[Cs254f11] subst

Fri Nov 18 19:23:10 EST 2011

BTW I've been getting even better multicore utilization than pmap can provide with this:

(defn pmapall
  "Like pmap but: 1) coll should be finite, 2) the returned sequence
   will not be lazy, 3) calls to f may occur in any order, to maximize
   multicore processor utilization, and 4) takes only one coll so far."
  [f coll]
  (let [agents (map #(agent % :error-handler (fn [agnt except] (println except))) coll)]
    (dorun (map #(send % f) agents))
    (apply await agents)
    (doall (map deref agents))))

One drawback is that calls to pmapall cannot be nested, but if you're careful about how you use it this can be great. Check out recent loads on fly:compute-4-2 (and some other 4-*) to see what I mean.

 -Lee

On Nov 18, 2011, at 6:56 PM, Wm. Josiah Erikson wrote:

> Ooh neat, I will try changing something to a pmap after thinking about which one makes sense. Probably the mapping of error down the population....
> 
> Connected by DROID on Verizon Wireless
> 
> 
> -----Original message-----
> From: Lee Spector <lspector at hampshire.edu>
> To: "Wm. Josiah Erikson" <wjerikson at hampshire.edu>
> Cc: cs254f11 at lists.hampshire.edu
> Sent: Fri, Nov 18, 2011 22:08:57 GMT+00:00
> Subject: Re: [Cs254f11] subst
> 
> 
> Doesn't it feel breezy clean? :-)
> 
> And I think you're now ready to go multicore pretty easily, maybe just by judiciously changing a map to a pmap or something along those lines (the details of which will take a little thought about what you want to run concurrently, but in any event it should be quick and safe).
> 
> -Lee
> 
> On Nov 18, 2011, at 4:45 PM, Wm. Josiah Erikson wrote:
> 
> > Yep, now my error function is just (note this is 1/4 the size it was before, I think):
> > 
> > (reduce + (map #(Math/abs (- (int (eval (postwalk-replace % individual))) %2)) terminal-map scores)))
> > 
> > That postwalk-replace just pulls out my placeholder function names that were created by calls to random-code, and replaces them with the appropriate numbers for the song, and since it's inside a function that's being mapped down my terminal-map and scores, two lists that correspond to each other, it does it appropriately for each song.
> > 
> > terminal-map is a list of maps where the keys are function names and the values are what they evaluate to when passed the corresponding song.
> > 
> > So I now am free of any global re-defing. Yay!
> > -Josiah
> > 
> > 
> > On 11/18/11 2:37 PM, Lee Spector wrote:
> >> Fabulous!
> >> 
> >> -Lee
> >> 
> >> 
> >> 
> >> On Nov 18, 2011, at 2:30 PM, Wm. Josiah Erikson wrote:
> >> 
> >>> Yeah, it's perfect (this isn't exactly what I'm going end up doing, but it shows the concept). This will actually be really easy! Check it out:
> >>> 
> >>> critic_evolution.core=>
> >>> (use 'clojure.walk)
> >>> nil
> >>> 
> >>> critic_evolution.core=>
> >>> (zipmap list-of-retrieval-functions list-of-analysis-functions)
> >>> {(p_i_m) percent_in_mixolydian,
> >>> (a_v) average_velocity,
> >>> (l_c_r_n) least_common_real_note,
> >>> (p_i_l) percent_in_lydian,
> >>> (a_n) average_note,
> >>> (l_n) low_note,
> >>> (p_o_pb) percentage_of_pitchbend,
> >>> (p_i_b) percent_in_blues,
> >>> (m_c_r_n) most_common_real_note,
> >>> (n_o_d_r_n) number_of_different_real_notes,
> >>> (h_n) high_note,
> >>> (a_r_n) average_real_note,
> >>> (s_d_m) std_dev_mostcom,
> >>> (p_o_p) percentage_of_percussion,
> >>> (p_i_p) percent_in_pentatonic}
> >>> 
> >>> critic_evolution.core=>
> >>> (def function-lookup (zipmap list-of-retrieval-functions list-of-analysis-functions))
> >>> #'critic_evolution.core/function-lookup
> >>> critic_evolution.core=>
> >>> (def mark (random-code 5))
> >>> #'critic_evolution.core/mark
> >>> critic_evolution.core=>
> >>> mark
> >>> (max
> >>> (-
> >>> (min (* (p_i_p) (- (l_n) (s_d_m))) (l_c_r_n))
> >>> (pmod
> >>> (n_o_d_r_n)
> >>> (pquot (* (p_i_m) (n_o_d_r_n)) (min (p_i_p) (p_o_pb)))))
> >>> (pd
> >>> (pquot
> >>> (pd (- (h_n) (p_i_l)) (- (a_v) (p_i_p)))
