Help OC Ryzen 5 3600

[WatEagle]

Hi, as title I would like an advice on how to overclock more my cpu if it is possible. Now i'm at 4.2 Ghz on 1.3v. My mobo is the Msi b450 tomahawk max, running the latest bios. For the cooling I have an Arctic freezer 33 esport.

Now I'm fine with the results I get, temps are fine in gaming and stress test, except prime95 where I reach about 85C. How can I push it more?

 

[Zach_01]

A screenshot of HWiNFO64 during that 85C?

This kind:

 

[WatEagle]

A screenshot of HWiNFO64 during that 85C?

This kind:
View attachment 142325

As I get home I'll post

As I get home I'll post

here it is what you requested

 

[Zach_01]

You know you exceeding almost all silicon limits of the CPU right?

PPT: 88W - 99!
TDC: 60A - 60
EDC: 90A - 114!

I’m not sure how healthy that is for the longevity of the CPU...

 

[Xzibit]

Speak up. I don't think he can hear you over the cpu fan close to 3k rpm

 

[tabascosauz]

I mean, it's pretty simple...you're running fixed clocks with presumably high LLC. If it isn't stable at that voltage, it isn't stable. If it is, it is. There's not much more to it...if you want to go further, dial up the Vcore if you can handle the heat, but judging from the temps I think not.

All-core 4.2 @ a hair under 1.3v. You're treading on thin ice there if you're running it day in and day out at those settings and benching it all the time. Go past 1.3V and I can't say if your chip will still sustain the same clocks / even still be here in a year's time. Hell, I'm not even sure about 1.29v.

None of these 3000 chips are actually designed for all-core boost on 6 or 8 cores past 41.5x. None of them. Period. You can't buy the cheapest SKU in the family and expect a well-binned chip. If you had a 3600X, 3700X or 3800X, looking at the stats I'd say you have a sliver of a chance to end up with a golden chip doing 4.3GHz at 1.3v, but even then it's slim, and still inadvisable for long-term longevity. Even the 3950X comes down to 41.5x-41.25x between 6-8 cores load.

The family has only been out for less than half a year, and there are plenty of accounts of severe degradation from running fixed clocks constantly at around 1.3V and higher. Add to that temps into the 80s and higher, and well...

 

[biffzinker]

I've found the sweet spot is 4.0 GHz at a Vcore of 1.25 for doing an all cores overclocked. Anything beyond is risking the longevity of the 3600. If plan you on replacing it with a Ryzen 4xx0 (Zen 3) then don't worry about it. I would rather wait to see what Zen 3 offers in reviews then decide.

 

[WatEagle]

Ok so it’s at limit probably won’t last that long...so what can I do to oc it but safely?
set 41.5 and and?

 

[oobymach]

I mean, it's pretty simple...you're running fixed clocks with presumably high LLC. If it isn't stable at that voltage, it isn't stable. If it is, it is. There's not much more to it...if you want to go further, dial up the Vcore if you can handle the heat, but judging from the temps I think not.

All-core 4.2 @ a hair under 1.3v. You're treading on thin ice there if you're running it day in and day out at those settings and benching it all the time. Go past 1.3V and I can't say if your chip will still sustain the same clocks / even still be here in a year's time. Hell, I'm not even sure about 1.29v.

None of these 3000 chips are actually designed for all-core boost on 6 or 8 cores past 41.5x. None of them. Period. You can't buy the cheapest SKU in the family and expect a well-binned chip. If you had a 3600X, 3700X or 3800X, looking at the stats I'd say you have a sliver of a chance to end up with a golden chip doing 4.3GHz at 1.3v, but even then it's slim, and still inadvisable for long-term longevity. Even the 3950X comes down to 41.5x-41.25x between 6-8 cores load.

The family has only been out for less than half a year, and there are plenty of accounts of severe degradation from running fixed clocks constantly at around 1.3V and higher. Add to that temps into the 80s and higher, and well...

