Open MPI logo

MTT Devel Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all MTT Devel mailing list

Subject: Re: [MTT devel] Submission totals graph
From: Josh Hursey (jjhursey_at_[hidden])
Date: 2007-11-30 10:20:55

Attached is a version with the calendar week data. Note that there is
some flex in the calendar week since at the beginning and end of a
month a week is not necessarily defined as 7 days. I'm not sure how
this anomaly in the data can be ironed out.

Month data uses the left axis, and week and day data use the right axis.

I just committed the updated scripts with the -week option.


On Nov 30, 2007, at 9:49 AM, Josh Hursey wrote:

> On Nov 28, 2007, at 5:01 PM, Ethan Mallove wrote:
>> Nice!
>> The double y-axis threw me off there. I'm confused why the
>> the per-day-tuples plot (green line) is so jagged, I thought
>> we were more or less on steady 24-hour testing cycle.
> Yeah it threw me a bit too, but I wonder if this isn't the effect of
> not testing as much on the weekends. a couple of days at low numbers
> would cause the sudden down burst of traffic. I'm not completely sure
> what is happening here. Maybe if we viewed it by week it would smooth
> out a bit more.
>> I
>> couldn't decode the the new stats/*raw-data{.pl,.plot} files
>> to answer my own question :-)
> I used the scripts as follows (I admit they are a bit rough):
> shell$ ./ -day >
> shell$ ./ -month >
> shell$ gnuplot graph-raw-data.plot
> You may have to change the terminal in that gnuplot file from
> 'aqua' (which is OSX specific I believe) to the postscript version
> that is commented out.
> -- Josh
>> -Ethan
>> On Wed, Nov/28/2007 11:49:58AM, Josh Hursey wrote:
>>> For no other reason than I was curious to see what it
>>> would look like, I created a graph of the per day and per
>>> month submission totals to MTT and graphed it. This is the
>>> total number of MPI Install, Test Build, and Test Run
>>> submissions. Attached is the graph if anyone is
>>> interested.
>>> It looks like we peaked in July 2007, then dropped off
>>> quite a bit there after. This is likely due to Cisco's
>>> transition to a new cluster setup so I'm not terribly
>>> concerned.
>>> An interesting graph none the less.
>>> Cheers,
>>> Josh
>>> _______________________________________________
>>> mtt-devel mailing list
>>> mtt-devel_at_[hidden]
>> _______________________________________________
>> mtt-devel mailing list
>> mtt-devel_at_[hidden]
> _______________________________________________
> mtt-devel mailing list
> mtt-devel_at_[hidden]