AMD Ryzen Discussion Thread.

[Durvelle27]

Not to start a panic but I would do some reading around with the MSI before you update it. On the Titanium people were losing OC contrl and I read one guy blaming 1.4 for bricking his Carbon. Just might pay to do some research. I'd check the MSI forum if I were you

I have the Gaming Pro

 

[Johan45]

OK, I'm just passing on info I have read, up to you what you do with it.

 

[Mr.Scott]

AFIK no one has gotten any Ryzen to 4000. I've seen 3800+ but I just don't think the Chip can take it

http://hwbot.org/submission/3588222_

 

[Durvelle27]

OK, I'm just passing on info I have read, up to you what you do with it.

I'm reading buddy. Don't get your panties in a bunch

 

[Nuckles56]

http://hwbot.org/submission/3588222_

But to do it, the CPU was clocked at 800MHz

 

[Chicken Patty]

Well, I don't know if it's the replacement CPU that's a better overclocker, or the CH VI board is more stable than my X370 Taichi, but before my max stable overclock was 3.7 GHz at 1.392v. This thing is crunching along at 100% load at 3.7 GHz and 1.329v. I ain't going to complain.

 

[Mr.Scott]

But to do it, the CPU was clocked at 800MHz

Common place when benching memory frequency.
Nothing was said about being 24/7.

 

[EarthDog]

Dont bother.. this place doesnt seem to understand.

Anyway, here is a spi run at 4ghz 4k mem...http://hwbot.org/submission/3592791

 

[Cvrk]

Somebody commented
"I had the SAME problem. Dimitry, disable the overtemp protection on your motherboard. I did so and it was stable 24/7 after. It's all about the incorrect temperature reading of the CPU. Your motherboard is thinking it's overheated when it's not. Hwinfo, HWmonitor, Aida64 all reported CPU temps of 60c when I'm at 3.9ghz, but using Cam (Nzxt x62 Kraken reported 109c. I realized it's the same temp that the Tctl was reporting; which was causing the crashes). I also thought it was the memory, but swapped it out with another kit and same thing. I also did change the GPU/mobo/another 1700x to rule out all variables. But in the end, it was the overtempt bs readings."

 

[Chicken Patty]

As it is I'm still unsure what is the most accurate temperature reading software to use for Ryzen. I get all sorts of different readings... What do you guys use?

 

[R-T-B]

As it is I'm still unsure what is the most accurate temperature reading software to use for Ryzen. I get all sorts of different readings... What do you guys use?

Ryzen Master and AIDA64 here.

 

[Chicken Patty]

AIDA64 seems to match my Asus Suite III software, so that's what I'm guiding myself by now. Also for CPU vcore I get 0.650v on CPU-z always. The only two software that get matching results is AIDA64 and Asus Suite III.

 

[R-T-B]

AIDA64 seems to match my Asus Suite III software, so that's what I'm guiding myself by now. Also for CPU vcore I get 0.650v on CPU-z always. The only two software that get matching results is AIDA64 and Asus Suite III.

Sounds correct.

 

[Chicken Patty]

What should I be watching in terms of CPU temperature. I used to always go by the Tctl/Tdie reading, or the CPU Diode. Or should I be worried by the other CPU temp reading more? This is all the current software I have installed now. Just wondering which one should I be monitoring?

 

[Sempron Guy]

aida64 detects cpu temp not core. CPU tctl/tdie is your core temp. My cpu temp usually has around -15c variance vs the tdie.

 

[Chicken Patty]

aida64 detects cpu temp not core. CPU tctl/tdie is your core temp. My cpu temp usually has around -15c variance vs the tdie.

Right, that's the impression I was under. So when overclocking or knowing when to "stop" per say. Which one would we go by?

 

[Sempron Guy]

i think the 95c thermal limit imposed on most boards which is enabled by default refers to the cpu temp so I'd go with that I guess. For me though, as my tdie is significantly hotter I couldn't imagine what would my tdie be if ever my cpu temp reaches 95c. So to sum it up, I have no idea tbh which is which.

 

[Chicken Patty]

i think the 95c thermal limit imposed on most boards which is enabled by default refers to the cpu temp so I'd go with that I guess. For me though, as my tdie is significantly hotter I couldn't imagine what would my tdie be if ever my cpu temp reaches 95c. So to sum it up, I have no idea tbh which is which.

I guess for now I'll just go by my highest one just to be on the safe side.

 

[R-T-B]

On my AIDA64, the Thermal Diode reading is correct and matches Ryzen Master, which is known to obey the tctl offset on the thermal core. So that's the one I'd use if not using Ryzen Master. Mind you only the latest AIDA64 corrects the Thermal Diode reading for the 20C offset.

 

[Chicken Patty]

On my AIDA64, the Thermal Diode reading is correct and matches Ryzen Master, which is known to obey the tctl offset on the thermal core. So that's the one I'd use if not using Ryzen Master. Mind you only the latest AIDA64 corrects the Thermal Diode reading for the 20C offset.

In my case both CPU readings are within 4-5º of each other, so it's not a huge difference but I was always told the CPU Diode is the accurate reading. Anyhow, I have room on both so I'm not too worried about it for now.

 

[infrared]

I like to go off the Tdie reading in HWinfo.

 

[Chicken Patty]

I like to go off the Tdie reading in HWinfo.

That's what I was going by. But it's the lower of my readings so... just kinda gets me thinking if it's really higher than that.

 

[Johan45]

Tctl is the actual reading from the board sensor ( with 20° offset for "X" ) Tdie AFIK is not a "real" reading

 

[Chicken Patty]

So since no one really knows for sure we should just not over clock at all.

 

[R-T-B]

So since no one really knows for sure we should just not over clock at all.

Frankly, AMD doing the temp offset thing for a "consistent fan profile" always struck me as really really strange. Yeah, it's confusing as if they were trying.