nice work. A basically like to multiply n matrices B (400,400) with one Matrix A (400,400). I used arrayfun before, but your code is about 2.5 times faster. Soon I will get a Tesla NVidia graphic card and I would like to ask, is there any possibility to run the calculation on the GPU instead on the CPU?
I tried to call mtimesx with tww GPU Arrays but as my computation times is still fast (on my slow GPU) I guess he was not using my GPU right now.
Perhaps I also missed this, but to my understanding mtimesx is using the CPU only, or??
I had a question about how the flags 't' and 'c' are used by sparse matrix arguments. I assume that the flags are there to circumvent explicit transposes and its associated overhead. Yet, it doesn't appear that sparse mtimesx operations make use of them. In the test below, I see basically the same execution speeds in all 3 versions. Is it true that sparse operations can't make use of the flags, and is it deliberate/inevitable? It would be a great benefit to be able to avoid transposing tall sparse matrices because of the high memory consumption that comes with that.
From my understanding, the file "mex_C_win64.xml" replaces "mexopts.bat".
"Error using mtimesx_build (line 169)
A C/C++ compiler has not been selected with mex -setup". I replaced line 169 by : mexopts = [prefdir '\mex_C_win64.xml'];. It did work with a few warnings...
Thanks, James. This thread describes where the information has moved to
I can post those xml files somewhere, if you like. However, is there no way to get mtimesx to compile simply with the "mex" command, maybe with a small sacrifice in performance?
Hmmm ... Well, I don't have access to R2014 so I will have to code in the blind on this. My understanding is that MATLAB will in real time make a guess as to the correct compiler to use based on source code extension. So the information that was in mexopts.bat (user had already selected a compiler) is no longer available for my automatic build stuff. I will put out an update shortly, but may have to ask for info direct from the user for compiler stuff. I will try to get it out there sometime next week.