Giving all my money to charity isn't a part of my ethics.

Increasing net utility (or something of the kind) is one of the things I care about. So the fact that something increases net utility is a reason to do it, and the fact that something decreases net utility is a reason not to. But net utility isn't the only thing I care about, so a thing that increases net utility isn't necessarily a thing I think I should do.

What I insist on, though, is another matter again. That's a matter of Schelling points and traditions and the like, optimized (inter alia) for being easy to remember and intuitively plausible.

So:

• Giving $1M to Joe: increases his utility, decreases mine, probably not a win overall in terms of net utility. Fails various other tests too. Not in any sense any sort of moral obligation.
• Giving $100 to Joe, who is much poorer than me: net utility increase, might be a good thing to do on those terms. Probably reasonable not to do simply on the grounds that I care more about my own utility than that of strangers, that if I'm trying to do maximum good there are others who need the money much more than Joe, etc.
• Giving $100 to a carefully chosen effective charity: close to the best thing I can do for net utility with the money. I still care more about my own utility than about strangers', though, so not necessarily obligatory even "internally".
• Giving at least a few percent of one's income to effective charities, provided one is reasonably comfortable financially: almost always a big net utility gain, not too burdensome, has the same form as various traditional practices, easy to remember and to do. I'd be comfortable recommending this as a principle everyone should be following.

The attentive reader will notice that not killing people just for being annoying clearly fits into the same category as the last of those.