I've been pushing my 3600x over 1.3v since I got it, no degradation whatsoever (although I'm not benching it day in day out). Gaming gets it up over 60 degrees but not by much, benching with p95 pushes it up over 80 degrees but that's to be expected, it's a torture test.

As long as you're using good cooling you can really push these chips hard, max tdp is 95 degrees so as long as it's not shutting down during use you're fine.

The default voltage in bios is like 1.44v but that's using the fluctuating frequency not a set one.

 

[Zach_01]

I've been pushing my 3600x over 1.3v since I got it, no degradation whatsoever (although I'm not benching it day in day out). Gaming gets it up over 60 degrees but not by much, benching with p95 pushes it up over 80 degrees but that's to be expected, it's a torture test.

As long as you're using good cooling you can really push these chips hard, max tdp is 95 degrees so as long as it's not shutting down during use you're fine.

The default voltage in bios is like 1.44v but that's using the fluctuating frequency not a set one.

Do not compare 3600 with 3600X. Different silicon limits....

-----3600------3600X
PPT: 88W------128W
TDC: 60A-------90A
EDC: 90A------125A

When a CPU has exceed 2 out of 3 limits or all 3 then its definately not good... The silicon degradation can occur in 3 months or 6 or 9... Just because you did not see it now does not mean you will not in future or someone else wont... Do not encourage OC, just because... Each user should know the specifics of the chip and do this on their own.
These CPUs are not for OC past 4.0~4.2GHz... depending the SKU, you like it or not, thats the deal and you have to know about it, accept it, and take your own risk.
The CPU is intelligent enough (auto mode) to know when must raise clocks/voltages in any given temp or load or any other situation.
Static OC kills its protection and silicon management...

 

[oobymach]

Do not compare 3600 with 3600X. Different silicon limits....

-----3600------3600X
PPT: 88W------128W
TDC: 60A-------90A
EDC: 90A------125A

When a CPU has exceed 2 out of 3 limits or all 3 then its definately not good... The silicon degradation can occur in 3 months or 6 or 9... Just because you did not see it now does not mean you will not in future or someone else wont... Do not encourage OC, just because... Each user should know the specifics of the chip and do this on their own.
These CPUs are not for OC past 4.0~4.2GHz... depending the SKU, you like it or not, thats the deal and you have to know about it, accept it, and take your own risk.
The CPU is intelligent enough (auto mode) to know when must raise clocks/voltages in any given temp or load or any other situation.
Static OC kills its protection and silicon management...

Yes and no, 1.3v isn't going to kill your cpu, and in all my years of overclocking I have yet to see silicon degradation even when pushing beyond intended limits. I ran my 8350 @ 4.7ghz 1.6v for years without issue, also had my nb oc'd to 3.1ghz same system for years of abuse, mind you I had a fan blowing in my case and a silverstone heligon he01 cooler but that's not the point. You won't kill your cpu by overclocking (unless you use stock cooler or have no idea what you're doing, then overclocking isn't recommended).

I've seen users push past 4.3ghz with 3600's, they don't use stock cooling obviously but it's well within tolerances. If using stock cooling you shouldn't push past 4.2ghz but again, if you have a decent cooler you can easily achieve higher speeds with minimal if not zero risk to your cpu.

 

[Zach_01]

Comparing ZEN in general and particularly ZEN2 with the FXs is irrelevant. ZEN2 is something entirely new and not as sturdy as those FX.
I owned an FX8370 for 7 years and the 5 of them worked at 180~200+W. Still works fine at stock to where I sold it.

