Видео

@@kylergeston Hello! That's for the message! I'm not certain what you mean by how I ran the water pump by itself, are you asking about the tube routing or the power? The pump has its own power header connected to the power supply and then the fan header is plugged into the motherboard and speeds are controlled directly by the motherboard. I added a temperature sensor to the header on the motherboard so that I can adjust pump speed based on temperatures, although that doesn't really matter as long as the speed is consistent and the cooling is good. The water does need changing and should be done once per year. I am using distilled water and some anti-corrosion and anti-fungal additives via Mayhems XT Nuke. I have a few other additives I am testing to see if they have comparable corrosion resistance. The biggest issue is mixed metals, the copper ions that the water slowly absorbs over time like to attach themselves to the nickel plating on the cooling blocks which "Browns" the nickel a bit. Regular flushes can prevent that, although it doesn't hinder performance.

@@UnhingedSystems When you were filling up the system with water you were running the pump to get things circulated and get the air bubbles out. I'm just wondering how you ran the pump without the motherboard being turned on. Is there a way to just run the pump without powering up the motherboard? I see why you do this, in case there is a leak it doesn't take out any of the electronic components, but I just wanted to know how. So many tips & tricks in doing this project that are not all covered in the video. Keep up the great work and do let us know what additives work best.

@@kylergeston The power supply isn't connected to any of the components except for the water pump. I use a PSU test switch (which plugs into the 24pin cable) to signal the PSU to turn on. This allows the pump to run and nothing else has power 👍 I'm possibly doing a new build (on the fence on upgrading from 5950X to 9950), and if I do it will be more comprehensive as I will have help with camera work this time around 👍 It's been a hot minute since I made a video, but that's because I work all the time haha.

@@UnhingedSystems Ok Good to know. I will research on how to do that. Looking forward to the new build. I'm sure there are plenty of your viewers that would be interested in the 9950X. Me for one.

@@stanislavilkinovam It's all luck of the draw. Each chip has its own variation in spec from the manufacturing process at a microscopic level which causes different tolerances for internal resistance and voltage sensitivities. The "silicon lottery" as it has been coined for decades. Glad to hear you got yourself a golden sample 💪

​@@UnhingedSystems I also guess that Ryzen 7000 series is way more capable to go lower with this voltage curve. As far as I now most people sets 20 negative curve and they are good to go.

Thank you for the kind comment! I appreciate it!

Me too. I just followed the standard "logic". It simply does not work. PC might be stable under 100% load for days on end but will fall over even at the lightest load.

Glad you found it useful! Thanks for stopping by!

I'll be retiring my 6700K to a display case this year 🫡

@@UnhingedSystems It'll become a future office PC for me in a year or two depending on what CPUs come out. This 6700k has gotten me through my bachelors, masters, and now medical school, as well as gaming. It's done it's job fabulously.

That's so "computers". Reseating the CPU can sometimes (rarely) resolve an issue, so it sounds like you inadvertently made a better connection on the reseat. Memory is finicky, it always has been. I try hard to always choose RAM on the QVL list in the combination I am using. That has guaranteed stability for me over the years. From a rate time to time I deviate from the QVL list and about 50% of the time it doesn't run at the correct spec.

To be honest even my 3090 isn't very satisfactory at 1440p in high-end games. It is playable for sure and generally over 60FPS, but that is also paired with a 5950X CPU so my CPU isn't struggling to run the game's engine. I'm going to call it, the 6700K is officially dead as of 2024. It can't run any modern AAA "Ultra Graphic" titles without major struggling. I also can't multi-task anymore with the higher demands. For games in the 2015-2020 era, sure the 6700K and a 3070Ti could chug along a little longer. The main reason I upgraded when I did was because having Discord open while gaming was bringing my system to a halt. It just couldn't do both anymore. The 3070Ti will run older games at 1440p just fine, but any newer title it just isn't capable enough. The 3090 is still 60FPS+ in most new titles with a stronger CPU, and the 4070Ti will provide a similar experience. The 4090 is where 1440p gaming really works well in all scenarios, but 4K is still questionable in some games. 4K is amazing on the 4090 @ 60FPS+ in most cases to be sure, but I still prefer 1440p. The 5000 series should hopefully make 1440p mainstream at all levels and get us to 4K @ 120FPS+. The current spec I recommend to people for a gaming PC with some intended longevity is a current or 1-gen previous CPU with a 16GB 4070Ti and 32GB of RAM. My 6700K system hasn't powered up in over 2 years, and it probably won't until I retire it and take it for one last spin. Good luck! Have fun!

