Ryzen Owners Zen Garden

[LifeOnMars]

OK good to know thanks, it's on pretty much 24/7 but I do let it sleep and then wake it when I use it. Using the Ryzen balanced power plan with the standard 10 minutes display/30 minutes sleep setting.

 

[tabascosauz]

OK good to know thanks, it's on pretty much 24/7 but I do let it sleep and then wake it when I use it. Using the Ryzen balanced power plan with the standard 10 minutes display/30 minutes sleep setting.

I am right in understanding that you're essentially doing a 4.2 @ 1.325V static OC?

If that's the case, and you're seeing 1.12V in HWInfo under the SVI2 Vcore reading (which, by the way, is a lot of Vdroop lol), there's not a lot of harder tests than P95 Small FFT with AVX enabled. With the stock boosting algorithm enabled, it'll automatically cut down your clocks and Vcore if AVX is enabled because Small AVX is brutal.

You could also try tweaking a custom run length in IntelBurnTest, or use LinpackXtreme for tough CPU tests.

For 1.12V SVI2 Vcore (our AMD equivalent of die sense), it would be fine, but remember that temperature is another factor that accelerates degradation at the top end, so keep it below 85C. Also, even given how P95 behaves, you only really need to run it for an hour or two at most before you move onto a different stress test. Running 12 hours of P95 Small on end @ 85C is pretty much silicon abuse

 

[LifeOnMars]

I am right in understanding that you're essentially doing a 4.2 @ 1.325V static OC?

If that's the case, and you're seeing 1.12V in HWInfo under the SVI2 Vcore reading (which, by the way, is a lot of Vdroop lol), there's not a lot of harder tests than P95 Small FFT with AVX enabled. With the stock boosting algorithm enabled, it'll automatically cut down your clocks and Vcore if AVX is enabled because Small AVX is brutal.

You could also try tweaking a custom run length in IntelBurnTest, or use LinpackXtreme for tough CPU tests.

For 1.12V SVI2 Vcore (our AMD equivalent of die sense), it would be fine, but remember that temperature is another factor that accelerates degradation at the top end, so keep it below 85C. Also, even given how P95 behaves, you only really need to run it for an hour or two at most before you move onto a different stress test. Running 12 hours of P95 Small on end @ 85C is pretty much silicon abuse

TBH I literally just set 4200, 1.1v in the bios, 1.05 vscoc, RAM to XMP and then set cpu fan to silent mode. It certainly doesn't look static when checking in HW Info x64 clock speeds change, as do voltages.

In games it holds 4200 throughout. Benchmark figures for cinebench and CPUz look spot on. Temps during games only hit mids 60s or lower and then mid 70s very max in Mafia 3 which hammers all the cores constantly.

I'll definitely try those other tests you suggested but this quite honestly looks and feels like a very stable setup. For my own uses I think I've got lucky as I hate having to endlessly tweak in order to get stability so I'll take it.

 

[PaulieG]

Finally got the 5600x in a case with the EK AIO 240 installed. This is what I've accomplished thus far. 4.6 is rock solid. Time to see what I'll need to get 4.7 stable without heat getting in the way.

 

[freeagent]

Good luck! I am hoping for 4700+ with mine when it gets here!

We shall see though.. I am not usually that lucky like everyone else

 

[kapone32]

Good luck! I am hoping for 4700+ with mine when it gets here!

We shall see though.. I am not usually that lucky like everyone else

Mine runs at 4700 all day long

 

[PaulieG]

Mine runs at 4700 all day long

All core? Vcore, SOC and IOD voltage?

 

[Space Lynx]

All core? Vcore, SOC and IOD voltage?

i can't get 4700 stable for life of me. but i run 4625 all day long 24/7 at 1.360v, 1.150 soc. i never break 60 celsius and no downclocking. (the games im playing aren't cpu intensive, like WoW Shadowlands) im sure if i play witcher 3 or something it will jump to 70 celsius. i have a nice cooler though and high fan cruve on it.

 

[PaulieG]

My Asus ROG Strix B550 board HATES bclk over 100 so I can't overclock it to 101 like I can on my Asus TUF B550 which is odd, since it's the same manufacturer and chipset, nevermind that the STRIX is supposed to be a higher end board. Maybe I'll mess around with with underclocking BCLK if I don't like the vcore it takes to get 4.7ghz.

 

[freeagent]

Honestly, I probably won’t feel too bad if it doesn’t oc that well. My XT boosts up to 4625 on half of the cores the other half scale to 4475, heavy load 43-4400 so I could manually run 44-4500 with lowish volts, 46-4700 with a more abusive voltage, so I didn’t hang out in that region for too long. The TPU review got 4600 on the 5600x.. so I have to wonder about the reviews of 4800mhz overclocks and their actual stability.. for play that would be ok, but I am more interested in actual overclocks that you can run daily. If I have to move to water I might.. I am positive there is not much more I can do with air cooling without moving to 200cfm+ class fans. Honestly, I’m getting too old for that shit, I can barely tolerate 100cfm fans at 12v these days. If I can squeeze 4600-4800 out of it I will be happy. That’s what I’ve run my last few Intel CPUs at, and it would scratch my cpuz itch. Of course more is better.. so I will remain hopeful and keep the cynic in me gagged and bound

Also I saw that review by the Aussie and he got pretty high with something like 1.2v.. which got me wondering if nicer bins go to countries like that because of heat and power reasons while cooler regions get the rest..