ZEN2 was made as it is for a reason. Just because a CPU's temp is under throttle limit doesnt mean its ok. Its all relative but some users tend to not understand it.
Clock, voltage, current, temp... individually or half of them, mean a big nothing... All together as a whole have a meaning that only the CPU's silicon FITness manager can fully understand... apparently

 

[tabascosauz]

Yes and no, 1.3v isn't going to kill your cpu, and in all my years of overclocking I have yet to see silicon degradation even when pushing beyond intended limits. I ran my 8350 @ 4.7ghz 1.6v for years without issue, also had my nb oc'd to 3.1ghz same system for years of abuse, mind you I had a fan blowing in my case and a silverstone heligon he01 cooler but that's not the point. You won't kill your cpu by overclocking (unless you use stock cooler or have no idea what you're doing, then overclocking isn't recommended).

I've seen users push past 4.3ghz with 3600's, they don't use stock cooling obviously but it's well within tolerances. If using stock cooling you shouldn't push past 4.2ghz but again, if you have a decent cooler you can easily achieve higher speeds with minimal if not zero risk to your cpu.

You might want to do a bit of reading. There's plenty of complaints of serious degradation in just a few months from people running fixed freq between 1.3-1.325v. After release, the Stilt speculated that 1.325v was the FIT limit, so people automatically assumed it was safe and set their 3600s to 1.325v. All without realizing that as soon as you set static clocks, the gloves come off and the chip no longer regulates itself with respect to a voltage limiter like FIT. Tales of woe ensued, and continue to manifest as people slowly learn the hard way that fixed freq is not the way to go, especially on low end SKUs and air cooling.

This is DUV 7nm Zen. It's a very different playing field out here. This is not 32nm Core, or 32nm Piledriver. Those FX handle 1.425v out of the box.

Everything is magnified on this process, including silicon variation and leakage with higher temperatures. Just because someone is running 4.3GHz @ 1.3V under water or more extreme cooling doesn't mean it's long-term safe for you to do the same on a D15. Barring silicon inconsistencies, 1.3V might very well be safe at full load 60°c, but most people can't dream of reaching those kinds of temperatures at that Vcore. At 80°c, 1.3V is a very different story.

 

[WatEagle]

Guys, I was a bit worried about degradation, and I set in the bios core multiplier and Vcore to auto...I didn’t know also because it’s the first ryzen I own and before I got only Intel. I didn’t know how sensible was 7 nm.

But still, how can I oc it without degradation if it is possible?

 

[oobymach]

I'm not wrong, I looked into it and it turns out degradation is a big issue with ryzen but 1.3v is safe. More than 1.3v isn't unless you have an aftermarket cooler and even then not much more. I'm running 1.33v so pretty sure I'm safe, though I won't be pushing it to its limits after reading this.

From reddit:

https://www.reddit.com/r/overclocking/comments/bbxot0

Guys, I was a bit worried about degradation, and I set in the bios core multiplier and Vcore to auto...I didn’t know also because it’s the first ryzen I own and before I got only Intel. I didn’t know how sensible was 7 nm.

But still, how can I oc it without degradation if it is possible?

I would start with your current oc 4.2ghz and see if you can get that stable with 1.2-1.25v and don't stress your system for hours on end with p95. A 15 min test is more than enough and I usually only run for a minute to see if oc causes immediate failure of a worker, otherwise if everything is good after a minute it's good enough to game on imo.

 

[Zach_01]

I'm not wrong, I looked into it and it turns out degradation is a big issue with ryzen but 1.3v is safe. More than 1.3v isn't unless you have an aftermarket cooler and even then not much more. I'm running 1.33v so pretty sure I'm safe, though I won't be pushing it to its limits after reading this.

From reddit:

https://www.reddit.com/r/overclocking/comments/bbxot0

I would start with your current oc 4.2ghz and see if you can get that stable with 1.2-1.25v and don't stress your system for hours on end with p95. A 15 min test is more than enough and I usually only run for a minute to see if oc causes immediate failure of a worker, otherwise if everything is good after a minute it's good enough to game on imo.

I really hope that you’re not wrong.
Still I believe there is a great misconception about auto voltage and static voltage. It’s not the same 1.33V static OC and auto boosting. The voltage regulator monitors several aspects of the silicon and can fluctuate the voltage and the clock hundreds of times within a sec as needed, not only to achieve boost but to preserve silicon also.

