Well since high performance mode has the same harddrive settings than the balanced power plan, that didnt solve it... i mean the freezes happen a lot less than they used to a few days ago but 5min ago i had another freeze (the more the pc is running, the less freezes i got)... but every freeze is just so god damn annoying since the pc was expensive and i want a 100% working pc :S
I tested it without oc (so default settings) which also resulted into a few freezes... dunno why stresstest run perfectley fine... currently running a p-state oc with my ryzen using 1,177 - 1,199 volt @ 3,8 Ghz
Since my first rig with the same ram also had various freezes --> do you think the ram may cause problems? although every ram stresstest resulted without errors? ( i only let the ram test run for like 3 hours, so i cant say for 100%, but since the new pc has the same problems with everything new except the ram, that can be the only explanation i have)
Anything can theoretically be a problem; the trick is isolating it, and diagnosing it. Without any spare hardware to play with, it's harder to figure out, but here we go.
Put all hardware to default (no overclocking anywhere), if you have watercooling, make 10000% sure the pump is always 100% (I'd probably just connect it directly to the PSU), and remove all non-essential hardware (random Bluetooth dongles, fancy soundcards, SATA expander, USB hubs, USB sticks, external harddrives, garmin watch, etc etc. If possible, you want PSU, motherboard, CPU, RAM, and just your main drive (but don't remove secondary drives if it's going to break things).
Make sure all components are operating in their happy place; power levels are good, and temps are good. This can all be monitored with a ton of software, but I like HWMonitor personally; you'll have all of your fanspeeds, temps, and random other data all in one easy-to-read spot. From here, you have a base-line, knowing that all of the external factors to your system are basically ok.
From here, you pick a component to test. You said you already tested RAM, but RAM "CAN" be finicky. Since you're overclocking your chip, I presume you're "overclocking" your RAM (aka: not running it at 2133), so maybe try putting all of that back to AUTO and see if anything is resolved. Any memtest should be run "overnight" (aka: a long time); I personally believe you want to at least get through your RAM 3x with 0 errors; I usually shoot for "more than 5", if I have the time.
OCCT has a PSU test, but it doesn't really test your PSU (IMO), unless you're 100% sure everything else is stable. I'd skip testing PSU, and instead see what the PSU us doing during GPU/CPU tests.
If it's not RAM, its starting to sound like videocard crashing, and rebooting. I was getting something odd with my 1070 that was similar, but when it came back up, I got an audible crash report. With my 4770k, using the iGPU, it did EXACTLY what you are talking about, with no crash report, until the most recent drivers (most recent being probably 2 years ago now). Stupidly, with both videocards, it was actually the DP cables that were at fault. Sometimes problems lie in silly places.
CPU / GPU are super easy to test, thankfully. For CPU, I use OCCT; simple, and when it detects a crash, it makes some charts showing what basically every sensor in your computer was saying when it crashed. With those charts, you're looking for fanspeeds that drop randomly, voltages that are erratic, or really high temps. For GPU, I run FurMark; gets the card nice and hot, while you get to see information on what's going on with your card; in general, it doesn't take long for the card to start doing weird stuff. Both of these software packages might be pretty old by today's standards, so maybe some Googling would get you something better, but I like them, and use them when I have issues.
To me, it doesn't sound at all like motherboard/chip (at least, not that they're faulty). It does either sound like your videocard is crashing; or your bus speeds are messed up for some reason; or something else is crashing, holding your system bus hostage, which is then starving your videocard of much wanted data.
Again thanks for your answer!
Pump is running at 100% all the time (connected it via 3 pin on the aio_pump header that lets the pump run at 100%). I already put everything to default a few days ago except the ram.. so i should definitley do that! (last time i overclocked the ram again since i thought it would be cpu-oc related), if the problems are not occuring anymore i will prolly buy another one and send these ones back... power levels according to hwmonitor (i always have it open ) are ok, temps are :
CPU : ~29-33 idle (spikes durring windows to like 40 for like a second but i read that this is normal, isnt it?) | ~ 39-43 after long gaming sessions (fans barely spinning)| ~ max 60 after dunno but i guess around 2 hours of prime95 (fans at full speed)
GPU: ~28-33 idle | ~ around max. 53 (mostly 50-51ish) after long gaming sessions | ~ also max 60 ~ 63 (heavy load aka furmark) degrees after stresstests (radiator fan full speed at 60 degrees).
