-
Notifications
You must be signed in to change notification settings - Fork 24
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Seamlessly removing the use of backingpath for sub and attach. #56
Conversation
Wow ok, I've just understood that this PR updates whenever I update my repo. So, if you want I added a method to efficiently transpose a |
Hi @privefl, with transpose you are more in my territory with bigalgebra. My fork is the most recent. Please compare yours to the version I have with the armadillo backend. Things like transpose belong in that package. |
Hi @cdeterman, I've tested matrix multiplication with Armadillo backend and it is very slow (compared for example with Eigen, or directly by blocks within R). Moreover, I'm not sure that As you (and Mike) prefer. |
I agree that the default is quite slow but that depends entirely on the BLAS backend being used which by default is the R BLAS which is not very good. If you use OpenBLAS or ATLAS it is much better. This flexibility is the reason armadillo was chose and not Eigen or our own custom scripts. As for the proper place of transpose this discussion happened before hence why I made the statement here. That said if the others agree it would be better here I wouldn't argue. |
I'll redo some testing (for matrix multiplication) and open an issue in |
@kaneplusplus Not to be a stickler but did you want the entire merge here? I am mainly referring to the addition of |
If you don't want the
They should be independent files. |
@cdeterman Thanks for pointing this out. I'll remove now. |
So,
so that now, you can use
sub.big.matrix
andattach.big.matrix
without specifying thebackingpath
parameter (you still can for backward compatibility).There are lots of man files that have been modified (without real change in the documentation) due to the use of the latest version of roxygen2.
You can see the new test file if it's not clear enough.