Help With 2700X overclock safe PPT TDC and EDC?

[Shadowlid]

So long story short I thought I had a 100% stable 24/7 OC of 4.35ghz. Until I was listening to some youtube videos on the way back from work and decided to download Cinebench R20 and see how my system was stacking up. As the program instructed I disabled all other programs and ran the benchmark. To my surprise I was much slower than even a stock 1700X! So I opened Hwmonitor and all of a sudden when I would run the bench mark my CPU would downclock itself to 3.0ghz on all cores?? To be fair ive not ran to many stress tests other than the one that was included in CPUID which showed that it was taxing my CPU at 100% on all threads. And on CPUID stress test I had not had any problems with the CPU down clocking and scored nicely. Temps were also not a problem would always stay below 70C even after long testing. So I opened Ryzen master to check temps and clocks to make sure it was not HWmonitor problem and then i noticed the PPT was in the red saying 141W max and 100%of the 141 was being used. After lots of googling and things I figured out what the basics of each one of these meant PPT, TDC, and EDC. Though I do not fully understand them.

This brings me to my questions I started tweaking and googling but I do not want to damage my system so Im coming to my go to community to get some real answers.

Question 1- What is the max safe amount of PPT, TDC, and EDC this CPU can handle with a 24/7 OC. I put the PPT up to a max of 175W and ran one round of Cinebench R20 and scored well though slighty lower than other 2700Xs but dont want to push it if this is unsafe. Temps then jumped to 72C on the CPU everything else well below 70c.

Question 2- What is the safe settings for all this for my Motherboard? All temps in HWmonitor for the MOBO have never went over 55C.

Question 3- I read one post while googling about degradation at above 1.35v on the CPU where someone claimed that he tested like 4 ryzen systems im wanting to say all 2700x and every system that ran above 1.35 within 6 months had a 100mhz degradation to the cpu. This was not clearly explained maybe its was talking about boost speed, or stable OC? Not sure, I kind of feel like this is bull seeing on how ive ran my I7 system 24/7 OCd close to the max safe voltage with no problems after 5 years and its still running strong. But just wanted to make sure.

Question 4 - Why would Cinebench R20 cause these symptoms but not CPUID stress test if each of them tax the CPU to 100%?

P.S. Im not above reading about this stuff so if you would link me good quality info about it all I would like to understand why and what each of these mean PPT, TDC, EDC and any other information that would help me keep my system stable and safe.


As always thank you all for the help!

Shadow

 

[trents]

When stress test utilities say they are stressing the CPU to 100% you can take that with a grain of salt. It is obvious when you stress test with several different tools and watch temps and watts that even though they all say the CPU is running at 100% they are actually placing different amounts of stress on the system. So there must be 100%, 100%+, 100%++, etc. I say that tongue n' cheek.

I would suggest using HWInfo64 to monitor temps, voltages, frequencies and power draw. It gives a lot more info than HWMonitor.

Your core temps are not the problem in throttling but your VRM components may be getting too hot, causing down-throttling. In HWInfo64, look for the line in the motherboard section called "VRMOS." That will give you the temp of the mosfets, the chief component of the motherboard VRM. You may need to put a spot fan to blow air on the VRM heatsink.

You really need to practice better stress testing before assuming your overclock is truly stable. I would suggest using the Realbench stress test for two hours and set the amount of RAM used in the test to half of the total installed on your system. The vast majority of 2700x will not achieve a stable overclock of 4.35 on any kind of safe voltage. You need to be thinking more in terms of 4.2 ghz.

 

[Zerileous]

With a 12 phase v-core section, I doubt the VRM is overheating. For PBO you can set all of those limits to max (if that's an option). I can't honestly answer you about degradation, but the boost periods are usually so short anyway I doubt it matters. Since you requested some reading material: https://www.gamersnexus.net/guides/3491-explaining-precision-boost-overdrive-benchmarks-auto-oc

Did you enable precision boost overdrive? Or are you just worried because it's red in Ryzen Master? Pretty much any overclock is probably going to bust that stock value of 142W, I would not worry about that specifically, just about voltage. If you're going to manually OC keep it to 1.35v. So yeah you can set your voltage to 1.35 if you want and start stress testing around 4.2GHz. If it fails go down 25MHz, if it passes go up 25MHz.

 

[bassnut]

There was a link to an overclocking guide here the other day posted from XS though I can't find it now. It was a different approach to Overclocking these RYZEN 3000 chips and In it they talked a little bout some of the settings you are worried about. Maybe someone will post up a link to the thread and post it is in.

 

[Shadowlid]

Thanks all for the info. I downloaded the listed programs and my VRmos never went above 52C. I ran real bench for 2 hours at 4.25ghz at 1.35V no problems temps got to around 75C max on the cpu. I just bought some new case fans yesterday because i had the cheap factory ones that come with the case in it so gonna run some more testing this week after work. Also thank you for the reading info ill read that tonight!

Again thanks all for the info and help!

Shadow

 

[Zerileous]

Nicely done. If you're happy with 4.25GHz you can try decreasing the vcore gradually to see what's going to be the lowest stable voltage at that clock. Or you can try increasing the clock by 25MHz increments until it becomes unstable. Or you can just leave it as is if you really don't want to dial it in.