Skip to content

fix: advance indices in unrolled loops of portable dcopy and scopy #16122

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

Open
mantrapatel05 wants to merge 1 commit into
base: develop
Choose a base branch
from
Open

fix: advance indices in unrolled loops of portable dcopy and scopy #16122

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 2 additions & 0 deletions lib/node_modules/@stdlib/blas/base/dcopy/src/dcopy_ndarray.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,8 @@ void API_SUFFIX(c_dcopy_ndarray)( const CBLAS_INT N, const double *X, const CBLA
Y[ iy+4 ] = X[ ix+4 ];
Y[ iy+5 ] = X[ ix+5 ];
Y[ iy+6 ] = X[ ix+6 ];
ix += M;
iy += M;
}
return;
}
Expand Down
29 changes: 29 additions & 0 deletions lib/node_modules/@stdlib/blas/base/dcopy/test/test.ndarray.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -296,3 +296,32 @@ tape( 'if both strides are equal to `1`, the function efficiently copies element

t.end();
});

tape( 'if both strides are equal to `1` and the offsets are non-zero, the function copies all elements spanning multiple unrolled blocks', function test( t ) {
var expected;
var lens;
var x;
var y;
var i;
var j;

lens = [ 14, 15, 16, 21, 28 ];
for ( j = 0; j < lens.length; j++ ) {
x = new Float64Array( lens[ j ]+5 );
y = new Float64Array( lens[ j ]+5 );
expected = new Float64Array( lens[ j ]+5 );
for ( i = 0; i < x.length; i++ ) {
x[ i ] = i + 1;
y[ i ] = -1.0;
expected[ i ] = -1.0;
}
for ( i = 0; i < lens[ j ]; i++ ) {
expected[ 3+i ] = x[ 2+i ];
}

dcopy( lens[ j ], x, 1, 2, y, 1, 3 );

t.deepEqual( y, expected, 'returns expected value when N='+lens[ j ] );
}
t.end();
});
29 changes: 29 additions & 0 deletions lib/node_modules/@stdlib/blas/base/dcopy/test/test.ndarray.native.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -305,3 +305,32 @@ tape( 'if both strides are equal to `1`, the function efficiently copies element

t.end();
});

tape( 'if both strides are equal to `1` and the offsets are non-zero, the function copies all elements spanning multiple unrolled blocks', opts, function test( t ) {
var expected;
var lens;
var x;
var y;
var i;
var j;

lens = [ 14, 15, 16, 21, 28 ];
for ( j = 0; j < lens.length; j++ ) {
x = new Float64Array( lens[ j ]+5 );
y = new Float64Array( lens[ j ]+5 );
expected = new Float64Array( lens[ j ]+5 );
for ( i = 0; i < x.length; i++ ) {
x[ i ] = i + 1;
y[ i ] = -1.0;
expected[ i ] = -1.0;
}
for ( i = 0; i < lens[ j ]; i++ ) {
expected[ 3+i ] = x[ 2+i ];
}

dcopy( lens[ j ], x, 1, 2, y, 1, 3 );

t.deepEqual( y, expected, 'returns expected value when N='+lens[ j ] );
}
t.end();
});
2 changes: 2 additions & 0 deletions lib/node_modules/@stdlib/blas/base/scopy/src/scopy_ndarray.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,8 @@ void API_SUFFIX(c_scopy_ndarray)( const CBLAS_INT N, const float *X, const CBLAS
Y[ iy+4 ] = X[ ix+4 ];
Y[ iy+5 ] = X[ ix+5 ];
Y[ iy+6 ] = X[ ix+6 ];
ix += M;
iy += M;
}
return;
}
Expand Down
29 changes: 29 additions & 0 deletions lib/node_modules/@stdlib/blas/base/scopy/test/test.ndarray.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -296,3 +296,32 @@ tape( 'if both strides are equal to `1`, the function efficiently copies element

t.end();
});

tape( 'if both strides are equal to `1` and the offsets are non-zero, the function copies all elements spanning multiple unrolled blocks', function test( t ) {
var expected;
var lens;
var x;
var y;
var i;
var j;

lens = [ 14, 15, 16, 21, 28 ];
for ( j = 0; j < lens.length; j++ ) {
x = new Float32Array( lens[ j ]+5 );
y = new Float32Array( lens[ j ]+5 );
expected = new Float32Array( lens[ j ]+5 );
for ( i = 0; i < x.length; i++ ) {
x[ i ] = i + 1;
y[ i ] = -1.0;
expected[ i ] = -1.0;
}
for ( i = 0; i < lens[ j ]; i++ ) {
expected[ 3+i ] = x[ 2+i ];
}

scopy( lens[ j ], x, 1, 2, y, 1, 3 );

t.deepEqual( y, expected, 'returns expected value when N='+lens[ j ] );
}
t.end();
});
29 changes: 29 additions & 0 deletions lib/node_modules/@stdlib/blas/base/scopy/test/test.ndarray.native.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -305,3 +305,32 @@ tape( 'if both strides are equal to `1`, the function efficiently copies element

t.end();
});

tape( 'if both strides are equal to `1` and the offsets are non-zero, the function copies all elements spanning multiple unrolled blocks', opts, function test( t ) {
var expected;
var lens;
var x;
var y;
var i;
var j;

lens = [ 14, 15, 16, 21, 28 ];
for ( j = 0; j < lens.length; j++ ) {
x = new Float32Array( lens[ j ]+5 );
y = new Float32Array( lens[ j ]+5 );
expected = new Float32Array( lens[ j ]+5 );
for ( i = 0; i < x.length; i++ ) {
x[ i ] = i + 1;
y[ i ] = -1.0;
expected[ i ] = -1.0;
}
for ( i = 0; i < lens[ j ]; i++ ) {
expected[ 3+i ] = x[ 2+i ];
}

scopy( lens[ j ], x, 1, 2, y, 1, 3 );

t.deepEqual( y, expected, 'returns expected value when N='+lens[ j ] );
}
t.end();
});
Loading