It's great when they are stable at max settings 👍

@@UnhingedSystems witch options do you recommend for "precision boost overdrive scalar"? i set my at 1x.

@@4AM551 I never exceed 1x personally. Any other settings didn't result in any kind of measurable gains anyways.

@@UnhingedSystems humm, tks!

That's odd. Have you updated the latest BIOS, reset to BIOS defaults and set your ram speed to the rated spec? 1.104V is quite low and would probably be slowing down the CPU significantly. Does the voltage ever go beyond that while under load? What motherboard are you running on?

@@UnhingedSystems I have Msi x570 carbon pro gaming board and it’s updated and when I try to run it at default setting it will just crash ,but when I run it at 1.3 v after putting this manually in bios it will run fine

@@gautamrocker69 it should be 100% stable on loaded defaults. That's very odd.

If you are stable at -20 then you are doing A-OK!

Facts! The higher EDC and TDC doesn't always result in better performance. Especially for EDC there seems to be a point where the performance actually comes down slightly when increasing EDC. This definitely depends on the type of workload on the CPU.

It won't make enough of a difference to matter to be honest. One could measure a difference, but one could not perceive a difference. It's all about that obsession for squeezing every last ounce of performance out.

My friend if THIS video is too complicated for you 😂you should just buy a pre-built and not touch anything

Yes, we shouldn't exceed 1.35V as a manual setting (Although stock the voltages will exceed 1.4V during boost). I did some testing at and above 1.4V to see if anymore performance can be gotten, but stability isn't improved beyond that.

@@bloodhound850 The thermal throttling will prevent damage, that's why it shut down / reset. What cooler do you have? I find it odd that you hit 115C.

@@bloodhound850 D15 isn't going to cut it for any kind of overclock on a 5950X. I did mine on a custom water loop.

@@bloodhound850 That is correct. With that cooler I would just run the CPU on all default settings, no overclock, no manual voltage. The 1.35V is the maximum AMD recommended for safe long-term use, though you would be better in the long run just letting the CPU manage the power settings.

@bloodhound850 also u can go into bois and change your tjmax aka thermal throttle limit from 90c factory to 80 or 85. I would recomend going with an aio water cooler from corsair. I have one and able to pbo oc let the mb control voltage. Do all core co -10. And ur 5950x will be happy

Thank you for the kind comment!

If it is stable, you are in good shape. If you start experiencing random lockup, blue screens, reboots, or shutdowns, then adjust it back up until it is stable again 👍

@@UnhingedSystems I'm running cinebench r24, playing many games and still stable without seeing a blue screen. I'm thinking of increasing the offset to be able to reduce the temperature even more

I've spent months playing with various settings, I would leave them on auto when using curve optimizer. It isn't worth the time tweaking all the settings for <1% gains IMO.

What does yours pull stock? Are you using R23? They have different weighting on the scores from version to version.

@@UnhingedSystems yes im using r23. i dont remember exactly what it was at stock but 1-2k less for sure. :P Now it sits at 29-30k with an undervolt

Interesting proposition. That might come with its own issues such as unnecessary heat and power draw, or potentially shortening the lifespans of the CPU. I know I used to do that as a regular practice for Intel all-core overclocks 10 years ago, but I'm not sure how well AMD will handle it. More testing to do.

