Support block matrices in several BLAS1 routines #226
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
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
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
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
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
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 +1,135 @@ | ||
| /* | ||
| Copyright (c) 2009-2016, Jack Poulson | ||
| All rights reserved. | ||
|
|
||
| This file is part of Elemental and is under the BSD 2-Clause License, | ||
| which can be found in the LICENSE file in the root directory, or at | ||
| http://opensource.org/licenses/BSD-2-Clause | ||
| */ | ||
|
|
||
| #include <El.hpp> | ||
| using namespace El; | ||
|
|
||
| template <typename T, DistWrap W> | ||
| void TestColumnTwoNorms(Int m, Int n, const Grid& g, bool print) | ||
| { | ||
| // Generate random matrix to test. | ||
| DistMatrix<T, MC, MR, W> A(g); | ||
| Uniform(A, m, n); | ||
| if (print) | ||
| Print(A, "A"); | ||
| DistMatrix<T, MR, STAR, W> norms(g); | ||
| ColumnTwoNorms(A, norms); | ||
| if (print) | ||
| Print(norms, "norms"); | ||
| for (Int j = 0; j < A.LocalWidth(); ++j) | ||
| { | ||
| T got = norms.GetLocal(j, 0); | ||
| T expected = 0; | ||
| for (Int i = 0; i < A.LocalHeight(); ++i) | ||
| { | ||
| T val = A.GetLocal(i, j); | ||
| expected += val * val; | ||
|
There was a problem hiding this comment. Explicitly accumulating the squares and then square-rooting should be fine for random data but can lead to very low accuracy in extreme cases. There is the routine There was a problem hiding this comment. My main goal there was to provide a simple baseline that ensures there's no major issues (e.g. block vs element distributions causing problems) while not just reimplementing the method that's being tested. If you think it would be better to switch to |
||
| } | ||
| expected = mpi::AllReduce(expected, g.ColComm()); | ||
| expected = Sqrt(expected); | ||
| if (Abs(got - expected) > 10 * limits::Epsilon<El::Base<T>>()) | ||
|
There was a problem hiding this comment. The accumulated error should be a function of the number of entries (with said function being logarithmic if a tree-based scheme is used, but linear with the current single-process implementation): would you mind multiplying the right-hand side by the number of entries? Also, it would be better to use a relative bound and to diving the left-hand side by the maximum of the norm and 1 (so that it still behaves well for zero norms). There was a problem hiding this comment. Just want to clarify: The new check should be |
||
| { | ||
| Output("Results do not match, norms(", j, ")=", got, | ||
| " instead of ", expected); | ||
| RuntimeError("got != expected"); | ||
| } | ||
| } | ||
| } | ||
|
|
||
| template <typename T, DistWrap W> | ||
| void TestColumnMaxNorms(Int m, Int n, const Grid& g, bool print) | ||
| { | ||
| // Generate random matrix to test. | ||
| DistMatrix<T, MC, MR, W> A(g); | ||
| Uniform(A, m, n); | ||
| if (print) | ||
| Print(A, "A"); | ||
| DistMatrix<T, MR, STAR, W> norms(g); | ||
| ColumnMaxNorms(A, norms); | ||
| if (print) | ||
| Print(norms, "norms"); | ||
| for (Int j = 0; j < A.LocalWidth(); ++j) | ||
