Originally posted by: Shaq
Temp 1 shows up as CPU on MBM5. It would be nice to know what temp1 and temp2 are though. I'm showing 30 on 1 and 26 on 2 at idle. Regardless, one of them is probably external CPU and the other the NB. I'd like to find a program that will give SB temps as well. I also don't know know what vcore 2 is, I assume the NB. And for some reason it doesn't read my 12v rail correctly.
According to
http://www.hardforum.com/showp...030788516&postcount=12:
With Speedfan vcore2 is actually vdimm.
Temp 2 is related to cpu temp, best quess is that it is a thermal probe located under the cpu in the "well" in the middle of the socket. This is easy to verify as you can run some cpu stress testing software and it quickly raises temp 2. However it does not exactly correspond to the cpu core temps and reacts a bit slower, thus the conclusion its a thermister under the cpu.
Temp 1 is also a thermister somewhere on the board that is reading "system temp" I really dont quite know what exactly, its near something warm on the board but rises no where near a quickly as the cpu related ones. It is more of a general "case temp" kinda thing.
Temp 3 is hooked to nothing and is completely "wacko" disregard it.
(I don't know about the "system temp" being "case temp", since I've seen it go above CPU for a bit. That could be NB or SB, if you consider those the "system".)
Originally posted by: grateful ned
I don't understand enough about this stuff to know why it's important to have the FSB and DDR frequency 1:1 - seems a shame to run this 533 MHz memory at only 375... but can you more experienced folks convince me the bump in CPU speed is worth it? Haven't done any benchmarking yet, guess I can sort it out that way.
AFAIK, the 1:1 is generally only to make sure you're not hitting against the
memory's limits (because you probably don't want to keep thinking about the memory's speed to make sure you're not going over). After you're done overclocking the CPU/FSB, you can probably fiddle with that a bit to see how you'd like to balance things. I would think of the divider as limiting the memory's speed, not raising the CPU's speed or FSB's speed. (Besides, I'd be stunned if you managed 500+ Mhz FSB...
TomsHardware only managed 459.) There are some claims that 1:1 ends up "better" somehow, but I don't really buy those; you can always use benchmarking tools to find out for yourself. (Remember, of course, that a increase in benchmarking score is only really surprising if you either decreased or held constant the speed of both your RAM and CPU...I've seen a post claiming 1:1 is better, but that was because they then proceeded to increase FSB (and CPU and RAM) speed to compensate, which would obviously boost performance...) IMHO, I wouldn't do any divider that turns my memory into the limiting factor. For example, I'm using a divider (2.40 in the Gigabyte BIOS, or 5:6) that overclocks my RAM slightly and pretty much maxes my FSB.
For your situation... If you're running at 375 MHz FSB, 533/375 * 2 = 2.84 divider would put your RAM at specifications (but that divider doesn't exist in my BIOS... ). The 3.00 divider would put your memory at 562.5 MHz, which is quite a bit over.... (Use at own risk, etc. etc. )