第二段
/* 3th */
seems like the speed at (1.65v vcore, and 133mhz bus) tops out at 1666mhz. which is about 12.5x (actually 12.526x)...probably cause the core is at 1.66v realistic, not the nominal 1.65v.
although disappointed, i'm not surprised. it posts the cpu as an xp2000+ which should run at 1729mhz (133*13), but it runs at 1666mhz cause shorting the VSS and NC pins (like connecting the 5th L3 bridge) only unlocks 12.5x (and below) multipliers on the cpu, but nothing above that, so i guess the 13x setting in the bios is redundant, since you'll never set the multiplier higher than 12.5x using this unlocking method because of the processor itself.
at least i got some result though, (starting out with a blank screen when i first got the cpu and board together). i guess that's at least a good amount of progress, which led to the understanding (for me) of how a lot of this stuff works, and although it was a pain in the ass, it was well worth the knowledge gained. now we wait for iWill to get off its ass and release a good bios with autodetect 13.5x multipliers and above, in 0.5x increments.
so i guess now only choice is fsb changing and vcore changing. which still all depend on the the accuracy of the cpu temp monitors which i still would like confirmed if possible.
thanks.
hope you guys learned something from this. i know i learned much more than i wanted to when i decided to make this cpu work on the board.
Leon.
/* 4rd */
Update:
i haver removed my jp31 and upped the fsb to 166, pci divider is auto, and stepped the multiplier down to 11x.
so now i'm running on 166x11 = 1826mhz. and the cpu temp is constant at 43 degrees. (Thermaltake 6cu+ heatsink w/fan. lapped heatsink, arctic silver 3)
the plus side to this is that now my pc2700 (333mhz) ram is actually running at 333mhz without any jumper changes for ram voltage. SWEET!. computer is stable as hell too, at 43 deg., obviously. so cpu is slightly overclocked and the ram is running at optimal speed.
further overclocking efforts will be explored later.
btw. in winXP, i dicovered that CPUCool works best as far as a hardware monitor for XP333-R hardware sensors. use it!
will keep you guys updated.
Leon
/* 5th */
...also, i noticed that the processor is now atually recognized by the bios as a 2200+, pretty cool. WCPUID 3.1 verifies that, including all speeds. i guess the bios just looks at the actual final frequency for id of posted cpu, rather than a lookup table for the fsb/multiplier combination, as i previously thought.
one strange thing. neither WCPUID nor CPU-Z recognize the mobo, they cant run the ID function or something. why? anyone know?...very strange.
Leon
/* 6th */
some pics:
initial setup: xp2200+ @ 1,666mhz (recognized as 2000+), pc2700 (333mhz)ram @ 266mhz
jp31 at 133mhz, fsb in bios at 133, multiplier @ 13x 42 deg. core temp
http://www.cloud-9.net/Leeoniya/2200+.jpg
current setup: xp2200+ @ 1,826mhz (recognized as 2200+ now), pc2700 (333mhz)ram now @ 333mhz.
jp31 removed, fsb in bios at 166, multiplier @11x 43 deg. core temp
http://www.cloud-9.net/Leeoniya/2200+2.jpg
pci dividers and voltage are both "auto" in bios.
heatsink lapping:
before:
http://www.cloud-9.net/Leeoniya/unlapped.png
after: (1000 grit + water)
http://www.cloud-9.net/Leeoniya/lapped.png
never did use that copper shim...would have degraded heat transfer from core to heatsink, as opening in heatsink is smaller than copper slug, would have made poor contact w/core.