 

[Space Lynx]

My Asus ROG Strix B550 board HATES bclk over 100 so I can't overclock it to 101 like I can on my Asus TUF B550 which is odd, since it's the same manufacturer and chipset, nevermind that the STRIX is supposed to be a higher end board. Maybe I'll mess around with with underclocking BCLK if I don't like the vcore it takes to get 4.7ghz.

I wasn't aware Zen 3 owners were messing around with BCLK... i have never touched that before, I thought we are not supposed to touch that?

 

[PaulieG]

I wasn't aware Zen 3 owners were messing around with BCLK... i have never touched that before, I thought we are not supposed to touch that?

I experimented some. I was able to get up to 102 to cooperate on the 3600x/TUF B550 board, which was quite handy for smaller incremental clock increases. How are you setting your clocks to get 4625? Are you talking about your boost clock?

 

[Deleted member 178884]

I only did the diagnosis or whatever is it called to see what my sample was. I presume it didn't apply any settings?

Hello, as long as you didn't apply to a profile or at startup, it should be fine. Either way thanks to @biffzinker for sharing the screenshots over here that I had sent in the discord, and to elaborate upon this that 1.5-1.55v Vcore is NOT using the protections from AMDs PBO system and manual removes these protections, 1.55v static through a single core is not the same as PBO and 1usmus removes the protections during this process and runs a Prime AVX1 test (more suitable towards RAM/IMC) for the testing process which drives the core to the ground and will cause degradation to the processor in most scenarios, Anyone using this with renoir avoid it period - unless you want your processor degraded
Some more fun of CTR2.0:

And a new article written on the matter from "Chip":

CTR Safety, Revisited

There are times when being a journalist is exciting.Your team writes something important, people engage with it and it generates a large response. Unfortunately, those are often the exact situation…

chipsandcheese.com

Either way, do not trust this tool and run PBO instead or manual with safe vcore.
Edit: (another screenshot from a renoir sample)

@trickson Care to explain why you're bombarding posts with the 'haha' reaction on? Is there something we're doing wrong with CTR2.0 for it to degrade processors and consistently damage CPUs along with shoving down 1.5V into a 3970X? Or running completely unsafe all-core load vcore?????????? Care to elaborate?

 

[Gmr_Chick]

Jesus Christ...

I was thinking of running CTR just to see what tier my CPU was but not anymore. Why the hell would any Ryzen owner risk the longevity of their CPU by using this?

 

[Space Lynx]

I experimented some. I was able to get up to 102 to cooperate on the 3600x/TUF B550 board, which was quite handy for smaller incremental clock increases. How are you setting your clocks to get 4625? Are you talking about your boost clock?

its an all core clock no boosting or downclocking. I just set all core to 4625, 1.360 man voltage with a -25 amd offset and 1.150 on soc and LLC to mode 4. been rock solid for months.

Jesus Christ...

I was thinking of running CTR just to see what tier my CPU was but not anymore. Why the hell would any Ryzen owner risk the longevity of their CPU by using this?

I think the first version was trusted, so sites advertising it for free basically just took it for granted... I think that's what happened. I have not used it, thankfully.

 

[Mussels]

My Asus ROG Strix B550 board HATES bclk over 100 so I can't overclock it to 101 like I can on my Asus TUF B550 which is odd, since it's the same manufacturer and chipset, nevermind that the STRIX is supposed to be a higher end board. Maybe I'll mess around with with underclocking BCLK if I don't like the vcore it takes to get 4.7ghz.

I saw an extra BIOS setting hidden away once, that kept boost enabled when messing with BCLK

My asus B450 went like a year with being able to OC with 103, then all of a sudden 101 a BIOS update changed and it would lock it to stock multi (37x) and drive me bonkers. a few BIOS later i found a setting that specifically said it kept boost active when messing with BCLK, and then i never saw that setting again.
It could be one of those stupid ones that you have to set the OC, reboot, and only then does it appear (ECO mode works like that, gotta enable PBO and reboot before it appears)

 

[8BitZ80]

Jesus Christ...

I was thinking of running CTR just to see what tier my CPU was but not anymore. Why the hell would any Ryzen owner risk the longevity of their CPU by using this?

The results aren't very surprising anyway. It seems to be that six cores are usually Bronze, with the eight core chips as Silver. I'd imagine that Gold samples are reserved for Threadripper and Epyc.