| { | ||
| T got = norms.GetLocal(j, 0); | ||
| T expected = 0; | ||
| for (Int i = 0; i < A.LocalHeight(); ++i) | ||
| expected = Max(expected, Abs(A.GetLocal(i, j))); | ||
| T r; | ||
| mpi::AllReduce(&expected, &r, 1, mpi::MAX, g.ColComm()); | ||
| expected = r; | ||
| if (got != expected) | ||
| { | ||
| Output("Results do not match, norms(", j, ")=", got, | ||
| " instead of ", expected); | ||
| RuntimeError("got != expected"); | ||
| } | ||
| } | ||
| } | ||
|
|
||
| int main(int argc, char** argv) | ||
| { | ||
| Environment env(argc, argv); | ||
| mpi::Comm comm = mpi::COMM_WORLD; | ||
| try | ||
| { | ||
| const Int m = Input("--m", "height", 100); | ||
| const Int n = Input("--n", "width", 100); | ||
| const bool print = Input("--print", "print matrices?", false); | ||
| ProcessInput(); | ||
| PrintInputReport(); | ||
|
|
||
| const Grid g(comm); | ||
| OutputFromRoot(comm, "Testing ColumnTwoNorms"); | ||
| TestColumnTwoNorms<float, ELEMENT>(m, n, g, print); | ||
| TestColumnTwoNorms<float, BLOCK>(m, n, g, print); | ||
| TestColumnTwoNorms<double, ELEMENT>(m, n, g, print); | ||
| TestColumnTwoNorms<double, BLOCK>(m, n, g, print); | ||
| #if defined(EL_HAVE_QD) && defined(EL_ENABLE_DOUBLEDOUBLE) | ||
| TestColumnTwoNorms<DoubleDouble, ELEMENT>(m, n, g, print); | ||
| TestColumnTwoNorms<DoubleDouble, BLOCK>(m, n, g, print); | ||
| #endif | ||
| #if defined(EL_HAVE_QD) && defined(EL_ENABLE_QUADDOUBLE) | ||
| TestColumnTwoNorms<QuadDouble, ELEMENT>(m, n, g, print); | ||
| TestColumnTwoNorms<QuadDouble, BLOCK>(m, n, g, print); | ||
| #endif | ||
| #if defined(EL_HAVE_QUAD) && defined(EL_ENABLE_QUAD) | ||
| TestColumnTwoNorms<Quad, ELEMENT>(m, n, g, print); | ||
| TestColumnTwoNorms<Quad, BLOCK>(m, n, g, print); | ||
| #endif | ||
| #if defined(EL_HAVE_MPC) && defined(EL_ENABLE_BIGFLOAT) | ||
| TestColumnTwoNorms<BigFloat, ELEMENT>(m, n, g, print); | ||
| TestColumnTwoNorms<BigFloat, BLOCK>(m, n, g, print); | ||
| #endif | ||
| OutputFromRoot(comm, "Testing ColumnMaxNorms"); | ||
| TestColumnMaxNorms<float, ELEMENT>(m, n, g, print); | ||
| TestColumnMaxNorms<float, BLOCK>(m, n, g, print); | ||
| TestColumnMaxNorms<double, ELEMENT>(m, n, g, print); | ||
| TestColumnMaxNorms<double, BLOCK>(m, n, g, print); | ||
| #if defined(EL_HAVE_QD) && defined(EL_ENABLE_DOUBLEDOUBLE) | ||
| TestColumnMaxNorms<DoubleDouble, ELEMENT>(m, n, g, print); | ||
| TestColumnMaxNorms<DoubleDouble, BLOCK>(m, n, g, print); | ||
| #endif | ||
| #if defined(EL_HAVE_QD) && defined(EL_ENABLE_QUADDOUBLE) | ||
| TestColumnMaxNorms<QuadDouble, ELEMENT>(m, n, g, print); | ||
| TestColumnMaxNorms<QuadDouble, BLOCK>(m, n, g, print); | ||
| #endif | ||
| #if defined(EL_HAVE_QUAD) && defined(EL_ENABLE_QUAD) | ||
| TestColumnMaxNorms<Quad, ELEMENT>(m, n, g, print); | ||
| TestColumnMaxNorms<Quad, BLOCK>(m, n, g, print); | ||
| #endif | ||
| #if defined(EL_HAVE_MPC) && defined(EL_ENABLE_BIGFLOAT) | ||
| TestColumnMaxNorms<BigFloat, ELEMENT>(m, n, g, print); | ||
| TestColumnMaxNorms<BigFloat, BLOCK>(m, n, g, print); | ||
| #endif | ||
| } | ||
| catch (exception& e) | ||
| { | ||
| ReportException(e); | ||
| } | ||
| } | ||
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you noticing this crucial portion of the extension!