I'm not sure how easy or difficult it is for a ROM cooker to make his software false report about the version of Android it's based on, and I can't think of any legitimate reason a hacker would go to the trouble without ensuring a big hoopla ensued. So it's definitely interesting that GreenComputing found evidence of Android 3.0, 3.0.1, and 3.2 running their popular benchmark software, Linpack for Android:
"Android 2.2 (FroYo) took the Linpack lead for the most runs this last week. According to my analytics, Android 2.2 was used 43.1% of the time. Android 2.1-update1, the previous leader was used 39.8% of the time. So what was used for the rest of the 17% of the time? How about 158 other ROMs! That is a lot of variety to choose.
Mixed in with the lot is some curious ROMs; Android 3.0, Android 3.0.1 and Android 3.2." (Emphasis added)
Are some Google engineers testing the performance of versions as far out (or close?) as 3.2? I don't know, but there's something comforting about seeing those numbers. Thanks to Jonathan for sending this in.