Since I do not want to crack open my Lenovo, I decided to use Vista's ability to lower the amount of RAM available to the OS. If you want to set this yourself, click Start | Run and run MSCONFIG. Click the Boot tab and int eh advanced options dialog there is a setting for Maximum Memory. Enable it and set it for the amount of memory Vista will be able to use.

First I set it for 1 GB and ran my tests. With RB on, start times for OneNote average about 8.1 seconds. With RB off, 11.3 seconds. A 39% boost - that was definitely visually notable and RB made a huge difference.

Then I tried 512 MB. Yes, Vista ran slower. The biggest drawback I noticed was speed when running multiple applications - running one application at a time was acceptable, but trying to have any 2 apps open at the same time was painful (but possible).

My testing again showed RB allowed OneNote to start faster (19.6 compared to 21.1 seconds) but did not show nearly the speed boost expected. I had misread our help article about readyboost suggesting to use 1-3 times the system RAM for the cache. Instead of using 1.5GB, I was manually setting the cache size to 512MB. I changed the cache size to "Let windows choose" and re-ran my tests. Results of starting were MUCH faster - I didn't bother to accurately record the times since it was visibly and noticeably faster.

Is this good performance testing? No, not really. It gives me a good "feel" for the effects of ready boost, but there is more I could do to narrow down the data:

  • I could start other applications between trials of starting OneNote to keep the cache from getting overly optimized for OneNote.
  • I was noticing a drop in start times without readyboost enabled. I figure this was due to the cache on the hard drive controller getitng optimized, but did not follow up on this. The percentage increases could drop after multiple starts of OneNote if the controller cache could get fully optimized for OneNote.
  • Readyboost is very much hardware dependent. I should have tried a variety of different memory devices (thumb drives, SD cards, compact flash, etc…) with different data transfer speed ratings. Since this test was "self financed," though, I stuck to my $20 card. It's now down to $15 at Fry's Electronics.

The list goes on and on. Faster hard drives speeds might have narrowed the gap. Different hard drive controllers likely have different cache technologies and the numbers may be different.

Still, I'm now able to quantify in some way the speed increase I get from my investment.

One final tip: my tablet has ports for all kinds of devices, and when trying to figure out which device is mapped to which drive letter, I get confused. There is a great solution to this at http://lifehacker.com/software/usb-drive/assign-a-custom-icon-to-your-flash-drive-188852.php. Basically, you can assign a custom icon to each device which will show in Windows Explorer. Thanks folks!