Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.soft-sys.math.mathematica > #13485 > unrolled thread
| Started by | Frank Iannarilli <frankeye@cox.net> |
|---|---|
| First post | 2013-02-14 07:09 +0000 |
| Last post | 2013-02-14 07:09 +0000 |
| Articles | 1 — 1 participant |
Back to article view | Back to comp.soft-sys.math.mathematica
This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by
below is the oldest one visible, not the original post.
Re: Win7 "not responding"; WinXP no problem w/ my Manipulate Frank Iannarilli <frankeye@cox.net> - 2013-02-14 07:09 +0000
| From | Frank Iannarilli <frankeye@cox.net> |
|---|---|
| Date | 2013-02-14 07:09 +0000 |
| Subject | Re: Win7 "not responding"; WinXP no problem w/ my Manipulate |
| Message-ID | <kfi2ij$d14$1@smc.vnet.net> |
I think I finally have the correct diagnosis and solution(s), below.
The Windows Ghosting "solution" I posted above is not a solution. However, the link:
http://blogs.technet.com/b/askperf/archive/2010/09/10/de-ghosting-your-windows.aspx
does describe the underlying cause of why Win7 is less forgiving then WinXP of Manipulate and Dynamic cells whose evaluation-loop computations run too long when in (default) SynchronousUpdating->True mode (i.e. clogging the preemptive front-end/kernel link, thus blocking the front-end and triggering Win7's treatment of unresponsive programs).
So here are my findings/solutions for accommodating long-running (i.e. > 5 seconds) computations comprising the Manipulate or Dynamic argument, probably recommended in general but specifically geared to Win7:
* Use the option SynchronousUpdating->False.
* Don't rely on the idiom of wrapping Manipulate within Style[..., DynamicEvaluationTimeout->10000]. This would prevent the $Abort timeout, but under Win7 makes the front-end susceptible to Windows Ghosting.
* If a Button included as a Manipulate control invokes long-running computation, use the Button option Method->"Queued".
* Be mindful with "interior" Dynamic cells, e.g. Dynamic inside Manipulate, for by default these use SynchronousUpdating->True. Heavy computation with in their arguments can lock-up the front-end. (I was heedless of this point, which is why in my original post, I gained no relief merely by employing SynchronousUpdating->False in the "outer" containing Manipulate -- I also needed this option in some of my interior Dynamic cells).
* Realize that with ContinuousAction->False (and SynchronousUpdating->False, the user can still muck with controls in a manner that generates excessive delay, which may be confusing. For example, for the cell:
Manipulate[Pause[10]; x, {x, 0, 10}, ContinuousAction -> False,
SynchronousUpdating -> False]
the user can drag the slider in all manner during the Pause, but with each release of the mouse, she generates another evaluation loop that gets cued to the main link, therefore the overall manifested pause delay can be tens of seconds, not merely the nominal Pause[10] 10 secs.
Back to top | Article view | comp.soft-sys.math.mathematica
csiph-web