You can totally do that. I have entire runs with several bends that have only fittings at the start end end points. Patience and time. Drawing out and measuring out your patching is important. Measure twice, cut once, bend once. The challenge is that if your measurements are off, or you aren't sure where to bend it, you can end up being really far off from your target. Tube bending requires as much spatial awareness and spatial thinking as it does a good respect for measurements and planning. Some people never get good at it, some people ace it out of the gate. Trial and error, patience, and spare tubing is recommended.

@@UnhingedSystems okeee sounds like I gotta yeet it and present it! Edit: but no for real Ty for the response I for some reason went with hardline as my first build ever and I’m only running into more and more problems. Your video explained way more than anything else I’ve seen THANKS!

What temperatures are you seeing on the GPU?

@@UnhingedSystemswhen playing monster hunter world on 4k it was getting core 84/85 memyr were fine around 88 max tho, but i had the card with 120 core clock oc and 400 men oc, i went and undervolt it and now is giving me 75 max but instead of going 4k i went back to 2k

What paste did you end up with instead? The less dense pastes tend to pump out within a year or so. I agree it can be a hassle, especially on the slippery GPU die surface, but its one and done, no need to re-paste every year.

​@@UnhingedSystems nth2 from noctua. Had it for 3 years on my CPU (granted, which is not as toasty) but when I went to repaste last week it was still pretty much like new. I honestly don't mind repasting my GPU every year as it's like a 30 minutes process (and I deep clean it every year anyways) but not having to spend 30 of those minutes just to apply the thermal paste. Again, you had difficulties with an heat gun. Imagine someone like me that does not. It's pretty much impossible on a smaller GPU die (on the CPU it might be easier). Even more so if the die is not perfectly straight (like mine is) it's impossible to spread on sides, which would create hotspots very easily. Oh and my card is not liquid cooled, so it's not as much of a pain to open up for a clean and repaste like liquid cooled ones would be.

@@davidepannone6021 Alternatively you could apply the paste to the heatsink side instead of the die, would be pretty much the same result, albeit a bit messier potentially.

Unfortunately the only event viewer errors I got were unexpected shutdowns. With my system I didn't get WHEA errors, just black screen lock ups. I actually didn't get any WHEA errors since AMD patched their platform BIOS a couple of months after launch. Used to get WHEA errors and random application crashes all the time on the early BIOS revisions even on default BIOS settings, but it's been solid since. 👍

@@UnhingedSystems weird, what agesa version? I haven't had a problem and I started doing this about a year and a half ago, although occasionally it didn't show up after a crash, but would eventually.

I've also been pushing my ram in the past few months, when I clock higher (the ram) it shows up core 0 as the problem because the cpunb voltage isn't high enough

​​​@@MarkknightexeterIt was a long way back, right after launch. My original BIOS revision was in the 0400 range so somewhere in Agesa 1.0.0.0a, before they released any bios updates for the board. The first official BIOS I could download was in the 0800 range. The BIOS it shipped with was such a hot mess I almost RMA'd the board, but they fixed the problems after several months with just a BIOS / AGESA update. I heard the new 4702 is really unstable and breaking things again...le sad.

Use a robust paste like TFX and spread it for thin but complete coverage using a spreader. I'll never not use the spreader method, it just gets much more consistent results. You want to take advantage of every square millimeter of that heat spreader.

Thank you for watching! Glad you found the video useful!

Try resetting the bios to default settings and run it without making any changes in the BIOS. If it still locks up then you may have cooling issues or lower delivery inconsistencies. Let me know!

@@UnhingedSystems Good. I loaded the factory settings of my bios. When I turn on my computer and open the ryzen master program, the PC directly freezes. I pass tests of occt, cb R23 etc. No problem. But with ryzen master I don't understand what happens. I don't think it is due to a poor cooling problem since my temperatures do not exceed 75°c under stress test. Likewise, tomorrow I will change my mx-4 thermal paste for kryonaut. I will inform you. Thank you for your time.