I tried out 40d5, and ran into a major bug, it seems the gfps is dependent on the fps cap, so that if you put in a high fps cap, it ends up updating the screen very slowly.
For example, for testing this, I set the fps cap to 10000, and left the gfps cap at 50. When the game is paused, it actually runs at around 10000 fps, and the graphical updating is fine. But when I unpause it, and the fps drops down to around 40something, the graphical updates drop way way down. I timed it, and it's updating the screen about once every 4-5 seconds, it seems to actually be running at the fps it says it is, dwarves jump around during that time, but it's only updating the screen every 5 seconds or so. I also found, that when I was designating, the fps would drop down to about 125, and there would be a similar, but smaller delay, updating the screen about once every second and a little bit.
After that, I dropped the fps cap down, I set it at 1000, and the delays dropped by about an order of magnitude, when paused, it was as smooth as ever, but unpaused, at about 40+ fps again, it was still very jerky, only updating the screen at about 5fps or something. Dropping the fps cap down to 100 nearly solves the problem, running at 30fps, it's still not updating the screen as smoothly as it does in 40d.
So something got mixed up somewhere, it seems to only update the screen at the full amount when it's running at the fps cap, if it's running slower then that, then it updates the screen less frequently then gfps says it should. And how slow it is on updating the screen is based on the ratio between the current fps and the fps cap.
There's some strange relationship going on between the fps cap, the gfps cap, and the current fps. If I set my gfps cap to 500, and my fps cap to 10000, I get about the same effect that I do with my gfps cap at 50 and my fps cap at 1000 (except the intro movies play at about 20fps, and paused drops down to about 2500fps, unpaused fps remains about the same) If I set the fps cap to 100 while the gfps cap is at 500, it drops down to around 13-20fps, the same as the movie, so the movie is probably running slow in the earlier one, because it's capped at 100 fps, which makes the actual gfps high, and in the rest of the game the gfps is much lower then it's supposed to be.
So it looks like there's something setting the gfps wrongly, setting it based on the current fps, and the fps cap and gfps cap settings.
Other then that, my FPS improved a good amount:
On a map I've been playing with 80some dwarves I think, on a 3x15 map (I wanted a deep enough chunk of ocean to build some stuff in and got carried away)
40d with partial print on, 0:
running: 32-34ish fps
paused: 700-800ish fps
escape menu: 2000ish fps
40d with partial print off:
running: 23-25ish fps
paused: 370ish fps
40d5 with partial print on, 0:
running: 40-42ish fps
paused: 10000! fps, limited by the cap
escape menu: 2000ish fps
40d5 with partial print off:
running: 40-44ish fps
paused: 10000! fps, limited by cap
escape menu: 2500ish fps
I had another fortress that I tested, with over 200 dwarves, and lots of CPU eating things, on a 6x6 map, and it showed an improvement from around 5fps to around 6fps.
When it's paused and only doing graphical stuff, the sky's the limit it seems, oddly enough when it's on the escape menu it runs slower then paused and looking at the map. Also unusual, in 40d, having partial print on speeds things up, whereas in 40d5, having partial print on slows things down slightly. It gives a nice fps boost on this oversized map, an extra 10fps, from 33->43 or so. Also something nice for people still using crts, 40d5 runs dwarf fortress at my current refresh rate, 40d always ran it at 60hz. And I didn't have any flickering with partial print on or off with either version.
Oh yeah, my computer:
Athlon 64, running at about 2.2Ghz, Radeon 3850 AGP card, catalyst 8.10 AGP hotfix drivers.