Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
44a6721
commit e5ff14e
Showing
2 changed files
with
140 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,61 @@ | ||
// Copyright ©2018 The Gonum Authors. All rights reserved. | ||
// Use of this source code is governed by a BSD-style | ||
// license that can be found in the LICENSE file. | ||
|
||
package gonum | ||
|
||
import ( | ||
"gonum.org/v1/gonum/blas" | ||
"gonum.org/v1/gonum/blas/blas64" | ||
) | ||
|
||
// Dlauu2 computes the product | ||
// U * U^T if uplo is blas.Upper | ||
// L^T * L if uplo is blas.Lower | ||
// where U or L is stored in the upper or lower triangular part of A. | ||
// Only the upper or lower triangle of the result is stored, overwriting | ||
// the corresponding factor in A. | ||
func (impl Implementation) Dlauu2(uplo blas.Uplo, n int, a []float64, lda int) { | ||
switch { | ||
case uplo != blas.Upper && uplo != blas.Lower: | ||
panic(badUplo) | ||
case n < 0: | ||
panic(nLT0) | ||
case lda < max(1, n): | ||
panic(badLdA) | ||
case len(a) < (n-1)*lda+n: | ||
panic("lapack: a has insufficient length") | ||
} | ||
|
||
// Quick return if possible. | ||
if n == 0 { | ||
return | ||
} | ||
|
||
bi := blas64.Implementation() | ||
if uplo == blas.Upper { | ||
// Compute the product U*U^T. | ||
for i := 0; i < n; i++ { | ||
aii := a[i*lda+i] | ||
if i < n-1 { | ||
a[i*lda+i] = bi.Ddot(n-i, a[i*lda+i:], 1, a[i*lda+i:], 1) | ||
bi.Dgemv(blas.NoTrans, i, n-i-1, 1, a[i+1:], lda, a[i*lda+i+1:], 1, | ||
aii, a[i:], lda) | ||
} else { | ||
bi.Dscal(i+1, aii, a[i:], lda) | ||
} | ||
} | ||
} else { | ||
// Compute the product L^T*L. | ||
for i := 0; i < n; i++ { | ||
aii := a[i*lda+i] | ||
if i < n-1 { | ||
a[i*lda+i] = bi.Ddot(n-i, a[i*lda+i:], lda, a[i*lda+i:], lda) | ||
bi.Dgemv(blas.Trans, n-i-1, i, 1, a[(i+1)*lda:], lda, a[(i+1)*lda+i:], lda, | ||
aii, a[i*lda:], 1) | ||
} else { | ||
bi.Dscal(i+1, aii, a[i*lda:], 1) | ||
} | ||
} | ||
} | ||
} |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,79 @@ | ||
// Copyright ©2018 The Gonum Authors. All rights reserved. | ||
// Use of this source code is governed by a BSD-style | ||
// license that can be found in the LICENSE file. | ||
|
||
package gonum | ||
|
||
import ( | ||
"gonum.org/v1/gonum/blas" | ||
"gonum.org/v1/gonum/blas/blas64" | ||
) | ||
|
||
// Dlauum computes the product | ||
// U * U^T if uplo is blas.Upper | ||
// L^T * L if uplo is blas.Lower | ||
// where U or L is stored in the upper or lower triangular part of A. | ||
// Only the upper or lower triangle of the result is stored, overwriting | ||
// the corresponding factor in A. | ||
func (impl Implementation) Dlauum(uplo blas.Uplo, n int, a []float64, lda int) { | ||
switch { | ||
case uplo != blas.Upper && uplo != blas.Lower: | ||
panic(badUplo) | ||
case n < 0: | ||
panic(nLT0) | ||
case lda < max(1, n): | ||
panic(badLdA) | ||
case len(a) < (n-1)*lda+n: | ||
panic("lapack: a has insufficient length") | ||
} | ||
|
||
// Quick return if possible. | ||
if n == 0 { | ||
return | ||
} | ||
|
||
// Determine the block size. | ||
opts := "U" | ||
if uplo == blas.Lower { | ||
opts = "L" | ||
} | ||
nb := impl.Ilaenv(1, "DLAUUM", opts, n, -1, -1, -1) | ||
|
||
if nb <= 1 || n <= nb { | ||
// Use unblocked code. | ||
impl.Dlauu2(uplo, n, a, lda) | ||
return | ||
} | ||
|
||
// Use blocked code. | ||
bi := blas64.Implementation() | ||
if uplo == blas.Upper { | ||
// Compute the product U*U^T. | ||
for i := 0; i < n; i += nb { | ||
ib := min(nb, n-i) | ||
bi.Dtrmm(blas.Right, blas.Upper, blas.Trans, blas.NonUnit, | ||
i, ib, 1, a[i*lda+i:], lda, a[i:], lda) | ||
impl.Dlauu2(blas.Upper, ib, a[i*lda+i:], lda) | ||
if n-i-ib > 0 { | ||
bi.Dgemm(blas.NoTrans, blas.Trans, i, ib, n-i-ib, | ||
1, a[i+ib:], lda, a[i*lda+i+ib:], lda, 1, a[i:], lda) | ||
bi.Dsyrk(blas.Upper, blas.NoTrans, ib, n-i-ib, | ||
1, a[i*lda+i+ib:], lda, 1, a[i*lda+i:], lda) | ||
} | ||
} | ||
} else { | ||
// Compute the product L^T*L. | ||
for i := 0; i < n; i += nb { | ||
ib := min(nb, n-i) | ||
bi.Dtrmm(blas.Left, blas.Lower, blas.Trans, blas.NonUnit, | ||
ib, i, 1, a[i*lda+i:], lda, a[i*lda:], lda) | ||
impl.Dlauu2(blas.Lower, ib, a[i*lda+i:], lda) | ||
if n-i-ib > 0 { | ||
bi.Dgemm(blas.Trans, blas.NoTrans, ib, i, n-i-ib, | ||
1, a[(i+ib)*lda+i:], lda, a[(i+ib)*lda:], lda, 1, a[i*lda:], lda) | ||
bi.Dsyrk(blas.Lower, blas.Trans, ib, n-i-ib, | ||
1, a[(i+ib)*lda+i:], lda, 1, a[i*lda+i:], lda) | ||
} | ||
} | ||
} | ||
} |