OE2 Benchmarking by Users: Difference between revisions

From GO Wiki
Jump to navigation Jump to search
No edit summary
No edit summary
Line 56: Line 56:
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):<br>
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):<br>


1)  45 seconds.
1)  45 seconds.<br>
2) displays immediately
2) displays immediately<br>
3) 65 seconds
3) 65 seconds


These timings were found to be consistent over 2 replications.  
These timings were found to be consistent over 2 replications.  


Apple Desktop (Mac OS X 10.4.11)<br>
Apple Desktop (Mac OS X 10.4.11)<br>
Line 68: Line 67:
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):<br>
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):<br>


1) Uses a single CPU core only, and takes 25 seconds.  
1) Uses a single CPU core only, and takes 25 seconds. <br>
2) displays immediately
2) displays immediately<br>
3) Uses a single CPU core only, and takes 35 seconds
3) Uses a single CPU core only, and takes 35 seconds


Line 77: Line 76:




==Question arising==
==Questions arising==


1) Could we be using more than one core for these processes at a time?  
1) Could we be using more than one core for these processes at a time?  

Revision as of 09:32, 12 January 2009

Initial Reasoner Run

Jen Deegan

Windows laptop (32-bit XP):
single core Intel processor 1.73 GHz,
1.99 GB RAM (-Xmx1152M)
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):

Total reasoner time = 40051.836756 ms
time in rule (org.obo.reasoner.rbr.SubPropertyRule@1c1364a) = 0.0 ms
time in rule (org.obo.reasoner.rbr.LinkCompositionRule@1624cae) = 27629.723889 ms
time in rule (org.obo.reasoner.rbr.PropertyIntersectionRule@ec120d) = 0.0 ms
time in rule (org.obo.reasoner.rbr.IntersectionRule@954f73) = 0.481067 ms
reasoning finished in 41422 milliseconds

Apple Desktop (Mac OS X 10.4.11)
2 x 3GHz Dual-Core Intel processor
2GB RAM (-Xmx1152M)
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):

Total reasoner time = 21993.166 ms
time in rule (org.obo.reasoner.rbr.SubPropertyRule@f7ceaa) = 0.0 ms
time in rule (org.obo.reasoner.rbr.LinkCompositionRule@5df5c4) = 15683.829 ms
time in rule (org.obo.reasoner.rbr.PropertyIntersectionRule@7b28c5) = 0.0 ms
time in rule (org.obo.reasoner.rbr.IntersectionRule@b1c5c3) = 0.248 ms
reasoning finished in 22293 milliseconds

Editing with Reasoner

In order to figure out the speed of editing with the reasonser across the array of machines that we use, I am going to set up some standard tests.

I am assuming these settings on the verification manager:


And this layout, with my standard computer equipment and memory settings as noted with the results.

Editing actions to test:

1) Copy biological adhesion to be an is_a child of cellular process. The editing manoeuvre is carried out with in a single OTE panel on global selection setting. 2) Click on cell aging and see how long it takes to display. 3) Add new child to biological adhesion.

These actions should be done in order, straight after starting the application, loading the file and running the reasoner, and with no other intervening steps.

Results:

Jen Deegan

Windows laptop (32-bit XP):
single core Intel processor 1.73 GHz,
1.99 GB RAM (-Xmx1152M)
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):

1) 45 seconds.
2) displays immediately
3) 65 seconds

These timings were found to be consistent over 2 replications.

Apple Desktop (Mac OS X 10.4.11)
2 x 3GHz Dual-Core Intel processor
2GB RAM (-Xmx1152M)
Using gene_ontology_write.obo cvs revision 1.377 with Rule Based Reasoner (not including time to load file):

1) Uses a single CPU core only, and takes 25 seconds.
2) displays immediately
3) Uses a single CPU core only, and takes 35 seconds

These timings were found to be consistent over 2 replications.



Questions arising

1) Could we be using more than one core for these processes at a time?