSIC MAX, MIN
INTRINSIC CEILING, MAX, MIN, REAL
* ..
* .. External Functions ..
INTEGER ILAENV
REAL SCEIL
EXTERNAL ILAENV, SCEIL
EXTERNAL ILAENV
* ..
* .. Executable Statements ..

Expand Down Expand Up @@ -205,13 +204,13 @@ SUBROUTINE CGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
*
* So here 4 x 4 is the last T stored in the workspace
*
NT = K-SCEIL(REAL(K-NX)/REAL(NB))*NB
NT = K-CEILING(REAL(K-NX)/REAL(NB))*NB

*
* optimal workspace = space for dlarfb + space for normal T's + space for the last T
*
LLWORK = MAX (MAX((N-M)*K, (N-M)*NB), MAX(K*NB, NB*NB))
LLWORK = SCEIL(REAL(LLWORK)/REAL(NB))
LLWORK = CEILING(REAL(LLWORK)/REAL(NB))

IF( K.EQ.0 ) THEN

Expand All @@ -230,7 +229,7 @@ SUBROUTINE CGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )

ELSE

LBWORK = SCEIL(REAL(K)/REAL(NB))*NB
LBWORK = CEILING(REAL(K)/REAL(NB))*NB
LWKOPT = (LBWORK+LLWORK-NB)*NB
WORK( 1 ) = LWKOPT

Expand Down
11 changes: 5 additions & 6 deletions SRC/VARIANTS/qr/LL/dgeqrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,12 +172,11 @@ SUBROUTINE DGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
EXTERNAL DGEQR2, DLARFB, DLARFT, XERBLA
* ..
* .. Intrinsic Functions ..
INTRINSIC MAX, MIN
INTRINSIC CEILING, MAX, MIN, REAL
* ..
* .. External Functions ..
INTEGER ILAENV
REAL SCEIL
EXTERNAL ILAENV, SCEIL
EXTERNAL ILAENV
* ..
* .. Executable Statements ..

Expand Down Expand Up @@ -205,13 +204,13 @@ SUBROUTINE DGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
*
* So here 4 x 4 is the last T stored in the workspace
*
NT = K-SCEIL(REAL(K-NX)/REAL(NB))*NB
NT = K-CEILING(REAL(K-NX)/REAL(NB))*NB

*
* optimal workspace = space for dlarfb + space for normal T's + space for the last T
*
LLWORK = MAX (MAX((N-M)*K, (N-M)*NB), MAX(K*NB, NB*NB))
LLWORK = SCEIL(REAL(LLWORK)/REAL(NB))
LLWORK = CEILING(REAL(LLWORK)/REAL(NB))

IF( K.EQ.0 ) THEN

Expand All @@ -230,7 +229,7 @@ SUBROUTINE DGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )

ELSE

LBWORK = SCEIL(REAL(K)/REAL(NB))*NB
LBWORK = CEILING(REAL(K)/REAL(NB))*NB
LWKOPT = (LBWORK+LLWORK-NB)*NB
WORK( 1 ) = LWKOPT

Expand Down
86 changes: 0 additions & 86 deletions SRC/VARIANTS/qr/LL/sceil.f
View file
Open in desktop

This file was deleted.

11 changes: 5 additions & 6 deletions SRC/VARIANTS/qr/LL/sgeqrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,12 +172,11 @@ SUBROUTINE SGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
EXTERNAL SGEQR2, SLARFB, SLARFT, XERBLA
* ..
* .. Intrinsic Functions ..
INTRINSIC MAX, MIN
INTRINSIC CEILING, MAX, MIN, REAL
* ..
* .. External Functions ..
INTEGER ILAENV
REAL SCEIL
EXTERNAL ILAENV, SCEIL
EXTERNAL ILAENV
* ..
* .. Executable Statements ..

Expand Down Expand Up @@ -205,13 +204,13 @@ SUBROUTINE SGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
*
* So here 4 x 4 is the last T stored in the workspace
*
NT = K-SCEIL(REAL(K-NX)/REAL(NB))*NB
NT = K-CEILING(REAL(K-NX)/REAL(NB))*NB

*
* optimal workspace = space for dlarfb + space for normal T's + space for the last T
*
LLWORK = MAX (MAX((N-M)*K, (N-M)*NB), MAX(K*NB, NB*NB))
LLWORK = SCEIL(REAL(LLWORK)/REAL(NB))
LLWORK = CEILING(REAL(LLWORK)/REAL(NB))

IF( K.EQ.0 ) THEN

Expand All @@ -230,7 +229,7 @@ SUBROUTINE SGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )

ELSE

LBWORK = SCEIL(REAL(K)/REAL(NB))*NB
LBWORK = CEILING(REAL(K)/REAL(NB))*NB
LWKOPT = (LBWORK+LLWORK-NB)*NB
WORK( 1 ) = LWKOPT

Expand Down
13 changes: 6 additions & 7 deletions SRC/VARIANTS/qr/LL/zgeqrf.f
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@
C>\details \b Purpose:
C>\verbatim
C>
C> ZGEQRF computes a QR factorization of a real M-by-N matrix A:
C> ZGEQRF computes a QR factorization of a complex M-by-N matrix A:
C> A = Q * R.
C>
C> This is the left-looking Level 3 BLAS version of the algorithm.
Expand Down Expand Up @@ -172,12 +172,11 @@ SUBROUTINE ZGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
EXTERNAL ZGEQR2, ZLARFB, ZLARFT, XERBLA
* ..
* .. Intrinsic Functions ..
INTRINSIC MAX, MIN
INTRINSIC CEILING, MAX, MIN, REAL
* ..
* .. External Functions ..
INTEGER ILAENV
REAL SCEIL
EXTERNAL ILAENV, SCEIL
EXTERNAL ILAENV
* ..
* .. Executable Statements ..

Expand Down Expand Up @@ -205,13 +204,13 @@ SUBROUTINE ZGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )
*
* So here 4 x 4 is the last T stored in the workspace
*
NT = K-SCEIL(REAL(K-NX)/REAL(NB))*NB
NT = K-CEILING(REAL(K-NX)/REAL(NB))*NB

*
* optimal workspace = space for dlarfb + space for normal T's + space for the last T
*
LLWORK = MAX (MAX((N-M)*K, (N-M)*NB), MAX(K*NB, NB*NB))
LLWORK = SCEIL(REAL(LLWORK)/REAL(NB))
LLWORK = CEILING(REAL(LLWORK)/REAL(NB))

IF( K.EQ.0 ) THEN

Expand All @@ -230,7 +229,7 @@ SUBROUTINE ZGEQRF ( M, N, A, LDA, TAU, WORK, LWORK, INFO )

ELSE

LBWORK = SCEIL(REAL(K)/REAL(NB))*NB
LBWORK = CEILING(REAL(K)/REAL(NB))*NB
LWKOPT = (LBWORK+LLWORK-NB)*NB
WORK( 1 ) = LWKOPT

Expand Down