Timestamps during tracing

[José Fonseca]

On Tue, Apr 16, 2013 at 4:47 PM, Eugene Velesevich
<eugvelesevich at gmail.com>wrote:

> [Jose, apologies for the duplicated e-mail]
>
> On Tue, Apr 16, 2013 at 7:23 PM, José Fonseca <jose.r.fonseca at gmail.com>
> wrote:
> > It sounds like you already have something implemented. If so, could you
> > provide a screenshot of what you've done?
>
> See the attachment.
>

Is is true: sometimes a picture is worth thousand words.  Nice work.


> Apologies for a slight discrepancy: previously
> we've mentioned GPU timestamps, but the attached screenshot uses a
> build with the GPU timeline disabled.
>

I think we need to be clearer regarding the origin (recorded vs replayed)
of the timings.  For example, instead of "Timeline" we should have
"Recorded Timeline", "Replayed Timeline", etc, so one knows what figures
one is seeing.

Whether the timeline should be its own window, or in pane, it is really
orthogonal to the origin of the timings (that is, the replayed timings
could also be in a pane). I personally like it as on your screenshot: pane
seems to give a nice organization, especially as one switches between
drawables.


> This also shows how we implemented per-drawable call lists: you can
> see tabs that allow switching between different drawables touched by
> the application.
>

Yes, it does look nice. I can see the value of breaking things like that
when multiple widgets.

As I said in my other email, if you can make it so that people switch back
to old behavior (all drawables lumped together) I'm happy to accept it.

Jose
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/apitrace/attachments/20130416/8a0c3b7d/attachment.html>