IMO this reverse engineering is wrong for the main reason I just explained above.
The so called safe limit of 1.33V at 85C means absolutely nothing... The boost algorithm is not that shallow. If it was like that it should not be called algorithm but just a calculator.
1.33V with 85C in what short of load and with what power current?
And even if we new that... still the static OC/voltage cannot simulate the auto regulator of chip...

 

[tabascosauz]

I'm not wrong, I looked into it and it turns out degradation is a big issue with ryzen but 1.3v is safe.

Uh...

I really don't mean to offend, but if you came to that conclusion, you clearly didn't look very well at all. That post is from 9 months ago, and draws its assumptions from Stilt's Pinnacle Ridge analysis. Pinnacle Ridge is Ryzen 2000. 9 months ago, Ryzen 3000 was not on the market.

That isn't even the aforementioned erroneous/incorrect Stilt 1.325v Ryzen 3000 analysis, which we know to be dubiously dangerous for long-term operation on Ryzen 3000.

In no way is the recommended voltage range for 14nm Ryzen 1000 or 12nm Ryzen 2000 safe for Ryzen 3000. Those who OC'd 14nm Summit Ridge with 32nm Vishera know-how burned their chips with excessive voltage. Those who OC'd 12nm Pinnacle Ridge with 14nm Summit Ridge Vcore experience degraded their chips severely. The same happens for trying to apply Pinnacle experience to 7nm Matisse. End of story.

Guys, I was a bit worried about degradation, and I set in the bios core multiplier and Vcore to auto...I didn’t know also because it’s the first ryzen I own and before I got only Intel. I didn’t know how sensible was 7 nm.

But still, how can I oc it without degradation if it is possible?

First step is to take it off of fixed frequencies. Put multiplier back on Auto, put Vcore back onto either Auto or Normal, put core LLC back to Auto. Make sure Core Performance Boost is enabled, it pretty much enables boost, if you have the setting. Start with PBO Disabled or on Auto (essentially disabled). Test and see:

- what kind of scores you are getting in CB R20, single and multi thread
- what kind of clocks, and SVI2 TFN do you get in CB R20
- what kind of clocks, and SVI2 TFN Vcore do you end up on in Prime95 Smallest and Small
- what kind of temperatures in both

Once you establish a baseline, then you can start messing with PBO.

As a Ryzen 3000 owner, you need to embrace the stock boosting algorithm. These chips are like no other. There is negligible "OC" headroom in even the better binned chips; this is 7nm. And you need to first throw away what you know about "OC" (which usually entails fixed frequencies and fixed Vcore in traditional chips) if you come from Intel.

Testing fixed frequencies on Ryzen 3000 is useful for a variety of reasons. It's useful for determining the voltage limits of your chip (ie. absolute minimum Vcore required to sustain a certain load/clocks) and where the voltage "wall" lies, thus a general idea of how the chip was binned, or with a known astronomically excellent chip that can do 4300-4400MHz at 1.25V to 1.28V under strong watercooling, or running an HTPC at lower clocks that cares more for thermals and acoustics over 100% performance.

Using fixed freqs as a quick-and-dirty way to "OC" is not one of those reasons.

 

[Zach_01]

Guys, I was a bit worried about degradation, and I set in the bios core multiplier and Vcore to auto...I didn’t know also because it’s the first ryzen I own and before I got only Intel. I didn’t know how sensible was 7 nm.

But still, how can I oc it without degradation if it is possible?

I agree with everything @tabascosauz said and like to add just to give you a perspective and some thought...

What is you every day usage of the CPU beside synthetic benchmarks and stress tests?
Are you gaming? editing? rendering?
What does 4.2GHz static OC has gave you in real life games/apps that is worth risking your CPU’s longevity. Did you see any significant benefit from it?
Again, I’m not talking about scores of synthetics, but real life workloads...

Something must be killed... It’s either static OC or CPU “life” and the choice is in user’s hands...

