diff options
author | marco <marco@3ad0048d-3df7-0310-abae-a5850022a9f2> | 2011-05-31 10:56:28 +0000 |
---|---|---|
committer | marco <marco@3ad0048d-3df7-0310-abae-a5850022a9f2> | 2011-05-31 10:56:28 +0000 |
commit | e4cf596a7d96b54a4bfd4ef5191786990c04458e (patch) | |
tree | bcc2641208d07603e498ebd45a8f2f2fc2ae3e6e | |
parent | cd890348768ac3997f1b4249b3b1d2da9def830d (diff) | |
download | fpc-e4cf596a7d96b54a4bfd4ef5191786990c04458e.tar.gz |
--- Merging r17178 into '.':
U packages/univint/src/cblas.pas
# revisions: 17178
------------------------------------------------------------------------
r17178 | jonas | 2011-03-25 11:51:15 +0100 (Fri, 25 Mar 2011) | 3 lines
Changed paths:
M /trunk/packages/univint/src/cblas.pas
- reverted r17177 for univint because these units are automatically converted
from C headers, and manual changes make future updating more difficult
------------------------------------------------------------------------
git-svn-id: http://svn.freepascal.org/svn/fpc/branches/fixes_2_4@17616 3ad0048d-3df7-0310-abae-a5850022a9f2
-rw-r--r-- | packages/univint/src/cblas.pas | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/packages/univint/src/cblas.pas b/packages/univint/src/cblas.pas index b8a215c676..fd87c1e8ca 100644 --- a/packages/univint/src/cblas.pas +++ b/packages/univint/src/cblas.pas @@ -580,7 +580,7 @@ procedure vMultMatMat_32x32( ConstVectorFloat A[32][8]; ConstVectorFloat B[32][8 { ------------------------------------------------------------------------------------------------- These routines provide optimized support for common small matrix multiplications. They use - the scalar floating point unit and have no dependency on SIMD instructions. They are intended + the scalar floating point unit and have no dependancy on SIMD instructions. They are intended as complements to the AltiVec-only routines above. They do not check for parameter errors. They just do the multiplication as fast as possible. Matrices are presumed to use row major storage. Because these are all square, column major matrices can be multiplied by simply reversing the parameters. |