no, at the end is nothing missing:
220 -> 3000
225 -> 3057
226 -> 3069
228 -> - "-" meaning: doesn't work.
... this was with multi 11.
My MDT rams are not so special. So I allways had them on "optimal". After changing them the system wouldn't boot. Right now I also use 2,5/4/4/10 (with multi 9,5 as well as that time with multi 11).
But with a multi of 9,5 (now with epox september bios) I can quickly measure it:
220 -> 3140
225 -> 3211
228 -> 3255
230 -> 3274
233 -> 3343
Result: with a smaller multi I get 10% more memory performance! (Similar situation with a friends 8RDA3+ with multi-free XP 2500+.)
Why? This was my first question.
And my second question: for daily use (no 3D-games), what makes more sense?
a. multi 11 x 220 MHz fsb -> 2420 MHz @ 1,67V with 3000 MB/s, or
b. multi 9,5 x 233 MHz fsb -> 2216 MHz @ 1,475 V with 3340 MB/s?
(BTW: Real highspeed I only "need" for encoding mp3. My reference 18:15 min song with lame 3.92 vbr:
12,5 x 200 -> 2:42 min
9,5 x 233: 3:04 min
11,5 x 166 -> 3:14 min (XP 2800+, VIA KT 400)
15 x 133 -> 3:19 min (XP-M 2600+, VIA KT 266a)
Conclusion: the time I spend on experimenting on the latest Nforce 2 is a very lot more than the time waiting for the mp3s to be done on the good old Epox 8KHA+ with the Via 266a chipset. - But it's much more fun!
