[petsc-dev] comments on make streams

[Barry Smith]

On Mar 23, 2014, at 3:42 PM, Karl Rupp <rupp at iue.tuwien.ac.at> wrote:

> Hi,
> 
> nice work, Barry :-)
> 
> Is the plotting facility sufficiently reliable and available on target machines? The last thing we want to see is that 'make test' errors out because of a plotting failure.

   Read the code :-)  I have try’s around all the fragile places and it just prints a nice message about producing the graph when not possible.

> 
> >     I agree with Matt I don’t like the other plot overlaid on top. Too
>> difficult to understand. I also don’t like a separate plot, nor do I
>> want to remove the perfect scaling because on a decent system it
>> provides useful information. I have tried the following: the top of the
>> prefect scaling will be removed to make sure that at least 1/3 of the y
>> axis is used to display actual scaling. See attached image (of course it
>> could be changed to use 1/2 the y axis or whatever.) Is this a
>> reasonable compromise.
> 
> I like it, it is a good compromise. Two more suggestions:
> - What about using a dashed line (no points) for the ideal scaling? The current blue points suggest that this is actual measured data.

   Ok, removing the “points” from perfect speedup.

> - Add a (small) legend on the top right, because we don't expect to see actual data points there? If the legend covers the ideal scaling curve, it's not an issue.

    I don’t care about no stinking legends, feel free to add them yourself. matplotlib has already wasted enough of my time :-)

   Attached is the plot from es with 

   make streams NPMAX=32 MPIEXEC="${PETSC_DIR}/${PETSC_ARCH}/bin/mpiexec --bind-to socket”

    seems to run reasonably consistently.




  Barry
> 
> Best regards,
> Karli
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140323/7c17b082/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: scaling.png
Type: image/png
Size: 48535 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140323/7c17b082/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: scaling.png
Type: image/png
Size: 47089 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140323/7c17b082/attachment-0001.png>