I was reading some blogs about OneNote and stumbled across this one:
http://www.gottabemobile.com/ReadyBoostWithASDCardYesOrNo.aspx. In it, the author (Warner Crocker) makes a comment that OneNote loads much faster with ReadyBoost on Vista enabled. I'm starting to think ahead to performance testing, and decided to see if I could reproduce and quantify his results.
Plus, I bought a 2GB Sandisk SD Card ($20) for my Lenovo tablet for the specific purpose of running ReadyBoost and wanted to see if my investment was worth it, see if the author's claim was accurate and verify if ReadyBoost is really worth the system resources dedicated to it.
My Lenovo has a dual Intel processor running at 1.83 GHz and 2GB of RAM. I do not know what the hard drive speed is, but since it will be held constant for this test, I'm not too concerned. One point that stood out immediately is my RAM: I've anecdotally heard ReadyBoost has it's biggest boost when using less RAM and 2GB is big by today's standards.
My 3 notebooks I used to test are 60MB, 4MB and 9MB. They are all local - I do not want to test with a network.
I decided to use procmon.exe from our sysinternals toolset. You can get it at http://www.microsoft.com/technet/sysinternals/ProcessesAndThreads/processmonitor.mspx. I decided to start the timer when the OneNote thread started and end when the international support loaded - again, I'm holding the test parameters constant, the UI is definitely responsive at this point and OneNote is "idle." No other apps were running.
First I tried three runs with ReadyBoost enabled. The times (in seconds) were 5.53, 5.63 and 6.21, for an average start time of about 5.79 seconds. Not too shabby and feels about right based on "UI responsiveness."
Then I unplugged my SD card and tried again. The times were 6.15, 5.99 and 6.69, for an average of about 6.28 seconds.
This is about a half second faster with ReadyBoost enabled, or about 8% faster. Not too shabby, and thankfully jibes with what Mr. Crocker experienced. Next up I want to find out if ReadyBoost has a higher payoff if I run with lower memory.
Questions or comments are always welcome,