Lare111

Hi! I failed the install with Kryosheet. However, I replaced thermal pads, thermal paste (again) and installed 0.5mm washers. Temps have been very solid.

I have RX 6950 XT Red Devil which has severe pump-out issues. I've tried different thermal pastes (currently using Thermalright TFX), different amounts and different spread methods. At first I get great temps but then Edge temperature starts to creep up and suddenly I hit +80C degrees instantly. Yesterday I had to repaste it again and Edge dropped from 82C to 69C and all the corners of the die were dry even though I had spread the paste everywhere. I ordered Thermal Grizzly Carbonaut pad for the GPU. Has anyone here used it? At least Carbonaut wont pump out and I don't mind if I get couple degrees higher temps than with paste. Honeywell PTM7950 is not available in my country.

After my last post my GPU died and I also decided to upgrade CPU, motherboard, RAM and cooler. Very happy with 1440p gaming performance and this thing is very cool and quiet too. Intel i5-13600KF 5.2/4.2Ghz 1.18V Noctua NH-D15S chromax.black ASRock Z790 PG Lightning 32GB Kingston Fury Beast DDR5 overclocked to 6400Mhz CL32 Gigabyte RTX 3070 Ti Gaming 8GB 2040Mhz 0.981V Samsung 980 Pro 1TB WD Blue SN550 1TB Fractal Design Ion+ 760W Platinum Fractal Design Meshify 2 Compact with 4 x Be Quiet! Silent Wings 3 120mm PWM fans

Yeah. You are right. It was pretty cheap and I was able to test it before buying. It worked fine until now. Fortunately I will get a brand new RTX 3070 replacement from the store the previous owner bought it from.

Yeah, I did that. Hopefully I'll get the replacement 3070 early next week. My friend borrowed an RTX 3060 and it works without any issues on my PC now.

I bought a second-hand RTX 3070 last month and now it has started to crash and artifact. Driver install with DDU usually fixes it for couple days and then issues start again. For example, I was playing Dying Light 2 and noticed some artifacts and then the game crashed. This happens with stock clocks too. Previously there were artifacts on both monitors even when browsing web. Is this card broken? Since driver re-install fixed it for couple days I thought it couldn't be a hardware issue. But now it's happening again with different drivers. GPU runs cool in my case and I have two separate PSU power cables (Fractal Design Ion+ 760 Platinum). Fortunately it still has warranty left.

Recently updated 1440p gaming setup. Maybe next year I could upgrade to DDR5 and get an RTX 4000 series GPU. But my PC handles all games very well now and it's also quiet. Intel Core i7-11700F Be Quiet! Dark Rock 4 Asus TUF Gaming B560 Plus Wifi 32GB DDR4 3600Mhz CL16 (Dual Rank) EVGA RTX 3070 XC3 Ultra 2115Mhz Samsung 980 Pro 1TB WD Blue SN550 1TB Fractal Design Ion+ 760W Platinum Fractal Design Meshify 2 Compact with 4 x Be Quiet! Silent Wings 3 120mm fans Lenovo G27Q-20 (1440p, 165Hz IPS) Samsung S24F350H (1080p, 60Hz, IPS) Logitech G810 Logitech G502 HyperX Cloud II

Yeah. Cheaper 3060 Ti models are sold for 350€. And since I have to RMA my 3060 Ti, I will receive a brand new one as a replacement. I think it will be easier to sell than used ones.

Secondhand market is full of them. This one has 17 months of warranty left and it's nearby so I can pick it up myself.

My RTX 3060 Ti has fan bearing noise and I need to RMA it which can take up to 2 weeks. My CPU doesn't have integrated graphics. Does it make any sense if I buy an EVGA RTX 3070 XC3 Ultra for 420€, RMA my current 3060 Ti and then sell the replacement 3060 Ti for 370€? 15% more performance for 50€ and since I play at 1440p, any performance increase is welcome. The RTX 3070 still has 17 months of warranty left. I was initially going to wait for an RTX 4070 but it seems like it won't be released this year. Also, it could be expensive and availability might be poor. Cheaper RDNA3 cards won't be available for a while either.

It's probably your 8GB RAM. Your games run smooth until RAM fills up. 16GB is the minimum for modern games nowadays. I would also upgrade CPU and motherboard too. The i5 4440 is a weak 4 core 4 thread CPU and even the cheapest 12th gen Intel i3 is much faster. There's no point in buying more DDR3 RAM anymore.

Gear 2 mode needs at least 4400Mhz RAM to match for example 3200Mhz CL14 in Gear 1 mode. So, it's not worth it unless CPU is paired with very high frequency RAM. But yeah, people in the overclock.nets forum are running their CPUs with very high SA and IO voltages. I scrolled through and no-one said anything about IMC degradation or died CPU. So, if +1.4V voltages for them has been safe, I think I'm good with 1.32V IO and 1.24V SA.

