Dear list,
I've got two threads that update the same location. One of them takes a lot
of time. Given this, my question is if there is a reason to prefer an atom
or a ref?
i.e.:
1) with an atom:
(def a (atom []))
;; the slow thread
(future (swap! a #(do (Thread/sleep 5000) (conj % 1
;;
Atoms are simpler, and I believe more efficient, than refs. An atom can
update itself atomically, but if you have two atoms, you cannot update both
in the same atomic transaction. With two refs, you can.
In general, it's a good idea to use atoms unless you need the additional
functionality that
One update taking much longer than the other, isn't optimal for either
atoms or refs.
In atoms, the faster update wins, which could randomly lead to a stream of
faster updates starving few slow updates.
I believe refs behave similarly, though if you can make your updates
commutative, you can take
Hi,
On Tuesday, September 30, 2014 12:01:33 PM UTC+2, Joachim De Beule wrote:
Dear list,
I've got two threads that update the same location. One of them takes a
lot of time. Given this, my question is if there is a reason to prefer an
atom or a ref?
is it really just two threads? Then
