oh, good idea, R800 mode in openMSX:
220324
135344
4428
22705
13902
13841
Nowind working again (was my own fault), here's teh results on R800 on my own GT:
222366 (second run: 222105, third run: 222111)
136350 (second run: 136347, third run: 136355)
4430 (also second run and 3rd run)
22706 (also second run, third run: 22705)
13902 (second run: 13903, third run: 13902)
13853 (second run: 13852, third run 13853)
So, the first number has quite some variation and thus not very reliable as an emulator benchmark, at least. The 2nd number also has some variation, but it's a lot more stable. The other numbers only vary a tiny bit (1 in my runs). So they should be reliable.
Looks like openMSX fails mostly for the first two numbers, the rest is pretty good. (blueMSX seems to come with consistently too high numbers.)
Let's try for Z80 mode on my GT:
120149 (2nd run: 120212, 3rd run: also 120212)
77709 (also 2nd run, 3rd run: 77718)
6917 (also 2nd run, also 3rd run)
25011 (2nd run: 25009, 3rd run: also 25009)
18050 (also 2nd run, 3rd run: 18049)
8849 (also 2nd run and 3rd run)
Comparing with:
on openMSX (FS-A1GT):
120140
77702
6916
25006
18048
8848
I'm pretty proud of openMSX here! 
(And blueMSX too, it's practically the same.)
@mars,
I dont know, I just feel it sounds too good to be true
It would mean the MSX2+ Turbo is 6/5 times faster in otir, accidentaly same ratio as pal vs ntsc timings,
that feels suspicious.
When on the real machine the display ratio is still same as always on 60Hz
(or TV can tell it is 60Hz) and you hit the E, then you got a winner.
a crash? how does the crash look like? no red paint?
@mars, that pic is no crash, it is a very fast machine