I have been spending some time this week hacking GalSim and noticed a few places where builtins are known to be faster than the numpy equivalents generally. This is not likely to move the needle in terms of the overall efficiency, but this is something I had to do anyway during the hack and I thought I might as well try to get it merged here.
Oh, I don't doubt that one bit. These are some changes that I had to do anyway to get my hacky version running. Btw, I'm not in any hurry and this need not result in a patch release immediately.
I have been spending some time this week hacking GalSim and noticed a few places where builtins are known to be faster than the
numpy
equivalents generally. This is not likely to move the needle in terms of the overall efficiency, but this is something I had to do anyway during the hack and I thought I might as well try to get it merged here.