which chip on a motherboard controls the base clock (BCLK) and the CPU multiplier? is there a crystal oscillator that supplies the base clock?
>>1086361
CPUs (and GPUs) come with their own oscillators since a long time. They even have multiple ones for different parts of the chip (for the newest GPU generations its somewhere in the ballpark of 5 if I remember correctly). I don't know how exactly it works for RAM though. I think the BCLK is also generated on the CPU package nowadays, and the board just uses multipliers of that for the RAM. Boards with wider frequency support just come with more multipliers.
>>1086365
>come with their own oscillators
on the die?? how does that work?
There is usually something relatively low frequency (tens to hundred MHz) which the PLLs on the CPU die use. Intel might have some more information available.
>>1086365
>CPUs (and GPUs) come with their own oscillators since a long time.
No, they don't, and, until someone manages to shove a crystal onto a die, they never will. On-die oscillators, even when calibrated, are not accurate enough for the tight timing requirements of the components in a PC. FFS, just LOOK at your motherboard; there's going to be at least three crystals on it, near the northbridge, PCI slots, and a couple of the I/O ports. You may be confusing the actual clock source with a phase-locked loop clock generator, which just takes an existing clock signal and outputs a higher-frequency signal based on the input clock's timing (phase), and IS located on-chip.
The base clock is generated by a crystal oscillator, and then given to northbridge. The northbridge has a PLL generator in order to produce the base clock for the FSB. The CPU multiplies this frequency again, via some additional PLL circuitry of its own in order to generate its faster internal clock. The factor of these multipliers is set by the BIOS and given to the CPU at startup, if adjustable at all.
I have no idea why you'd want to know this, given that, if you have to ask, you're probably well out of your depth.
>>1086386
Pic was supposed to be related, dammit.
>>1086386
thanks for great explanation
> The base clock is generated by a crystal oscillator
is the crystal oscillator somehow programmable, since BCLK can be changed in BIOS?
> I have no idea why you'd want to know this
learning to overclock - wanna know how shit actually works before i start monkeying around with all the knobs
>>1086394
>is the crystal oscillator somehow programmable, since BCLK can be changed in BIOS?
No, the BCLK is generated from the crystal oscillator using a multiplier and a phase locked loop, usually a single chip referred to as CLKGEN.
>>1086361
>>1086374
The crystals look like this
>>1086527
Can I use them to make meth?
Also WHAT the hell am I looking at in the OPs picture. Ice?
>>1086539
Yes
That is frost accumulating because some overclocker is using liquid nitrogen to cool his cpu