 

[WatEagle]

Uh...

I really don't mean to offend, but if you came to that conclusion, you clearly didn't look very well at all. That post is from 9 months ago, and draws its assumptions from Stilt's Pinnacle Ridge analysis. Pinnacle Ridge is Ryzen 2000. 9 months ago, Ryzen 3000 was not on the market.

That isn't even the aforementioned erroneous/incorrect Stilt 1.325v Ryzen 3000 analysis, which we know to be dubiously dangerous for long-term operation on Ryzen 3000.

In no way is the recommended voltage range for 14nm Ryzen 1000 or 12nm Ryzen 2000 safe for Ryzen 3000. Those who OC'd 14nm Summit Ridge with 32nm Vishera know-how burned their chips with excessive voltage. Those who OC'd 12nm Pinnacle Ridge with 14nm Summit Ridge Vcore experience degraded their chips severely. The same happens for trying to apply Pinnacle experience to 7nm Matisse. End of story.



First step is to take it off of fixed frequencies. Put multiplier back on Auto, put Vcore back onto either Auto or Normal, put core LLC back to Auto. Make sure Core Performance Boost is enabled, it pretty much enables boost, if you have the setting. Start with PBO Disabled or on Auto (essentially disabled). Test and see:

- what kind of scores you are getting in CB R20, single and multi thread
- what kind of clocks, and SVI2 TFN do you get in CB R20
- what kind of clocks, and SVI2 TFN Vcore do you end up on in Prime95 Smallest and Small
- what kind of temperatures in both

Once you establish a baseline, then you can start messing with PBO.

As a Ryzen 3000 owner, you need to embrace the stock boosting algorithm. These chips are like no other. There is negligible "OC" headroom in even the better binned chips; this is 7nm. And you need to first throw away what you know about "OC" (which usually entails fixed frequencies and fixed Vcore in traditional chips) if you come from Intel.

Testing fixed frequencies on Ryzen 3000 is useful for a variety of reasons. It's useful for determining the voltage limits of your chip (ie. absolute minimum Vcore required to sustain a certain load/clocks) and where the voltage "wall" lies, thus a general idea of how the chip was binned, or with a known astronomically excellent chip that can do 4300-4400MHz at 1.25V to 1.28V under strong watercooling, or running an HTPC at lower clocks that cares more for thermals and acoustics over 100% performance.

Using fixed freqs as a quick-and-dirty way to "OC" is not one of those reasons.

Thanks, I set everything to auto, even resetting the bios and now everything It‘s good. In cinebench I get about 3500 pt with all auto, with oc I got 3900 but I don’t want anymore risk the lifespan of the cpu.

I agree with everything @tabascosauz said and like to add just to give you a perspective and some thought...

What is you every day usage of the CPU beside synthetic benchmarks and stress tests?
Are you gaming? editing? rendering?
What does 4.2GHz static OC has gave you in real life games/apps that is worth risking your CPU’s longevity. Did you see any significant benefit from it?
Again, I’m not talking about scores of synthetics, but real life workloads...

Something must be killed... It’s either static OC or CPU “life” and the choice is in user’s hands...

Yeah I use it mostly for gaming, browsing internet for university and sometimes I edit videos just for fun.
I did test yet the difference, but on the there is a video about r5 3600 at 3600mhz vs 4200 and the difference is about 5 FPS. Now I’m happy with the performances I get, so I think I‘ll just leave everything to auto and then in the future I will consider again OC

 

[Zach_01]

Thanks, I set everything to auto, even resetting the bios and now everything It‘s good. In cinebench I get about 3500 pt with all auto, with oc I got 3900 but I don’t want anymore risk the lifespan of the cpu.


