My not be any different. In that case it would be a simple project of creating a new page on the Luajit site that shows the Lua code that does it with ffi along side the equivalent native C program and some bench marks that show the results side by side. The sample / example would allocate the array, Fill it with a set of doubles read from a file, Run the sum and return both the sum and average value. If we are successful the Luajit code and the C code should yield very similar results when ran across a few million items. If not then we know exactly how much slower it will be for this use case. This seems like a trivial use case but I do a lot of this kind of work in our ML project and we run this look mil. We actually have a 3 deep array the first two levels contain pointers to the next level array. On Wed, Dec 10, 2014 at 9:56 AM, Alex <initrd.gz@xxxxxxxxx> wrote: > > Implement operator to initialize high density array of type with fact > access similar to C > > How is this different from FFI arrays? I.e. `ffi.new("double[?]", maxele)` > or `ffi.cast("double*", ffi.C.malloc(ffi.sizeof("double")*maxele))` > > -- > Sincerely, > Alex Parrill > -- Joseph Ellsworth Principal Engineer 206-601-2985