Re: slow agenda view and scrolling through headlines

[Susan Addy <susan.e.addy@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 4413 bytes --]

Hi Nick,
    I have looked through the archives and tried what I found, including
the linum-mode and property-inheritance, which didn't affect the problem. I
also tried to change to the latest version of org mode and (I think) emacs,
which didn't help. I don't recall seeing the font-lock problem, though I
may have tried it earlier. I'll look for that one again. Most of the
archive questions had a slow agenda all the time, and not only sometimes.

   I tried your commands below (thank you!) and have some interesting
results, but I don't know enough to interpret them. If I am interpreting
right, org-scan-tags  sometimes takes much longer, but I have no idea why
it would be fast sometimes and slow others.

Here are the top 10 listings when operation is fast (about 3 seconds)

Function Name                                                 Call Count
Elapsed Time  Average Time

============================================================  ==========

org-agenda                                                    1
1.943013      1.943013

org-agenda-run-series                                         1
1.590022      1.590022

org-let2                                                      2
1.1610179999  0.5805089999

org-tags-view                                                 1
0.826727      0.826727

org-scan-tags                                                 5
0.760431      0.1520862

org-let                                                       2
0.426281      0.2131405

org-agenda-get-restriction-and-command                        1
0.3529        0.3529

org-agenda-list                                               1
0.334261      0.334261

org-prepare-agenda-buffers                                    2
0.3264879999  0.1632439999

org-prepare-agenda                                            3
0.3113339999  0.1037779999

org-get-entries-from-diary                                    1
0.266385      0.26638


When agenda is slow:

Function Name                                                 Call
Count  Elapsed Time  Average Time

============================================================  ==========

org-agenda                                                    1
12.133429     12.133429

org-agenda-run-series                                         1
11.768025     11.768025

org-let2                                                      2
10.856247999  5.4281239999

org-tags-view                                                 1
10.427989     10.427989

org-scan-tags                                                 5
10.402578000  2.0805156000

org-let                                                       2
0.908944      0.454472

org-finalize-agenda                                           3
0.8176239999  0.2725413333

org-agenda-get-day-entries                                    10
0.6845889999  0.0684588999

org-agenda-get-scheduled                                      10
0.4447109999  0.0444711

org-agenda-list                                               1
0.428234      0.428234

On Sun, Nov 13, 2011 at 2:58 PM, Nick Dokos <nicholas.dokos@hp.com> wrote:

> Susan Addy <susan.e.addy@gmail.com> wrote:
>
> > Thanks Bernt. I checked and Linum-mode is not enabled.
> >
> > Also, I using emacs 22.3.1 and org-mode 7.7 (if it helps).
> >
> >
> > On Sun, Nov 13, 2011 at 2:16 PM, Bernt Hansen <bernt@norang.ca> wrote:
> >
> >     Susan Addy <susan.e.addy@gmail.com> writes:
> >
> >     > Below is a list of my agenda command - maybe it is too long? When
> it
> >     > is slow, it is slow no matter what TODO I am trying to compile.
> >     >
> >     > Any thoughts?
> >     > Thank you!
> >
> >     If you have linum-mode enabled you should turn that off.  Others have
> >     reported slow navigation when that is enabled.
> >
> >     HTH,
> >     Bernt
> >
> >
>
> There have been various discussions on agenda slowness in the past -
> check the mailing list for some guesses (linum-mode, font-lock, property
> inheritance, code compilation and perhaps others have been suggested as
> possible causes). Sometimes the guesses pan out, sometimes not.
>
> The most important thing you can do, imo, is to profile the code, first
> when it's behaving well and then when it's behaving badly.
>
> All you have to do is
>
>    M-x elp-instrument-package RET org RET
>
> run your agenda command and then
>
>    M-x elp-results RET
>
> to get the results.
>
> Nick
>

[-- Attachment #2: Type: text/html, Size: 5980 bytes --]