Yeah I use it mostly for gaming, browsing internet for university and sometimes I edit videos just for fun.
I did test yet the difference, but on the there is a video about r5 3600 at 3600mhz vs 4200 and the difference is about 5 FPS. Now I’m happy with the performances I get, so I think I‘ll just leave everything to auto and then in the future I will consider again OC

While scoring 3500 vs 3900 in R20 is significant difference, it does not help in gaming and you are just an occasional “just for fan” editor. It’s not really worth it.

Yeah, take videos with a grain of salt...
Do your own gaming benchmarks and see if the benefits worth the trouble. I bet you would see less than 2~3% increase from stock to 4.2GHz OC.

Instead of OC you could improve slightly the auto clock with specific manual PBO settings and let the chip do it’s thing when it can. Although the perf. gains still would be marginal.

 

[WatEagle]

While scoring 3500 vs 3900 in R20 is significant difference, it does not help in gaming and you are just an occasional “just for fan” editor. It’s not really worth it.

Yeah, take videos with a grain of salt...
Do your own gaming benchmarks and see if the benefits worth the trouble. I bet you would see less than 2~3% increase from stock to 4.2GHz OC.

Instead of OC you could improve slightly the auto clock with specific manual PBO settings and let the chip do it’s thing when it can. Although the perf. gains still would be marginal.

Yes, the difference in game benchmarks is in the best case 5%.
I enabled PBO, but I don’t understand how it works and how to set it manually

 

[Zach_01]

Post a screenshot again of HWiNFO when running CB-R20 and w'll see...
I would like to see the PPT/TDC/EDC, temp, coreVID, effective clock. The max/avg values for all of those, but you should do the following...
Start bench, then right after reset HWiNFO values and take screenshot before it ends. Bench takes about 80sec to finish on R5 3600 so you take the shot at about 70~75sec.
Should be like mine, on post #2.

You already said that all core score is 3500.
How about single score?

 

[Durvelle27]

Honestly i have a 360p and pushing for high clocks showed no benefits in gaming only benchmarks. Heavy work loads also don't improve much.

I achieved 4.275GHz on my 3600 but voltages were high and temps could get there.

Daily i settled for 4.1GHz as i could run lower volts with decent temps.

4-4.1 should be doable with no fuss

 

[WatEagle]

Post a screenshot again of HWiNFO when running CB-R20 and w'll see...
I would like to see the PPT/TDC/EDC, temp, coreVID, effective clock. The max/avg values for all of those, but you should do the following...
Start bench, then right after reset HWiNFO values and take screenshot before it ends. Bench takes about 80sec to finish on R5 3600 so you take the shot at about 70~75sec.
Should be like mine, on post #2.

You already said that all core score is 3500.
How about single score?

Here it is the screenshot during cb r20

 

[Zach_01]

Ok... I expected a little longer than 9sec run... (at least a minute) but this will do as long as future screenshots will be taken the same way.
This will be your starting point.

See that PPT/TDC/EDC values? (raw value and %)
The default CPU PBO limits are

-----default-----(yours)
PPT: 88W--------(83W)
TDC: 60A---------(46A)
EDC: 90A---------(77A)

PPT: Power Package Tracking (total CPU socket power draw)
TDC: Thermal Design Current (max Current draw when throttling >95C)
EDC: Electrical Design Current (max Current draw when normal)

Under the same workload mine was:

-----default-----(mine)
PPT: 88W--------(86W) +2
TDC: 60A---------(49A) +3
EDC: 90A---------(78A) +1

The difference between them is due to temp.
Yours 66~67C
Mine 62~63C

And all these are monitored and regulated by silicon manager. The lower the temp the more watt and ampere the CPU is allowed to draw.
But your temps are more than acceptable.
What you can do now and may try to improve things, is to cap EDC.
If you bring that 77A EDC down the silicon manager will see potential less silicon stress, and may try to increase power draw with more clock and voltage. Some how it tries to take the stress headroom created by EDC reduction.

So, go to BIOS under XFR Enhancement and set PBO to manual like this

PPT: 90
TDC: 0 (default)
EDC: 70
PBO Scalar: Auto

...and run the R20 again.