I dont have anything unnecessary connected. Only mouse, keyboard, 1 x dvi ,1 x hdmi cable and my lan-cable. So from this point i know that its not temperature / voltage related..
I think i skip testing the psu since power levels are perfect and i never had a problem with be quiet! psus , especially not with their best product-line.
As said i didnt downclock the ram and will do / test that (would u personally use memtest86 or just memtest (the one that lets u test the ram in windows)
Already tested cpu via prime95 and my graphicscard with benchmarks (valley, heaven and furmark, although i dont like furmark that much)
My first ryzen pc with a 1080 ti also crashed a few times :\ (i sent it back due to the 1080 ti from zotac with dual fans was running extremley hot in general).now i have a 1080 and i doubt that both were faulty... And thats why i am really depressed right now... there must be something im not doing right (i installed the chipset driver via express installation -> is there a driver i dont need?).
Both pcs were , besides ram, completley different, yet both have the exact same error.. so there cant be anything faulty besides the ram or i am too stupid to install everything accordingly (only driver i didnt install was the raid driver since i use my ssds in ahci mode)
But im curious: if the pc would have faulty components (or if the ram couldnt handle 2933 mhz oc), shouldnt it crash in high performance mode too? I mean if its not stable, shouldnt the pc crash regardless of thewindows settings?
Im no expert but in ryzen power plan ,i let the minimal frequency run at 20% ( so around 1500mhz per core) sometimes (even in windows) could it be that the processor reacts to slow for sudden load like loading up a game or videos ? so the frequency switches too slow? (i used a guide for overclocking (was new to p-state, although i think i know it now.. the guide-creator used the same config but added a bit more voltage since his cpu was worse at oc than mine, other than that i have identical settings (guide was highly liked and ppl were running their systems fine with it)
Ok nvm, so i just started cinebench again and let it run a few times... now following happened: cinebench showed me that a bug occured and for detailed information, i should check the log-document (will check it after a restart)
I have to restart since after this test, my pc froze and now some programs look like this : Imgur: The most awesome images on the Internet
I couldnt see numbers in hw monitor / cinebench nor paint. Mabye my oc isnt as stable as i thought ?
Checked my settings again, cinebench crashed since i had the offset wrong.. so the max voltage was 1,177 instead of 1,2 , now at 1,2 it is stable again. (although with 1,177 v high performance crashed once after the 8th try while balanced mode crashed 2 times in 5 trys.) Tomorrow i will try following: lower the ram frequency and run a few stresstests again with 1.2 just to make sure that i didnt mess up the voltage settings (i changed that yesterday night so that is not the reason for the freezes)
But the question that i just cant answer still stands: if it works in high performance, shouldnt it work in balanced mode (if that is so the ram cant be faulty and the issue should be software related)
Whatever timezone you are: have a nice day / night and i will be back in a few hours! Thank your for your answers in advance!
High performance vs not-high-performance is almost assuredly some random piece of hardware not playing nice.
You need to make your system stable @ all times. Then, from there, start adding items that might cause instability. I was recently trying to help a guy who had problems going to sleep; turns out his keyboard wasn't playing nice.
Forget about the message with the imgur link, i wrote that before the post with me messing up the voltage, but since the post had a link in it, it had to be approved.
I run the ram at 2400 mhz speed now. I will post again if i have any freezes (after that i will change the speed to 2133). If you say that its almost 100% hardware related and, as said, both pcs with completley different components except ram showed the same problems, it has to be the ram then.
I will see if 2933mhz is too much for ryzen atm (seems like ryzen had ram problems although its better now). 2400 is a slight overclock and if it crashes there i put it back to standard 2133. When it crashes again, there has to be sth else causing problems (which also means i was very unlucky and both pcs had a faulty component :S)
What would u recommend for a safe 24/7 voltage for the ryzen 1700 and which load-line-cal do u use? (i use level 3 since lvl 1 or 2 has a very high vdroop (i set the offset to 1,1875 + 0,025, @ lvl1 the voltage is 1,155 volt , at lvl 2 between 1,155 - 1,177v , at lvl 3 between 1,177-1,199v)