AW: [Haskell-cafe] Something like optimistic evaluation

[Daniil Elovkov]

Nicu Ionita wrote:
> 
> I don't know if this would be worth, but theoretically one could go on 
> and evaluate those thunks that:
>  
> a) would be evaluated anyway (after the current IO operation have been 
> completed)
> b) do not depend on the result of the current operation
>  
> And, of course, the GC could work in this time also.

Yes, and btw, this work would _definitely_ not be wasted, unlike 
evaluating thunks.



> Nicu
> 
>     -----Ursprüngliche Nachricht-----
>     *Von:* haskell-cafe-bounces at haskell.org
>     [mailto:haskell-cafe-bounces at haskell.org] *Im Auftrag von *Brent Yorgey
>     *Gesendet:* Dienstag, 29. April 2008 16:42
>     *An:* Daniil Elovkov
>     *Cc:* haskell-cafe at haskell.org
>     *Betreff:* Re: [Haskell-cafe] Something like optimistic evaluation
> 
> 
>     On Mon, Apr 28, 2008 at 6:09 PM, Daniil Elovkov
>     <daniil.elovkov at googlemail.com
>     <mailto:daniil.elovkov at googlemail.com>> wrote:
> 
>         Hello
> 
>         Somewhat on the topic of optimistic evaluation, I've just
>         thought of another way to evaluate thunks.
> 
>         When the program is about to block on some IO, what if we start
>         a thread to evaluate (any) unevaluated thunks. We'll have
>         additional system thread, but the blocked one will not actually
>         consume any processor time.
> 
>         This would take place only when the program is compiled as
>         threaded and run with -N k, k>1.
> 
>         The RTS knows at least about some operations that will block,
>         those which IO operations are implemented with. for example. It
>         could merely start a process of evaluating any (or something
>         more clever) outstanding thunks right before going into one of
>         those operations and stop it when it's back.
> 
>         Of course, it's not like optimistic evaluation because we don't
>         avoid creating thunks. But in a sense it's similar. It could
>         also be compared with incremental garbage collection :)
> 
>         Has something like that been done, discussed?
> 
> 
>     This sounds like it could be helpful in certain circumstances, but
>     in many cases it could probably lead to unpredictable (and
>     uncontrollable!) memory usage.  I could imagine a situation where my
>     program is running along just fine, and then one day it takes a long
>     time to do a read from the network due to latency or whatever, and
>     suddenly memory usage shoots through the roof, due to evaluation of
>     some infinite (or even just very large) data structure. 
> 
>     -Brent