LM99

Yes, I purchased my Ryzen 5 3600 last week. As for the 1.2V static, I did not hear about that but I've only seen it mentioned in that Reddit post. I did read a post that said people who ran 1.325V experienced degradation and that someone who had read some information on TSMC 7nm node found out that the recommended voltage for that node is 1.3V. But I couldn't find the sources of those statements. This is what I think: My Ryzen 3600 can run 4.4Ghz@1.175V or even lower (temperatures around the 60's). For Prime95 small FFTs I need 1.25V. Temperatures can then spike to 96°C(!). I do not think the CPU gets this hot because of bad cooling (I have a NZXT Kraken X62 280mm AIO) but because the 7nm process is so dense that the heat can't travel fast enough through the silicon into the IHS so it doesn't matter how much cooling you have (correct me if I'm wrong). When you run workloads like this at these temperatures, you will definitely degrade your cpu. In AIDA 64 temperatures are around 77°C and for gaming they are somewhere in the high 60's. However I do not plan on running workloads like Prime95 so in this case it seems perfectly fine to run 1.25V if you have enough cooling. This ensures me that my system is stable for my workloads without the degradation and if I ever need the heavier workloads, I know that my system isn't just going to shutdown. I can then adjust the voltage and clock later to lower the temperatures. But I haven't had enough time to descently overclock the cpu so I will definitly try and lower those temperatures and voltages if I can. The easier route would be to try 4.3Ghz and lower the voltage, but to me, that's less fun

I had the exact same experience yesterday and I didn't understand it at first because other people where pushing 1.35V or more to reach 4.4Ghz. I was at 1.175V and it was still stable so I just didn't go any lower but possebly could. Make sure to run a Prime95 test with the small FFts. The system was stable for hours on a normal stress test but with Prime95 small FFts I had to bump the voltage up to 1.25V.

It is indeed a good piece of silicon but I still find it strange that the heat transfer between the die and the IHS is so poor. At 1.275V with a normal stress test, the die gets to 76°C (IHS 56°C). I've seen people who ran at 1.4V with temperatures of around 80°C. If I put mine at 1.35V it gets to 85-90°C with a normal stress test. Prime 95 instantly shuts down due to the temperatures that are above 100°C.

I've been recently testing the overclocking possibilities of my Ryzen 5 3600 and came across some strange results (I used the BIOS for overclocking, not Ryzen Master). When I tried 4.4Ghz@1.35V I was surprised to see that the system booted. When stresstesting, the system was perfectly fine. I even tried 4.5Ghz@1.35V and that too worked but was a bit unstable. So back to 4.4Ghz and I started lowering the voltage until I hit 4.4Ghz@1.175V which was unreal because of the other results that I saw online. I didn't want to go any lower but possebly could... After an hour of AIDA64 stress testing, some Cinebench r15 and r20 the system seemed perfectly fine until I started Prime95. To get Prime95 to work properly I had to increase the voltage to 1.275V. I know that the small FFTs test is a real power virus, but the CPU package (so the actual CCD, the die) hit 96°C while the overall temperature of the CPU was 65°C. It seemed like the heat that the 7nm die produced couldn't be transfered fast enough through the IHS to the cooler (NZXT Kraken X62, 280mm AIO). I used Ryzen Master, HwInfo and CPUID HWMonitor to check the temperatures and voltages. Does anyone else have simular results with their Ryzen 3000 series CPU's with those high temperatures with Prime95 at those voltages or does this seem like some kind of error? Thank you in advance for your reply.

Found the problem. One of my memory sticks was broken. The strange thing is that windows gives error messages for DirectX when it actually is your memory. I looked this up and it is because Windows doesn't really know what the problem is, but because it is a graphical application it assumes that the graphics card is the problem. Anyway, new Ram solved everything. Don't think that that will be the problem on your laptop though...

Hi everyone, I've always been able to run Battlefield 1 just fine but recently the game won't launch anymore due to a pop-up message that says: DirectX function "GetDeviceRemovedReason" failed with DXGI_ERROR_HUNG_GPU: "", Driver 38528. This error is usually caused by the graphics driver crashing; try installing the latest drivers. Also, make sure you have a supported graphics card with at least 1024MB of VRAM. I have a 1050ti with the latest drivers installed, a ryzen 5 1600, asus b350 plus mobo and 8GB of ram. I've tried reïnstalling windows several times but nothing helps. Thanks.

OK, thank you very much

And do you have any idea for the pascal refresh? Or is that a fake rumour?

Well not that AMD put so much pressure on Nvidia, Vega isn't really what I expected and can you guess when those new cards could be announced? They don't ship to Belgium...

Have a look in your log files in windows, maybe you can find some more information there about this crash

Well in my country a 1050ti right now is 190 euros... Not exactly a deal lol

Hi everyone, I am planning to upgrade my gtx 760 but is it smart to buy a 1050ti or 1060 now or should I wait? I've read somewhere that in the early fall there could be a pascal refresh but there is no conformation about that... I am not waiting for Volta cause it aint coming very soon...

Yeah my German is much worse than my English, I speak Dutch natively. And OK thank you for your time!

Thanks, I will take a look at the video

Yes it is that video that made me a bit scared about those voltage spikes but I didn't quite understand it (english is not my native language) but I get it now, And I will take a look at the overclocking guide. Thank you!