spit051261CPU seems to go OK.
spit051261To run the 10900k at max speed , you need a chiller at the least.I need 1.45V core at the least to run 5.6 and thats with chiller set down to 1C .Have to do a couple of quick benches before I have to switch the chiller off because of condensation.
STR8_AN94BALLERspit051261CPU seems to go OK. What's the silicon prediction (SP) value of this chip?
j8t2bspit051261To run the 10900k at max speed , you need a chiller at the least.I need 1.45V core at the least to run 5.6 and thats with chiller set down to 1C .Have to do a couple of quick benches before I have to switch the chiller off because of condensation.This begets the question of: What about insulation? There must be some simple techniques, such as those applied to anything from plumbing to automotive-style air conditioning systems, to make the 24/7 operation of a chiller, operating at temperatures below the ambient dew point temperature (and thus actively producing condensation), a feasible feat. My first thoughts are simply to add some pretty thick (or just multiple layers!) insulation to your tubing/your loop's plumbing. Beyond that, would it still be necessary to implement safeguards elsewhere? Where, exactly? The CPU socket, the area just around the socket/coldplate/waterblock? Or? Where else? CPU Socket/Coldplate Area:- Kneaded art eraser- Oversized thermal pad between CPU & cold-plate to avoid condensation directly above the CPU-die, and redirect any such moisture to our 'bed' of kneaded eraser and shop towel... (?)- Shop towels- 'Shop vac' or similar suction [or reversed - the exhaust] head positioned and directed very near to the socket/plate/block head?-- This could also be used as adjunct or better yet, in line with, a pretty hefty, heavy duty and designed to manage large areas dehumidifier Other general ramblingsI have read about a bunch of techniques, from 'plasti-dipping' the motherboard (or using e.g. some type of nail polish that can form a dielectric barrier/seal to ultimately 'paint' over the PCB/PCB sensitive components), to using dielectric (non-electrically conductive) grease inside the actual CPU socket - yes, that means directly inside of the socket, between/on the pins and the bottom of the CPU: the pins still make contact with the pads on the bottom of the CPU; but it adds some added moisture barrier/safety from moisture. (Allegedly/apparently.) Others will do what seems to now be colloquially known as a 'chill box'. This is an environmentally controlled box that your entire system will reside in. All of the environmental variables will be modulated per your needs so that environmental parameters associated with our largest risks are proactively monitored, managed, and kept as such that the internal temperature, humidity, potentially even pressure (in the most extreme case) match your desired fluid (or if your tubing/plumbing is already insulated well, then the cold plate's 'head temp') so you won't face the same risk of humidity/moisture under a chill-box scenario as you would if you were to, hypothetically speaking, just set the chiller to hold and maintain 0 degrees centigrade in ambient conditions like with the A/C on and set to 78(F) or so. Hopefully others have some good ideas and can add to or demolish any non-sense I've reiterated from months of lurking in XOC subzero communities. FWIW, though, I don't think we really *need* even 0 deg. C. temps so much as perhaps locking the chiller/head temp to ambient or not so far below ambient to daily some really nice, otherwise unattainable overclocks... Revisit the video of Intel at Computex or wherever it was that, in '17/'18, they demoed the W-3175X (A 28C/56T CPU) on stage running an all-core 5.0GHz overclock with a single Hailea HC-500A chiller. Out of the box, a Hailea won't let you go near 0 degrees C. without hardware modding the unit, so, were they even using insulation - did they have to run it as low as it'd go? Then again, the workload they demonstrated - Cinebench, if I recall - is pretty bursty, so they'd only be in any sort of danger without insulation for a short period of time given the power of that CPU combined with its 6 channel DRAM controller :-)
spit0512613123 - spit051261 - Intel 10900k - 5700 @ 1.6v - Apex XII Z490 - G-Skill 4000 CL15-16-16-36-2T then 4700 CL14-14-14-28-1T @ 1.95v - chiller and direct die coolingIMC is very good on this CPU.As I said earlier though , without a chiller, you are not going to get the full benefit of this CPU.