Feeling stuck with the ASUS 970 Pro gaming setup?
Feeling stuck with the ASUS 970 Pro gaming setup?
That's really interesting! I’m not really keen on trying it myself, but for computers it’s similar to how racing affects cars—racing is about speed and control, and I enjoy both cars and racing. Could you clarify what LLC (load line control) does? I switched between low, medium, and normal settings to see the effect. Each adjustment lowered my temperature slightly, and beyond that, not much else changed. I have a basic grasp of voltage drop, which is when power drops before it recovers. LLC helps prevent that kind of sudden drop. That’s all I understand so far. Why does changing the LLC setting affect temperature? If I increase it, can I lower my voltage? Also, if you could show me a screenshot of my BIOS readings, what do those numbers mean and how do they connect to the hwinfo info? I have the manual and looked up voltage terms, but it keeps looping back to it. I’m curious about more details on heat. I imagine AMD or someone probably developed these features to ensure stability at high temps—maybe around 61 or 62 degrees for a long time. That was probably a decade or two ago. This tech feels pretty old compared to what’s available today; if something breaks, it won’t be the end of the world since I still have another processor nearby. Besides that unit, I’m trying to keep things under 90°C throttle. From what I read, a quick check says a 25% rise in heat could mean a similar drop in lifespan. Is that accurate? Or is it more unpredictable? With a max of around 70°C under heavy load, I’d expect a shorter life if pushed that hard. I’m not too concerned about lasting much longer, as I have another CPU. If I damaged something recently, I’d be devastated and upset. If you’re sticking to the 61°C limit for some reason, that’s fine too. I just got a new 120mm fan with a 9-blade design. The packaging says it runs at about 2000 RPM plus or minus 10%. I’m not sure if it can go much higher than 1300 RPM. Should I return it? It did lower my temps by around 4 degrees during a small test. With the new fan, I recorded 62.4°C at a certain voltage and 58.0°C with the old one after 45 minutes. Keep going, wherever you are, and stay safe—those little bugs can be really annoying!
Edited April 22, 2020 by Mt Bob (spelled a few times)
It's a good point about upgrading older parts—it's relatively affordable to find them. Socket 775 Intel or similar AM2-AM3 models are quite inexpensive. You might also consider using an earlier generation platform, perhaps an old slot 1 unit. As for the LLC regulator, you're correct; it can be adjusted and lower the voltage in certain situations. Of course, I'll try to explain what your voltages mean in the BIOS. Processors don't simply run at higher speeds without limits; they slow down or shut down to protect themselves from overheating. Degradation from prolonged high voltage and heat is possible, though I wouldn't say it's common. I can't provide a precise timeline based on increased temperatures, but if a CPU had a typical lifespan of 100,000 hours at normal speed, with average loads below 50% and occasional full shutdowns, you'd likely see performance drop significantly—perhaps 1/4 to 1/2 of its original capacity. That could mean anywhere from a few months to several years depending on many factors. Honestly, I'm just guessing at this point. Nobody can truly predict the outcome. To be honest, your power supply might fail completely tomorrow, and overclocking never meant to last. When we began this discussion, you were running well above 1.5 volts—likely a high-end 5GHz processor. Your cooling system was nowhere near sufficient. A useful tip: space about half an inch between your fan and the cooler to avoid airflow gaps in the center. This can help improve heat dissipation. You haven't mentioned actively cooling the VRM area directly, but placing a fan nearby could be effective. The zip-tie method is simple and works well; I've seen double-sided tape used on video card backs. A small 80mm strip would suffice. So far, it seems we're moving in the right direction for safer overclocking. Just want to confirm you're in a good range for increasing clock speeds with minimal voltage. I look forward to hearing your progress while you gradually lower the voltage at the current 4.5GHz setting. C19—staying cautious is key. I'm doing my part every day, taking precautions and caring thoroughly. Hope you're following suit. This one is definitely a challenge.
