Something that seems odd to me about these two graphs (and I've been seeing
this pretty consistantly): Hitting the same url rockets the memory
usage to ~30 in the first 3-5 hits. Hitting different urls causes
a much more gradual increase of memory consumption over the first 50 (or
more) urls, even though the average slope ends up steeper over the full
500 hits. This seems counterintuitive to me. Does it make sense
to others?
Curt
Tracy Walker wrote:
Here are the memory consuption test results from test run on builds
from
a few days ago.
First graph shows memory consumption while loading these
41 urls over
ten repeated cylces.
Next week we will begin publishing graph of daily slope and y-intercept
values over time as they are accumulated for each daily run
The following graph shows one site, in this case cnn.com loaded 500
times in a row:
