Meaning of Visual Graph with Venkman Profiler HTML output

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Meaning of Visual Graph with Venkman Profiler HTML output

Glenn Boysko
Hello:

I think I've read all of the various links and articles on the Venkman
profiler, but can't find a meaningful explanation for the graph that
appears in the HTML output.

>From what I can infer, it is showing the min, max and average
total-time values for a specific method.

What doesn't make sense is how those images relate to the larger
graph-body. I would expect that it shows those min, max and average
times in relation to the total time. However, doing the math (reading
the HTML sizes of the various images), it doesn't seem to match.

A better graph might be to show the total time in relation to the
overall page time. Or show the min, max, average times inside a larger
graph body that is sized to the overall page cost.

If anyone has more information on what this graph represents, please
pass it along!

Regards,
Glenn

_______________________________________________
dev-apps-js-debugger mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-apps-js-debugger
Reply | Threaded
Open this post in threaded view
|

Re: Meaning of Visual Graph with Venkman Profiler HTML output

James Ross
[hidden email] wrote:
> Hello:
>
> What doesn't make sense is how those images relate to the larger
> graph-body. I would expect that it shows those min, max and average
> times in relation to the total time. However, doing the math (reading
> the HTML sizes of the various images), it doesn't seem to match.

The graph fills in a dark block from min execution time to max execution
time, and then adds a light marker for the average execution time.
Simple. :)

> A better graph might be to show the total time in relation to the
> overall page time. Or show the min, max, average times inside a larger
> graph body that is sized to the overall page cost.

I don't know what "overall page cost" is, but I will say that the scale
on the graphs is constant within a single time block (e.g. 500 - 1000ms).

--
James Ross <[hidden email]>
ChatZilla Developer
_______________________________________________
dev-apps-js-debugger mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-apps-js-debugger
Reply | Threaded
Open this post in threaded view
|

Re: Meaning of Visual Graph with Venkman Profiler HTML output

Glenn Boysko
Hello James:

Thanks for the response. I was expecting the total graph size to be
sized to the total time for that method call. From your comments, you
are saying that the total graph size is based on the range that method
appears in?

So for example if a method is listed in the range of 500-1000ms, what
does the overall size of graph represent? 500ms? 1000ms?

It may be me, but the graph as it currently exists doesn't seem to
convey much. The min/max/avg values are great, but in relation to what?


Regards,
Glenn

James Ross wrote:

> [hidden email] wrote:
> > Hello:
> >
> > What doesn't make sense is how those images relate to the larger
> > graph-body. I would expect that it shows those min, max and average
> > times in relation to the total time. However, doing the math (reading
> > the HTML sizes of the various images), it doesn't seem to match.
>
> The graph fills in a dark block from min execution time to max execution
> time, and then adds a light marker for the average execution time.
> Simple. :)
>
> > A better graph might be to show the total time in relation to the
> > overall page time. Or show the min, max, average times inside a larger
> > graph body that is sized to the overall page cost.
>
> I don't know what "overall page cost" is, but I will say that the scale
> on the graphs is constant within a single time block (e.g. 500 - 1000ms).
>
> --
> James Ross <[hidden email]>
> ChatZilla Developer

_______________________________________________
dev-apps-js-debugger mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-apps-js-debugger
Reply | Threaded
Open this post in threaded view
|

Re: Meaning of Visual Graph with Venkman Profiler HTML output

James Ross
[hidden email] wrote:
> So for example if a method is listed in the range of 500-1000ms, what
> does the overall size of graph represent? 500ms? 1000ms?

The graphs are scaled based on the maximum end of the range (i.e. 1000ms
in your example). I don't know the exact scale used, but it goes from 0
at the left to some value at the right which is constant within a block.

> It may be me, but the graph as it currently exists doesn't seem to
> convey much. The min/max/avg values are great, but in relation to what?

The graphs very quickly convey how the average relates to the min/max
times. This is particularly useful when spotting calls that are always
slow, vs. ones that are only slow a small fraction of time, for example.

--
James Ross <[hidden email]>
ChatZilla Developer
_______________________________________________
dev-apps-js-debugger mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-apps-js-debugger
Reply | Threaded
Open this post in threaded view
|

Re: Meaning of Visual Graph with Venkman Profiler HTML output

James Ross
James Ross wrote:
> [hidden email] wrote:
>> So for example if a method is listed in the range of 500-1000ms, what
>> does the overall size of graph represent? 500ms? 1000ms?
>
> The graphs are scaled based on the maximum end of the range (i.e. 1000ms
> in your example). I don't know the exact scale used, but it goes from 0
> at the left to some value at the right which is constant within a block.

Ah, finally figured it out.

The graph is scaled so that the max time of the block (1000ms) is 90% of
the way across the image. Don't ask why, it's not my code! :P

--
James Ross <[hidden email]>
ChatZilla Developer
_______________________________________________
dev-apps-js-debugger mailing list
[hidden email]
https://lists.mozilla.org/listinfo/dev-apps-js-debugger