summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2009-04-27first version of configure(.hs)Alberto Ruiz
2009-04-27more defined (|>) and (><) as suggested by Tracy WadleighAlberto Ruiz
2009-04-24More defined list-to-V/M operators: (|>|), (>|<).Tracy Wadleigh
Provides analogs of the (|>) and (><) operators that first apply an appropriate 'take' to the given lists so that they may be safely used on lists that are too long (or infinite) -- a feature I find particularly useful from the interactive prompt. As these operators are more defined, I would ask the package maintainer to consider, rather than adding (|>|) and (>|<), just updating the definitions of (|>) and (><) with the new semantics.
2009-04-18Algorithms module reorganized to fix documentation structureAlberto Ruiz
2009-04-18added minimizeVectorBFGS2Alberto Ruiz
2009-04-17restored C trans and constant for comparisonAlberto Ruiz
2009-04-17removed static experimentAlberto Ruiz
2009-04-10finit not required for ghc >= 6.10.2Alberto Ruiz
2009-01-19credits to Daniel Shüssler and fix scale for complex trig testsAlberto Ruiz
2009-01-19Compatibility with QuickCheck 2 as well as QuickCheck 1Daniel Schüssler
Used the C preprocessor to make it compile against either QuickCheck version. Added some implementations for the new "shrink" method of class Arbitrary. The tests (runTests 10) succeed when compiled with QC 1. With QC 2 (runTests 10 too) the fourth "vector operations" test (complex trigonometry) now fails. I don't know whether this is a bug in my changes, or whether QC 2 just generates Doubles differently and thus uncovered a real bug.
2008-12-08support for the accelerate framework on OS X, contributed by Simon BeaumontAlberto Ruiz
2008-11-30QuickCheck<2, ifdef i386 finit, and credit to Simon BeaumontAlberto Ruiz
2008-11-15minor changesAlberto Ruiz
2008-11-15zipVectorAlberto Ruiz
2008-11-14subMatrix reimplemented and removed auxi.cAlberto Ruiz
2008-11-14conj using mapVectorAlberto Ruiz
2008-11-14constantDAlberto Ruiz
2008-11-11fix transdataAlberto Ruiz
2008-11-08replaced transdataAlberto Ruiz
2008-11-07old tag: 0.5.0.1Alberto Ruiz
2008-11-05improved constantAlberto Ruiz
2008-11-05improved diagRectAlberto Ruiz
2008-11-05diag using STAlberto Ruiz
2008-11-04multiply/trans okAlberto Ruiz
2008-10-27simplified READMEAlberto Ruiz
2008-10-27added luSolveAlberto Ruiz
2008-10-24added dgetrs support for solving linear systems from LU factorizationReiner Pope
2008-10-23NaN problem solved with asm(finit)Alberto Ruiz
2008-10-22-WallAlberto Ruiz
2008-10-17acknowledgement to Dylan Simon, updated readme and fixed mapCAuxAlberto Ruiz
2008-10-17fast signum/abs Vector instanceDylan Simon
2008-10-16improved benchmarkAlberto Ruiz
2008-10-16added examples/tests.hsAlberto Ruiz
2008-10-15debug info for the NaN bugAlberto Ruiz
2008-10-15expose-bug flagAlberto Ruiz
2008-10-14minor changeAlberto Ruiz
2008-10-14ok with ghc-6.10-rcAlberto Ruiz
2008-10-05workaround optionAlberto Ruiz
2008-10-02alternative multiply versionsAlberto Ruiz
2008-06-29foldVector benchmarkAlberto Ruiz
2008-06-27tested with ghc-6.8.3Alberto Ruiz
2008-06-19ST bug fixed with NOINLINE newVector/MatrixAlberto Ruiz
2008-06-19foldVector benchmarkAlberto Ruiz
2008-06-12minor fixAlberto Ruiz
2008-06-10other benchmarkAlberto Ruiz
2008-06-10added examples/benchmarks.hsAlberto Ruiz
2008-06-10Unpack Matrix type, and add -funsafe flagDon Stewart
-funsafe optionally compiles out the bounds checks on indexing matrices and vectors. Yields good speedups on tight loops. Not enabled by default.
2008-06-10make empty lines with backslashes CPP friendlyDon Stewart
2008-06-09Use mallocPlainForeignPtrBytes if availableDon Stewart
As for bytestrings, use PlainForeignPtrs if supported, which have less overhead, as values are allocated on the Haskell heap without any wasted finalisers needed. Should mean less resources used if many small vectors are created.
2008-06-09acknowledgement to Don StewartAlberto Ruiz
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-14 19:46:30 +0000