On Mon, Aug 16, 2010 at 11:47 AM, Romain Francois
wrote:
> Le 16/08/10 15:33, Douglas Bates a écrit :
>>
>> On Mon, Aug 16, 2010 at 8:10 AM, Romain Francois
>> wrote:
>>>
>>> Nothing really conclusive here ...
>>>
>>> I've added a preliminary version of SUGAR_BLOCK_3 (only for the case
>>> where
Le 16/08/10 15:33, Douglas Bates a écrit :
On Mon, Aug 16, 2010 at 8:10 AM, Romain Francois
wrote:
Nothing really conclusive here ...
I've added a preliminary version of SUGAR_BLOCK_3 (only for the case where
all inputs are vectors) so that one can write :
incl8<- '
inline double y_log_y(dou
Le 16/08/10 15:33, Douglas Bates a écrit :
On Mon, Aug 16, 2010 at 8:10 AM, Romain Francois
wrote:
Nothing really conclusive here ...
I've added a preliminary version of SUGAR_BLOCK_3 (only for the case where
all inputs are vectors) so that one can write :
incl8<- '
inline double y_log_y(dou
On Mon, Aug 16, 2010 at 8:10 AM, Romain Francois
wrote:
> Nothing really conclusive here ...
>
> I've added a preliminary version of SUGAR_BLOCK_3 (only for the case where
> all inputs are vectors) so that one can write :
>
> incl8 <- '
> inline double y_log_y(double y, double mu) {
> return y
Nothing really conclusive here ...
I've added a preliminary version of SUGAR_BLOCK_3 (only for the case
where all inputs are vectors) so that one can write :
incl8 <- '
inline double y_log_y(double y, double mu) {
return y ? y*log(y/mu) : 0.;
}
inline double yMu(double y, double mu) {
On Sun, Aug 15, 2010 at 10:34 AM, Douglas Bates wrote:
> On Sun, Aug 15, 2010 at 3:39 AM, Romain Francois
> wrote:
>> Hi,
>>
>> Could you provide some minimal example using inline, etc ...
>
> I enclose a comparison of several different versions that I wrote.
> Interestingly the old code, called
On Sun, Aug 15, 2010 at 3:39 AM, Romain Francois
wrote:
> Hi,
>
> Could you provide some minimal example using inline, etc ...
I enclose a comparison of several different versions that I wrote.
Interestingly the old code, called by fx3, does reasonably well,
although I'm not sure if this is a fai
Hi,
Could you provide some minimal example using inline, etc ...
Here are a few hints in the meantime. In the next version of Rcpp, I've
added a few things that help generating sugar functions. This is how for
example the sugar version of choose is currently implemented :
SUGAR_BLOCK_2(choos
In profiling code for generalized linear mixed models for binary
responses I found that a substantial portion of the execution time was
being spend in evaluating the R functions for inverse links and for
the deviance residuals. As I result I wrote C code (in
$RSRC/src/library/stats/src/family.c) o
