Whit Armstrong

also, which exact gcc version are you on? and this is on Fedora as well? -Whit On Sun, Jul 8, 2012 at 2:21 PM, jhjourdan [email protected] wrote: > I have...

Thanks for digging into this. can you send or post your test code? -Whit On Sun, Jul 8, 2012 at 5:53 PM, jhjourdan [email protected] wrote: > Le 08/07/2012 20:38, Whit...

The more I look at this the worse it seems. There is a snip here, which basically describes how gcc did this to be more efficient w/ stack usage: http://gcc.gnu.org/gcc-4.7/changes.html...

I think rvalues are the way to go. However, that could complicate the classes significantly. All the permutations of rvalues in the hyper-parameters would need to be handled... for instance,...

you're way ahead of me! Ok, let me check out your fork. So, this is working for you in your production code? -Whit On Mon, Jul 16, 2012 at 9:03...

Unfortunately the README file didn't keep up with the changing codebase, I've updated it to display the eight schools model, which is much less complicated than the herd model. Just...

Sure. It's basically as simple as adding a new distribution class. The distribution class just needs to provide a method for calculating the log likelihood and override the default jumping...

yes, it's there in the cpp library only. you can add it to the R package if you copy one of the other examples.

Yes, exactly. Just add a wrapper for the distribution and rebuild the package. Or, wait a few weeks and I'll try to add it. There are a few other distribution...

yes. Rcppbugs needs to be updated per all the changes in the c++ backend. Do you need to use it via R, or could you perhaps use the c++ version?