 

[tabascosauz]

Hello, as long as you didn't apply to a profile or at startup, it should be fine. Either way thanks to @biffzinker for sharing the screenshots over here that I had sent in the discord, and to elaborate upon this that 1.5-1.55v Vcore is NOT using the protections from AMDs PBO system and manual removes these protections, 1.55v static through a single core is not the same as PBO and 1usmus removes the protections during this process and runs a Prime AVX1 test (more suitable towards RAM/IMC) for the testing process which drives the core to the ground and will cause degradation to the processor in most scenarios, Anyone using this with renoir avoid it period - unless you want your processor degraded
Some more fun of CTR2.0:

And a new article written on the matter from "Chip":

CTR Safety, Revisited

There are times when being a journalist is exciting.Your team writes something important, people engage with it and it generates a large response. Unfortunately, those are often the exact situation…

chipsandcheese.com

Either way, do not trust this tool and run PBO instead or manual with safe vcore.
Edit: (another screenshot from a renoir sample)

Not too much a fan of that Chip article as it's pretty thin on verified sources, just parroting the same anecdote about the broken 4650G (not saying it didn't happen, just saying that you kinda need more proof than just one dude on Discord to make claims of any sort). I think the post on TPU I saw about the 4650G might have been the same dude.

I mean, yes, the obvious is obvious, 1.55V isn't safe regardless of the bullshit that Yuri spouts. But "1.55V isn't safe, Yuri's full of shit, CTR pushes 1.55V, everyone knows 1.55V isn't safe" could probably have made for a much more succinct article.

That said, without knowing the whole context about the communications between them, I'm surprised Yuri took it so personally. Kinda makes him look like an arrogant prick. He's not Jesus, not everything Ryzen revolves around him, and he's not always right (just take a look at some of the shit he writes about procODT on his "Ryzen Memory guides"). And Renoir support is brand new in CTR v2.0 (1.0 didn't support it), so you'd think he'd be more open to feedback about whether the new features work properly or not.

Maybe he thinks he's famous cause he made a dinky, often questionable timings calculator for people who can't be bothered to read other DRAM guides to learn the basics.

Jesus Christ...

I was thinking of running CTR just to see what tier my CPU was but not anymore. Why the hell would any Ryzen owner risk the longevity of their CPU by using this?

You should be fine if you only use the Diagnostic button. After all, that's the only thing that CTR really offers that's unique. I sure as hell wouldn't fucking trust him with stress testing some "OC profile" that the software determined.

At the same time, CTR did recommend "OC profiles" that were sub-4GHz for my 4650G, so I pretty much just find CTR to be an amusing pile of ____. At least DRAM Calc has membench, which is actually useful as a benchmark.

 

[Mussels]

I mean, my 3700x degraded at 1.2V
my brothers just had his 3700x start degrading at 1.3v

Maybe thats why they werent used as 3800X's, who knows. we dont even know if it was the boards VRMs, the chips, or changes in ambient weather (we had identical boards and CPUs)

 

[Space Lynx]

I mean, my 3700x degraded at 1.2V
my brothers just had his 3700x start degrading at 1.3v

Maybe thats why they werent used as 3800X's, who knows. we dont even know if it was the boards VRMs, the chips, or changes in ambient weather (we had identical boards and CPUs)

if chips are really degrading this fast something seems very wrong to me... my 2500k was able to do 4.5 ghz all core no downclocking for like 7 years straight.

 

[Mussels]

if chips are really degrading this fast something seems very wrong to me... my 2500k was able to do 4.5 ghz all core no downclocking for like 7 years straight.

thats intel. these are not intel.

All i can say is, with needing to raise volts more and more over time (i went from needing 1.2v to 1.21v to 1.25v when i moved to the 5800x) i realised something was up, and i didnt want dead chips on my hands. Was this a flaw with the 3000 series? the 3700x? the motherboard we both had? Who knows. But it made me get a 5800x with high turbo clocks and focus on that, instead.

 

[Makaveli]

The results aren't very surprising anyway. It seems to be that six cores are usually Bronze, with the eight core chips as Silver. I'd imagine that Gold samples are reserved for Threadripper and Epyc.

Nope.

 

[Gmr_Chick]

Reading that Chips article right now and boy does Yuri sound like, well, a douchebag. Just from reading his various comments in the comments section, I can tell he thinks rather highly of himself and insists the writers are "stupid users who didn't read the manual!" Gee, you guys think he doesn't like it when people criticize his program? Lol Like, forget about the fact that it's been shown to cause degradation. Nah, that's not the fault of his program -- everybody is just stupid and didn't read the manual

 

[8BitZ80]

thats intel. these are not intel.

All i can say is, with needing to raise volts more and more over time (i went from needing 1.2v to 1.21v to 1.25v when i moved to the 5800x) i realised something was up, and i didnt want dead chips on my hands. Was this a flaw with the 3000 series? the 3700x? the motherboard we both had? Who knows. But it made me get a 5800x with high turbo clocks and focus on that, instead.

You don't think it's due to bios updates? It could even be from Windows updates for all we know.

Since I upgraded my power supply I've actually found that my PBO undervolting stability has improved. Previously could only do negative 20 all cores stable, now I can do negative 30 all cores. So it could even be a power delivery degradation issue, nothing to do with the processor itself. It just seems like everyone is jumping to the conclusion that the silicon is to blame.

 

[freeagent]

Maybe I will just stick to IF and Memclk tuning..

You guys know your awesome right?