[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Changeset]: Add some single precision test code..
From: |
David Bateman |
Subject: |
Re: [Changeset]: Add some single precision test code.. |
Date: |
Mon, 02 Jun 2008 21:26:55 +0200 |
User-agent: |
Thunderbird 2.0.0.12 (X11/20080306) |
John W. Eaton wrote:
> On 28-May-2008, David Bateman wrote:
>
> | Ok, here is my first step at adding single precision test code.. This
> | going to take a while to add it all, but already this has been
> | profitable as I found some bugs that are also addressed in this patch.
>
> I applied this changeset.
>
> Thanks,
>
> jwe
>
The part of the src/ChangeLog
<quote>
2008-05-28 David Bateman <address@hidden>
* DLD-FUNCTIONS/betainc.cc: Move test code here. Add test code for
single precision type.
* DLD-FUNCTIONS/gammainc.cc: Ditto.
* DLD-FUNCTIONS/gcd.cc (Fgcd): Support single precision. Move test
code here. Add test code for single precision type.
* data.cc: Move test code here and add tests for single precision
type. Add tests for Fislogical.
(NATIVE_REDUCTION): Support the 'double' argument.
* mapper.cc: Move test code here and add tests for single precision
type.
* ov-float.cc (CD_SCALAR_MAPPER): Replace Complex with
FloatComplex.
(erf, erfc, abs, acos, asin, asinh, atan, atanh, ceil, cos, cosh,
exp, expm1, floor, log10, log1p, sin, sinh, sqrt, tan, tanh): Use
float version of base functions (ie. ::sinf and not ::sin)
(lgamma, acosh, log, log2, log10, log1p): Use single precision
infinity value.
* ov-flt-complex.cc (erf, erfc, expm1, log1p): Use the float
version of base functions.
* ov-flt-cx-mat.cc (DARRY_MAPPER): Replace NDArray with FloatNDArray.
(erf, erdc, gamma): Use FloatNDArray::dmapper and not
NDArray::dmapper.
* ov-flt-cx-mat.h (is_double_type): Delete.
(is_single_type): New method.
* ov-flt-re-mat.cc (lgamma, acosh, log, log2, log10, log1p,
sqrt): Use single precision infinity value.
</quote>
appears to be missing.
D.