> >>> (pmod (pmod (l_n) (n_o_d_r_n)) (pmod (p_o_pb) (p_o_pb))))
> >>> (pmod (+ (pmod (p_i_b) 78) (pd (n_o_d_r_n) (p_o_pb))) 32)))
> >>> 
> >>> critic_evolution.core=>
> >>> (postwalk-replace function-lookup mark)
> >>> (max
> >>> (-
> >>> (min
> >>> (* percent_in_pentatonic (- low_note std_dev_mostcom))
> >>> least_common_real_note)
> >>> (pmod
> >>> number_of_different_real_notes
> >>> (pquot
> >>> (* percent_in_mixolydian number_of_different_real_notes)
> >>> (min percent_in_pentatonic percentage_of_pitchbend))))
> >>> (pd
> >>> (pquot
> >>> (pd
> >>> (- high_note percent_in_lydian)
> >>> (- average_velocity percent_in_pentatonic))
> >>> (pmod
> >>> (pmod low_note number_of_different_real_notes)
> >>> (pmod percentage_of_pitchbend percentage_of_pitchbend)))
> >>> (pmod
> >>> (+
> >>> (pmod percent_in_blues 78)
> >>> (pd number_of_different_real_notes percentage_of_pitchbend))
> >>> 32)))
> >>> 
> >>> critic_evolution.core=>
> >>> 
> >>> On 11/18/11 2:26 PM, Wm. Josiah Erikson wrote:
> >>>> ..and if I'm reading postwalk-replace's documentation correctly, I can create a map of things to look for and things to replace and it will do all of them. From the documentation:
> >>>> 
> >>>> Recursively transforms form by replacing keys in smap with their
> >>>> values. Like clojure/replace but works on any data structure. Does
> >>>> replacement at the leaves of the tree first.
> >>>> 
> >>>> Source:
> >>>> (defn postwalk-replace
> >>>> "...docs..."
> >>>> {:added "1.1"}
> >>>> [smap form]
> >>>> (postwalk (fn [x] (if (contains? smap x) (smap x) x)) form))
> >>>> 
> >>>> So actually postwalk-replace will do all of my work for me. Yes?
> >>>> 
> >>>> -Josiah
> >>>> 
> >>>> 
> >>>> 
> >>>> On 11/18/11 12:29 PM, Lee Spector wrote:
> >>>>> Josiah and I discussed this in person, as a way to do what he's doing more simply, and I won't repeat all of our discussion here, but others might find it useful for similar or different purposes.
> >>>>> 
> >>>>> -Lee
> >>>>> 
> >>>>> (use 'clojure.walk)
> >>>>> 
> >>>>> (defn subst
> >>>>> "Returns the given list but with all instances of that (at any depth)
> >>>>> replaced with this. Read as 'subst this for that in list'. "
> >>>>> [this that lst]
> >>>>> (postwalk-replace {that this} lst))
> >>>>> 
> >>>>> (subst 0 1 '(0 1 2 3 (0 1 2 3 (0 1 2 3) 0 1 2 3) 0 1 2 3))
> >>>>> 
> >>>>> ; (0 0 2 3 (0 0 2 3 (0 0 2 3) 0 0 2 3) 0 0 2 3)
> >>>>> 
> >>>>> 
> >>>>> -- 
> >>>>> Lee Spector, Professor of Computer Science
> >>>>> Cognitive Science, Hampshire College
> >>>>> 893 West Street, Amherst, MA 01002-3359
> >>>>> lspector at hampshire.edu, http://hampshire.edu/lspector/
> >>>>> Phone: 413-559-5352, Fax: 413-559-5438
> >>>>> 
> >>>>> _______________________________________________
> >>>>> Cs254f11 mailing list
> >>>>> Cs254f11 at lists.hampshire.edu
> >>>>> https://lists.hampshire.edu/mailman/listinfo/cs254f11
> >>> -- 
> >>> Wm. Josiah Erikson
> >>> Network Engineer
> >>> Hampshire College
> >>> Amherst, MA 01002
> >>> (413) 559-6091
> >>> 
> >>> _______________________________________________
> >>> Cs254f11 mailing list
> >>> Cs254f11 at lists.hampshire.edu
> >>> https://lists.hampshire.edu/mailman/listinfo/cs254f11
> >> --
> >> Lee Spector, Professor of Computer Science
> >> Cognitive Science, Hampshire College
> >> 893 West Street, Amherst, MA 01002-3359
> >> lspector at hampshire.edu, http://hampshire.edu/lspector/
> >> Phone: 413-559-5352, Fax: 413-559-5438
> >> 
> > 
> > -- 
> > Wm. Josiah Erikson
> > Network Engineer
> > Hampshire College
> > Amherst, MA 01002
> > (413) 559-6091
> > 
> 
> --
> Lee Spector, Professor of Computer Science
> Cognitive Science, Hampshire College
> 893 West Street, Amherst, MA 01002-3359
> lspector at hampshire.edu, http://hampshire.edu/lspector/
> Phone: 413-559-5352, Fax: 413-559-5438
> 

--
Lee Spector, Professor of Computer Science
Cognitive Science, Hampshire College
893 West Street, Amherst, MA 01002-3359
lspector at hampshire.edu, http://hampshire.edu/lspector/
Phone: 413-559-5352, Fax: 413-559-5438