That's not the case with Rocket Lake and Gear 1. Most Rocket Lake CPUs can't go past 3600Mhz in Gear 1 mode even with very high voltages. I tried and couldn't make 3733Mhz boot even with 1.45V SA and IO. And I've read that some people don't even reach 3600Mhz on Rocket Lake. So, it's a matter of luck (and binning). Gear mode is the reason why I'm unsure about voltages. Old rules don't apply with Rocket Lake. 4000Mhz is impossible in Gear 1 and reviewers say that even 3600Mhz needs lot "juice" for the IMC. But I don't know what's really safe when they say that.

Igor's Lab said that 1.45-1.5V is the max safe voltage for SA and IO on RKL. Some other tech site stated that 1.15V is the maximum safe voltage but I can't even boot at 3200Mhz if I go under 1.2V. People also seem to push IO and SA over 1.4V in overclock.net forum so maybe it's safe. And btw, my RAM kit is Samsung D-die. It overclocks pretty well as long as I push enough VCCIO voltage. Otherwise I can't get past memory training.

I've read that 1.25V is the safe value for VCCIO for 24/7 usage. Some sites tell that 1.35V is the maximum safe voltage. But how about Rocket Lake? Gear 1 seems to need lot of juice to reach 3600Mhz and many sources report that Rocket Lake tolerates more voltage. This includes Vcore, SA and IO. Rocket Lake also splits the VCCIO to two, resulting in VCCIO and VCCIO2. Also, motherboards seem to push more voltage on auto. For example, my MSI B560 board pushes 1.25V SA and 1.31 IO when the 3200Mhz XMP is enabled. If I manually increase the frequency to 3600Mhz, SA increases to 1.3V and IO to 1.41V. Many other people also report high voltages on auto too. However, it's been almost a year since the release and I believe there would have been several reports on dead RKL CPUs if auto-voltages were really dangerous. Typicals user just enable XMP and don't look at the voltages. My 32GB 3200Mhz CL16 RAM kit can reach 3600-16-19-19-36 if I set VCCSA to 1.23V and VCCIO to 1.32V. Anything lower for the VCCIO and memory training fails during cold boots. Do you think this is safe for 24/7 usage?

Sorry for the late reply. I couldn't figure out the issue on the Phanteks P300A so I bought the Fractal Design Meshify 2 Compact instead. CPU temps are so much better now and GPU also runs couple degrees cooler. Now I have 3 x 120mm front fans blowing cool air in to the case so both GPU and CPU get enough fresh air. Only one 120mm rear fan seems to be enough for exhaust since positive air pressure pushes hot air out through top and rear.

It's blowing in the same direction as case fans.

It's definitely exhausting. It's visible in the picture too since the fan frame points backwards.

GPU was hitting the 216W power limit constantly during the paused benchmark so it can't really produce any more heat. Clock speed was pretty stable 1935Mhz. I'm still running tests and no matter how slow I run the upper front fan, it always increases GPU temps. That's just so weird.

I have an i7-11700F and an RTX 3060 Ti in a Phanteks P300A case. The case is equipped with 3 x Be Quiet! Silent Wings 3 120mm fans. They are actually very good fans. Quiet, good looking and have decent static pressure for the front mesh panel. I have blocked the top fan place with sound dampening material since it reduced noise levels considerably and didn't affect temps. However, the GPU has been running slightly hotter than I would have thought and removing the side panel always reduces the temperature. My GPU exhausts all the hot air through side vents so the small gap between the GPU and the side panel isn't ideal. Now I did some testing. I had Unigine Heaven running in the sedond monitor and I paused the benchmark in a scenario where the GPU was heavily power limited. I set GPU fan speed to fixed 1500RPM and tested different case fan speeds in Argus Monitor. I noticed that the faster I run the upper front fan which cools the CPU, the hotter the GPU runs! These are the results: Fan speed 1500RPM = GPU temperature 75C Fan speed 1000RPM = GPU temperature 72C Fan speed 0RPM = GPU temperature 69C So, what's going on in here? Why does the fan above the GPU affect temperatures so much? I suspect that the upper fan produces airflow which prevents the hot air coming from the GPU from raising to the exhaust fan. Or maybe there's just too much positive pressure in the case which disturbs the airflow. However, the rear of the case is very open and I can feel that there's air coming through. I have circled the fan which causes high temps.

Yeah. But I like the closed top. I have added sound dampening material on top of my P300A already since I don't have any top exhaust fans.

That's what I'm afraid of too. Phanteks P600S is definitely larger, 40mm wider and also has cool features. Maybe that one would be a great case for years. It's a bit expensive though.

They are all sligthly too big for me. I was looking for Lian LI LANCOOL II Mesh Performance or Corsair 4000D Airflow since they are both on sale. But I don't know how big the difference in temps will be.

Yeah. I'm just planning to overclock. Also, I would like to keep my PC quiet.

I tried but no difference. There's a small exhaust behing the GPU but most of the hot air seems to be exhausted from side. So I think a bigger case would be the only option.