Now I think I am really stuck. I had 4.5Ghz at about 1.4375 V running 60ish on Prime 95 running small fft's Today I am trying to run stability tests using Prime95 Blended core test and I keep losing core percentages. Maybe I don't understand how this works. In an effort to try to get more stability, I cranked up my llc to high after raising my CPUV to 1.45 previously. So now I am running hot and having cores drop down to less than 100%. I tried realbench but it refuses to cooperate. It thinks it is missing files. Back to LLC. How do I figure out how much I need. Can you explain why it raises temps? In regards to the vrm's. I have my cpu fan blowing across them, and I have a 140 mm top rear fan above them. They live in about a 4" space between my cpu cooler and my 140mm rear top exhaust fan. One would be tempted to think there is a butt load of air in that area, but most of it is probably hot. I haven't added a fan directly to them yet as I am still trying to puzzle out how. I did break down and move my whole office around. I took the case out from under my desk which opened up the top ventilation to the tune of 8C cooler temps. I have no less than 8 old ass computers, think Windows Vista old. I am trying to figure out how to add one of those heat sinks to each one of my vrm's. Looking at a google explanation of what and where they are, I see there are two sets of them. Both sets on my board have some kind of aluminum cover that I think is supposed to be a passive heat sink. I think more active cooling might do the trick. I do still have the 140mm fan blowing on the back of the socket. I would expect that same air to cool the backs of the vrms as well. Apparently that is not enough. I have come to the conclusion that 4.5Ghz might have to be good enough. Unless you convince me that I can get another 200 Mhz out of it with out letting the magic smoke out of the computer. Looking forward to your response...
Calibration of the load line is something I perform at the second highest setting to keep the v-core in place. I avoid raising it unless it starts sagging too much. If your board can handle the amperage, you shouldn’t see any sagging. "Loosing core percentages" refers to how much the core moves or shifts under load. At 200MHz it mainly comes down to how well the cooling system performs. Honestly, keeping things cool lets you push higher voltages. It’s really straightforward—just need to really value cooling. Some fans work, but a H212 isn’t top-tier. It’s decent but not exceptional. Likely suitable for 140-150 TDP, similar performance to an Opteron heat pipe cooler. I know the tricks to hit 200MHz easily. Try running two cores together as a quad; this should noticeably lower temperatures based on your cooler’s behavior at different temps. This photo is from about five years ago. I used Geothermal liquid cooling with a faucet tap drain setup. Just nylon tubing and a waterblock—no extra parts. This is my preferred method for bench testing, not for everyday use. It’s meant for benchmarking and proving what I mean: lower temps mean higher clock speeds.
You need the divine authority of an O.C. to achieve that frequency at that setting. It's unclear if performance drops when under stress. Core failures seem to occur more clearly. I'm planning to test P95 on a blended run for about 10 or 20 minutes, expecting core usage to fall below 100%. Small FFTs work well for the duration. I'm trying to keep stability in mind with this blended approach.
Very distant from a deity... The water delta was notably low during this period. February's cold and the negative temperatures pushed the water near freezing, resulting in a delta between 2-4c. This was a brief event; since Prime 95 isn't AVX instructions, it aids in maintaining lower temperatures. Does Prime 95 reduce CPU usage below 90%? That's acceptable too. You might also try OCCT for additional stress testing. Feel free to select different instruction sets to achieve slightly varied temperature outcomes.
Yes, but only on one main component. One core began dropping around 23%. It started this issue even on smaller FFTs, though it was stable 36 hours prior. Adjusting the LLC and voltages hasn't resolved it. I installed a 60mm fan on my VRM and noticed a massive 5000% noise spike. The unit spins at about 4600 RPM, and the dog began whining too. I suspect the GPU is the culprit. Currently, OOCT runs smoothly at 4.5Ghz with no problems. Core performance drops from 100% to 99.2% and then recovers. That seems acceptable?
There will be some changes in performance. The operating system will shift threads across cores for better use. This might occur, but it’s uncommon for a single core to drop about 75% during Prime95 testing. It doesn’t necessarily mean there’s a problem—OCCT seems to be handling it well. Smaller fans tend to produce more noise, so you’d have preferred checking VRM temperatures before and after installing the fan. Clearly, the noise isn’t worth the concern. How are the CPU temperatures in OCCT?
These screenshots show BIOS information and settings related to your system. The Voltage image displays power levels such as +12V and CPU voltage (likely Core Temp). The DIGI page shows your LLC configuration details. Understanding these parts helps you verify hardware status and settings without making any risky changes. If you need clarity, a detailed guide or link would be very helpful.