There are a couple of potentially significant performance related bug fixes in the 3.5 RC (there's a couple more being evaluated/investigated internally) and I wanted to start a thread for feedback on where we stand at this point.
I'm really hoping to keep this thread focused. That hopefully means no chatter on functional (as opposed to strictly performance) bugs, chatter and critiques on how we're approaching investigating and getting bugs fixed in dot releases, feedback on versions other than 3.5, or feedback that doesn't involve comparing 3.5 behavior to some earlier 3.x version (noting comparisons to 2.x in addition is fine, of course, but I'm focusing on evaluating how effective the changes in 3.5 are against the reported issues).
Example of the kind of feedback I'm looking to gather: "Compared to 3.4.1, when I do X it seems a bit faster. I still see high memory use and correspondingly sluggish performance when I do Y, though. My system is running Windows 7 64-bit and has 8 GB of RAM and 2 Xeon 3.0 Ghz Quad Core processors. My image files are on a striped RAID and I have two 1920x1080 monitors."
One other minor inquiry to note (at the risk of maybe doing too much in one thread): For those that do see sluggishness when brushing or applying dust spots, I'd be quite curious to know if it is improved if you clear the history states via the "X" in the upper right corner of the panel. There's a fix that's been confirmed to help that problem and we're considering putting it into the 3.5 final version, but some empirical confirmation that impacted customers would be helped by it would be worth knowing. A short addendum to the above post with "Clearing history [helps | hurts | has no obvious effect] on Develop adjustment interactive performance." will give me some useful information to use to evaluate the likely outcome of including that fix.
Thanks as always for your patience and participation,