Analyzing bamboo bechmarks–2013/3/13

Please refer to the generating index.Rmd file if you want to look at the code that generated this file. It should be in the same folder as index.html (which you are seeing now); most R code will be hidden for prettiness.

Lets see which host/url pair we'll be graphing here, and the last time the benchmarks were run.

## [1] "2013-02-18 13:19:51 EST"
## [1] "dyn-209-2-223-45.dyn.columbia.edu "

And a data-cleaning note. I throw away all commits that are erroneously empty string (“”). Here is how many of those there were:

## [1] "Discarding  0  out of  131 observations. Total left: 131"

Whats happening to performance over the last ten commits?

plot of chunk unnamed-chunk-5

Lower is better. Commit label is the first seven characters of the commit sha.

And over the last three commits?

plot of chunk unnamed-chunk-6

Lower is better. Commit label is the first seven characters of the commit sha.

And just for fun, the full point set, with a log scale on the y-axis.

plot of chunk unnamed-chunk-7